io.net Weekly AMA 2025/01/22 日本語翻訳版（要約＋全文翻訳）
io.net x Injective: the future of DeFAI
io.net×インジェクティブ：DeFAIの未来
今週のWeekly AMAはDefAIとAIエージェントの座談会的な内容でした。
— io.net (@ionet) January 22, 2025
ご注意：
本コンテンツは、ＡＭＡの内容を音声書き起こしからＡＩ翻訳したものとなります。最新情報等はio.netのDiscordサーバでご確認ください。
AMAセッション要約（日本語版）
このAMAでは、DeFiとAI、特に「DefAI」の概念と可能性についての議論が中心となりました。InjectiveとIonetのチームが参加し、それぞれの取り組みや業界の課題、将来のビジョンについて深く掘り下げました。
1. Injectiveの背景と目標
InjectiveはCosmosベースのレイヤー1ブロックチェーンとして、金融アプリケーションに特化して設計されています。開発者が容易に分散型アプリケーション（dApps）を構築できるように、SDKを提供し、カスタムWasMコードやEVM対応のメインネットをサポートしています。特に金融の効率性や互換性を重視し、dAppsが他のプロトコルやツールとシームレスに連携できるような仕組みを目指しています。
Injectiveのチームは、現在AIエージェントを活用したDeFiユースケースを模索しており、具体的には、資本効率を最大化するためのアルゴリズムや、複雑な取引を自動化する仕組みを構築しています。また、メインネットの拡張や新しいプロジェクトの実験を奨励し、プロダクト・マーケット・フィットを模索しています。
2. DefAIの概念とユースケース
DefAIは、AIエージェントがDeFiプロトコル上で自動化された金融取引を行う仕組みです。たとえば、AIエージェントが次のような役割を果たします：
パーペチュアル（永続）取引や利回り最適化
トークンの価格動向や取引量の分析
利回りの高いステーキングや流動性プールの探索
新規参入者向けのオンボーディングプロセスの簡略化
現在、この分野は非常に初期段階にあり、具体的なユースケースがまだ明確に確立されていません。ただし、AIエージェントが資本の効率的な運用や、意思決定の自動化に役立つと考えられています。
3. Ionetの役割と価値提案
Ionetは、分散型GPUコンピューティングネットワークを提供し、AIモデルのトレーニングや推論のための計算資源を低コストかつ迅速に提供するプラットフォームです。この会議では、Ionetの以下の特徴が強調されました：
許可不要なコンピューティング：開発者がAWSやGoogle Cloudのような従来の集中型サービスに依存する必要がなく、世界中の分散型GPUネットワークにアクセスできる点。
低遅延と高効率：AIエージェントがリアルタイムで市場データを処理し、最適な投資判断を行えるようにするための高速な応答時間を提供。
コスト削減：分散型の仕組みにより、計算コストを削減し、より多くの利益をユーザーに還元できる。
スケーラビリティ：需要の増加に柔軟に対応できるインフラを提供。
Ionetは、Injective SDKとの統合を通じて、開発者がコンピュートリソースを簡単に利用できる環境を作り出しており、両者の協力関係はDFAIエコシステムの基盤を構築しています。
4. 課題と未来への展望
会議では、DefAI分野の現在の課題についても言及されました：
セキュリティと信頼性：AIエージェントに多額の資本を預ける場合、スマートコントラクトの安全性や透明性が依然として大きな懸念。
アクセシビリティ：初心者がDeFiやAIエージェントを使いやすくするためのユーザーインターフェースの改良が必要。
規制の不確実性：DeFiとAIの融合に関する規制枠組みが明確でないため、採用が進みにくい。
相互運用性：複数のチェーンやプロトコル間でAIエージェントがシームレスに動作するための標準化がまだ不十分。
これらの課題を克服するために、InjectiveとIonetはSDKやプラットフォームの改良を進めており、開発者やユーザーが簡単に利用できるエコシステムを構築することを目指しています。
5. 未来の可能性
会議の終盤では、DefAIが業界全体に与える影響や、今後のユースケースの可能性について議論されました。特に以下のポイントが強調されました：
AIエージェントが金融取引の効率性を向上させるだけでなく、ゲーム、メタバース、不動産、再生可能エネルギーなど、他の多くの産業にも適用される可能性。
新しいタレントや開発者がWeb2からWeb3に移行し、斬新なアイデアや技術を持ち込むことへの期待。
バブルの可能性はあるが、過去の技術革新（インターネット、鉄道、ドットコムバブルなど）がもたらしたように、多くの実験が次世代の成長を支える基盤となるだろうという楽観的な見方。
全体として、この会議はDefAIの可能性と現実的な課題を深く掘り下げ、InjectiveとIonetがどのようにこの新しい分野を牽引しようとしているかを明確に示すものでした。また、業界全体の未来に対するビジョンが共有され、新しいタレントやパートナーシップがどのようにこのエコシステムをさらに強化するかに焦点が当てられました。
参考リンク
INJECTIVE
io.net
全文日本語翻訳
English: hello
Japanese: こんにちは
English: everyone how goes it
Japanese: みなさん、調子はいかがですか
English: doing well doing well it's pretty cold here in New York it's snowing a lot but pretty good pretty wild the weekend as well for crypto where all the Trump coin launches yeah
Japanese: 元気です。ニューヨークはかなり寒いです。雪がたくさん降っていますが、まあまあ良いです。仮想通貨界では週末もかなり騒がしく、トランプコインの立ち上げがありましたね。
English: I don't think we'll get into much about the new meme coins being launched by the administration on this space but plenty of time to talk about that continued part as we go on I'd like to thank all of our listeners and of course our panel today we're gonna be not doing community questions but we'll be hitting them the ground hard again as always if you have community questions please join our discord and ask away and the team will answer them very special space today we have injective and our IOT team let's start with a brief overview from each of our panelists objective let's start with you
Japanese: このスペースでは、政府による新しいミームコインの立ち上げについてはあまり触れない予定ですが、今後その点について話す時間は十分にあります。リスナーの皆さん、そして本日のパネルに感謝申し上げます。本日はコミュニティ質問を取り上げることはありませんが、またしっかり取り組みます。いつものように、コミュニティ質問があればぜひDiscordに参加して質問してください。チームが回答します。本日の特別なスペースでは、InjectiveとIOTチームをお招きしています。各パネリストから簡単に概要を聞いていきましょう。では、Injectiveから始めましょう。
English: yeah so names I'll be I'm on the business development team and objective and what injective is for the listeners who might not know is we are in cosmo was in based L1 right now developing our EBM side as well so every dollar of unit would be a stable coin or a dap token that you have on either end of the cosmo wasm or EBM they'll be interoperable you use them side by side but we're effectively built for finance and we look at finance by the means that anything of value can be transferred globally for three 24 seven but yeah thank you for having having us on
Japanese: 私の名前はAbhiで、Injectiveのビジネス開発チームに所属しています。Injectiveを知らないリスナーのために簡単に説明すると、私たちはCosmos WasmをベースとしたL1で、現在はEBMサイドの開発も進めています。Cosmos WasmやEBMのどちらのエンドでも、ステーブルコインやDAppトークンが相互運用可能であり、並行して使用することができます。私たちは基本的に金融向けに構築されており、価値あるものを24時間365日、世界中で移転できることを目指しています。今日はお招きいただきありがとうございます。
English: we were just talking a moment about you know Trump coin but one of the other big narratives that's been happening in our space over the last couple of weeks is defi AI and it is been increasing for those in the audience that don't have a full understanding or maybe have just heard this word in passing could someone give a description of what that is and how both IO and injective are positioned for it
Japanese: 少し前にトランプコインについて話していましたが、ここ数週間で私たちの領域で起きているもう一つの大きな話題はDefAIです。その関心が高まっています。オーディエンスの中で、この言葉をあまり理解していない、またはちらっと耳にしただけの方のために、これが何なのか、そしてIOとInjectiveがどのように関わっているのかを説明していただけますか？
English: yeah sure so what defAI is it's effectively you have a lot of these agents that are coming on blockchain and in the real world and web to business so what defAI does is that you it allows these agents to do financial transactions on behalf of you yourself so that could be a great a perp trade on a perp exchange like we have helix or that could be you have an agent then you tell them to I have a hundred dollars of USDT how can I maximize the yield and it kind of takes that money and puts it into different yield farms across crypto but there's no one exact use case I would say I think it's very experimental right now but there's a lot of untapped uses that defAI can't handle for everyone
Japanese: はい、もちろんです。DefAIとは、基本的にブロックチェーン上や現実世界、Web2ビジネスに登場する多くのエージェントのことです。DefAIの役割は、これらのエージェントがあなたの代わりに金融取引を行えるようにすることです。たとえば、Helixのようなパーペチュアル取引所での取引や、エージェントに「100ドルのUSDTをどのように最大限活用できるか」と尋ね、その資金をさまざまな暗号資産の利回りファームに配置することができます。ただし、明確なユースケースが一つだけあるわけではありません。現時点では非常に実験的であり、DefAIがすべての人に対応するには、まだ多くの未開拓の用途があります。
English: and before we start talking about some of those use cases to the IO team we've been talking about defAI but for those that haven't tuned in how is IO connected to defAI and why is this an exciting time for us
Japanese: それでは、これらのユースケースについて話し始める前に、IOチームに質問します。これまでDefAIについて話してきましたが、まだ参加していない方のために、IOがDefAIとどのように関わっているのか、そしてなぜこれが私たちにとってエキサイティングな時期なのかを教えてください。
English: sure I can jump in on that can you guys hear me that's a little bit of trouble with the mic loud and clear awesome awesome so in terms of how IO in particular relates to defAI right I think in terms of how to describe IO itself we're sort of like the foundational layer of decentralized compute underlying all AI use cases and applications right like whether it's model training inference fine-tuning things like that so when it comes to defAI I like how it was described before in terms of you know we don't necessarily know what the use cases are out there I think people understand the possibilities of what AI agents can do in defAI today but I think the actual design space is massive and I think what the wider adoption of this will look like is still anyone's guess right so I think in terms of where IO fits into it I think actually enabling a lot of these AI agents to conduct these defi transactions on chain where every time they make a decision to sort of optimize yield across one protocol or the other or do sort of financial analyses to decide where capital is best allocated on chain that's a significant amount of inference time that each of these AI agents need to constantly do in the face of changing financial data and prices that happen constantly on chain so I think with each of those actions that an AI agent takes that's an inference workload that can be run on IO net and IO can power it using our decentralized GPU network right so I think that's an area where I think IO fits in really nicely with how defAI is enabled
Japanese: もちろんです、それについてお話しします。マイクの調子が少し悪いようですが、聞こえていますか？はっきり聞こえますね、素晴らしいです。IOが特にDefAIとどのように関わっているかについてですが、まずIO自体を説明すると、私たちはすべてのAIユースケースやアプリケーションの基盤層として、分散型コンピューティングを提供しています。例えば、モデルトレーニング、推論、ファインチューニングなどです。DefAIに関して言えば、以前の説明のように、現時点ではユースケースがどのようになるか正確には分かりませんが、AIエージェントがDefAIでできる可能性については理解が進んでいます。しかし、実際のデザインスペースは非常に広大で、これがどのように広範囲に採用されるかはまだ誰にも予測できません。IOの役割について考えると、多くのAIエージェントがオンチェーンでDeFiトランザクションを実行できるようにすること、例えば一つのプロトコルから別のプロトコルへの利回りの最適化や、資本をどこに配分するのが最適かを決定するための金融分析を行う際に必要な大量の推論処理時間を提供します。これらのAIエージェントが行うアクションごとに、推論のワークロードがIOネットワークで実行され、IOは分散型GPUネットワークを使ってそれをサポートします。このようにして、IOはDefAIを効果的にサポートする役割を果たしていると考えています。
English: I think the other aspect that's really intriguing is one sort of the actual use case royals of AI agents being able to perform actions on chain and being able to be sort of like defAI agents sort of takes hold the next aspect of evolution will really be in terms of how these AI agents differentiate amongst themselves right because in the same way today people select where is you know like wealth managers or hedge funds or mutual funds where they choose to deploy capital based on their investment returns people will start allocating their capital to specific AI agents who are able to provide outsize returns compared to other agents or the broader market in general and the developers for buildings were these more sophisticated defAI agents will likely have to do a significant amount of fine-tuning or sort of like data structure optimization or or retrieval log-menter generation and things like this to make sure that their AI agents are constantly able to beat the market and maintain an edge and are able to sort of absorb more capital from capital alligators and thus have an edge in the market right so I think with each of those AI agent fine-tuning and sort of like optimizations that need to be done there's a significant amount of sort of like fine-tuning training compute that also needs to be used which can come from IO in a decentralized manner so I think there's a lot of foundational ways that IO can sort of support the overall defAI ecosystem and movement
Japanese: もう一つ非常に興味深い側面は、AIエージェントがオンチェーンでアクションを実行し、DefAIエージェントとして機能するユースケースです。その次の進化の段階では、これらのAIエージェントがどのように差別化を図るかに焦点が当てられるでしょう。現在、人々が資本をどの資産運用マネージャー、ヘッジファンド、またはミューチュアルファンドに配分するかを投資収益に基づいて選ぶように、特定のAIエージェントに資本を配分するようになります。そのAIエージェントが他のエージェントや市場全体と比較して優れた収益を提供できる場合です。より高度なDefAIエージェントを構築するための開発者は、市場を上回り続け、競争上の優位性を維持し、資本配分者からより多くの資本を吸収するために、ファインチューニングやデータ構造の最適化、情報検索の改良などを行う必要があります。これらのファインチューニングや最適化を行うには、多くのトレーニング計算能力が必要であり、IOの分散型ネットワークから提供される可能性があります。このように、IOはDefAIエコシステム全体を基盤から支える多くの方法を提供しています。
English: let me ask it was mentioned and it's been touched upon about use cases we've seen a lot of potential and existing use cases let's define that between your degens your heavy defi users and those that might be new onboarding to defi again going back to the attention that's been happening from individuals that don't know as much about crypto that are now getting into it what are the use cases that exist for both of these groups
Japanese: ここで質問です。これまでユースケースについて触れられ、多くの可能性と既存のユースケースがあることが述べられましたが、これを分けて考えてみましょう。DegenやヘビーなDeFiユーザー、そしてDeFiに新たに参入する可能性のある人々の間で、現在注目されているのは、仮想通貨についてあまり知らない人たちがこの領域に入ってきていることです。この両方のグループに共通するユースケースは何ですか？
English: yeah so I can take that so for let's start with degens the use cases for DJ as it is right now it could be suppose there are 20 different mean points that launched on crypto on crypto lines the defAI could be an entry that gives you metrics whether that be number of holders number of sniper bots that control supply number on the amount of volume for five minute volume and a volume whatever volume kind of gathers up all those metrics and then out of those 20 mean coins it gives you what it believes has the highest probability of going higher or the next set time period so I think that's a pretty cool use case for more of the future engine space one use case for more so traditional defi users and or more risk-averse users that are crypto native if a what I've mentioned prior is yield farming optimization suppose you have X amount of money and you want to earn risk-free yield or not risk-free but you want to earn yield on that capital without taking a lot of negative drawdown on your principle so it could be a defi agent that kind of scours through all of the yield that's available on chain and tells you which one is the best for you to park your money in given a certain time period and the last one you mentioned non crypto users so I think one of the biggest problems in crypto that's persisted is that it's still difficult the onboarding is so difficult for onboarding someone who has $0 in crypto kind of first buy his or her first crypto token open up a wallet and do all that so all this can be abstracted away if there is suppose a defi agent front-end
Japanese: はい、私が回答します。まず、Degenから始めましょう。現在のDegen向けのユースケースとしては、仮想通貨上で20種類のミームコインが立ち上げられたと仮定します。DefAIは、保有者の数、供給をコントロールするスナイパーボットの数、5分間の取引量などの指標を提供し、それらを収集して分析し、20のミームコインの中から次の一定期間で最も上昇する可能性が高いものを提示することができます。これは非常にクールなユースケースだと思います。次に、より伝統的なDeFiユーザーやリスク回避型の仮想通貨ネイティブユーザー向けには、先に述べた利回りファーミングの最適化が挙げられます。例えば、一定の資金を持っていて、その資金を元本を大きく損なうことなく利回りを得たい場合、DeFiエージェントがオンチェーン上で利用可能な利回りをすべて調べて、どこに資金を置くのが最適かを提示します。最後に、非仮想通貨ユーザーについてですが、仮想通貨で長く続いている最大の課題の一つは、オンボーディングの難しさです。仮に$0から始める場合、最初の仮想通貨トークンを購入し、ウォレットを開設し、それらをすべて行うのは難しいです。しかし、これらはDeFiエージェントのフロントエンドがあればすべて抽象化することができます。
English: wallet where you can put in like you can wire in or we can put in $100 and cash and then you can tell him or you can tell the agent I want to in I want an index of the best AI token for you without having to go across and do all that kind of stuff that is really difficult for non crypto natives to do
Japanese: ウォレットに、例えば$100を送金または現金で入金し、「最適なAIトークンのインデックスが欲しい」とエージェントに伝えるだけで、非仮想通貨ネイティブの人々には難しい手間のかかる操作をすべて省くことができます。
English: and to the IOT team you've all talked about what to look for on the dev and builder side when thinking about a compute network thinking about those case studies is there anyone in particular that seems more if you're building on any of those that should be something you should be thinking about decentralized compute network so if you mentioned some use cases do any of those stand out as priorities in it at all
Japanese: IOTチームへ。これまで、開発者やビルダーの視点からコンピュートネットワークについて考える際に注目すべき点について話されてきました。これらの事例を考慮した場合、特に分散型コンピュートネットワークを構築する際に優先すべきものはありますか？ユースケースの中で特に重要だと思われるものはありますか？
English: to be honest like the way the industry is headed I think in the near future you will have companies or industry let me defi or others who use AI or basically visit out over a period of time right and when we start talk about defi and the use cases right so or even the AI industry overall I think any vertical industry that is data intensive which requires transparency or can benefit from the decentralized infrastructure basically we be the adopters of the early adopters of defi technology because like basically when you are combining the AI intelligence with the defi openness and efficiency then the sectors can unlock the new value streams and can basically drive innovation at scale right and it will be in every sector whether you talk about institutional finance asset management let's say you take an example right like hedge funds what do they do they basically can use AI today to analyze liquidity pools basically across multiple chains identify arbitrage opportunities and execute trade basically autonomously you think smart contracts right now the whole decentralized concept itself will help this defi industry to bring trust so basically more adoption right
Japanese: 正直なところ、この業界の方向性を見ると、近い将来、DeFiやその他のAIを活用する企業や業界が一定の期間で進化していくと思います。DeFiやそのユースケース、さらにはAI業界全体について話を始めると、データ集約型で透明性を必要とし、分散型インフラストラクチャーから利益を得ることができるあらゆる業界が、DeFi技術の初期採用者になると考えています。AIのインテリジェンスとDeFiのオープン性や効率性を組み合わせると、各セクターが新しい価値の流れを解き放ち、スケールでのイノベーションを推進できるようになります。例えば、機関投資の金融や資産管理など、あらゆるセクターでその可能性が広がります。具体例を挙げると、ヘッジファンドは現在、AIを使用して複数のチェーンにわたる流動性プールを分析し、アービトラージの機会を特定し、自律的に取引を実行することができます。スマートコントラクトのように、分散型の概念自体がこのDeFi業界に信頼をもたらし、より多くの採用を促進します。
English: similarly in the gaming and metaverse side same thing could be said you can come to the real estate and the corresponding tokenization of the asset which is happening there same case you can talk about supply chain you can talk about renewable energy I think any market you touch right now you will see the decentralization penetrating that particular market and AI penetrating that market more and more and I think where Injective and IONET will succeed is that the IONET like the Injective team has created this whole like AI kit which will help the builders greatly to focus on just the business problem which people want to focus on and all the other problems statements and abstractions which they normally have to build for that will go away and where IONET helps them overall as a player for both Injective and the builders on top of Injective is that whenever the startup comes in they have to really think of the scale the resilience and the cost right and IONET brings those values to the table that are inventories decentralized so whenever there's a spike of requests for that AI model which is getting built on Injective they don't have to worry about that right the startups are coming in they don't have to worry about the cost of the infrastructure on which they're building right they don't have to think about there's any centralized players which can bring the whole company down in future if it starts to get traction and if there's anything gray area right because it's it's basically really resilient to censorship so overall you'll see all these industries benefiting in in in time going forward and I think Injective and the ION partnership which has happened that will play a major role in the months to come you'll see.
Japanese: 同様に、ゲームやメタバースの分野でも同じことが言えます。不動産や資産のトークン化が進行中の事例を考えることもできますし、サプライチェーンや再生可能エネルギーについても触れることができます。現在どの市場に触れても、分散化がその市場に浸透し、AIがさらに浸透しているのが見られるでしょう。InjectiveとIONETが成功する理由は、InjectiveチームのようにIONETもAIキットを作成し、ビルダーがビジネス問題だけに集中できるようにしている点です。通常は、構築に必要な他の問題や抽象化が取り除かれます。IONETは、Injectiveの上に構築するビルダーやInjective自体にとって、スケール、回復力、コストを考える際に大きな価値を提供します。Injectiveで構築されるAIモデルへのリクエストが急増しても心配する必要がありません。新興企業が登場しても、インフラコストを心配する必要がありません。中央集権的なプレイヤーが将来会社を危険にさらす可能性もないため、全体的に見て、これらすべての業界が今後恩恵を受けるでしょう。そして、InjectiveとIONの提携が、今後数ヶ月間で重要な役割を果たすと考えています。
English: Injective let me hand it over to you we're talking about the partnership where do you see it on your side of the equation?
Japanese: Injectiveさん、ここでお話をお任せします。このパートナーシップについて話していますが、そちらの視点ではどのように見えていますか？
English: Yeah so we see it in a very similar manner so the way we see it is that the Injective SDK that we built out well I think Hildiard just comes to right now we believe that we're in the very early experimentation stage of DefAI and how do you most benefit from the early experimentation days is that you give developers you decrease the friction to experiment and to build for developers so in that case what Injective did was we built out an SDK which allows developers kind of plug and play and easily build out DeFi apps on Injective and what we really loved about this ION partnership was that so okay now they have the SDK to build out these dApps but they also need compute and like to make that easier for them so they don't have to worry about where do we source compute for inference they can just come to IONet and use that(...) in a very similar plug and play manner where it's all kind of available to them really easily so all they have to focus on is experimenting and trying new stuff out to see what works and what doesn't work.
Japanese: そうですね、私たちも非常に似た見解を持っています。Injectiveが構築したSDKは、現在私たちがDefAIの非常に初期の実験段階にあると信じている中で、開発者が実験や構築をする際の摩擦を減らすことで最大の恩恵をもたらすよう設計されています。このため、Injectiveが行ったことは、開発者がプラグアンドプレイで簡単にDeFiアプリを構築できるSDKを構築することでした。このIONとのパートナーシップで特に気に入っているのは、開発者がdAppを構築するためのSDKを持っているだけでなく、推論のための計算リソースが必要な場合に、IONetにアクセスして簡単に利用できるようにした点です。同じようにプラグアンドプレイで利用可能なため、開発者は実験や新しいことを試すことだけに集中し、何が機能し、何が機能しないかを見極めることができます。
English: And speaking of friction first I'll ask you Injective where do you see the friction happening now and where do you see that evolving as the space continues to evolve?
Japanese: 摩擦に関してですが、まずInjectiveにお聞きします。現在、どの部分に摩擦が生じていると感じていますか？また、このスペースが進化していく中で、それがどのように変化していくと考えていますか？
English: Friction in what sense? In what part of stack?
Japanese: どのような意味での摩擦でしょうか？スタックのどの部分に関してですか？
English: Yeah the friction that you mentioned that you're solving with your SDK and the partnership with ION in building compute network.
Japanese: はい、SDKやIONとの提携を通じて解決しようとしている、コンピュートネットワークの構築における摩擦についてです。
English: Yeah so I think there's two key points where the friction lies so the first key point is just getting developers the tools to build out these experiments really fast and then I think the second friction point which is much later in the kind of curve is actually finding product market fit with these dApps, DeFi dApps. So I think right now we're trying to solve the first part of the friction where we're giving developers all the tools kind of build out an experiment and increase the amount of experiments launched every day every week on an every month basis to kind of increase the probability that one of the experiments going to find PMF and kind of like block subs over time.
Japanese: そうですね、摩擦が生じている主なポイントは2つあると思います。1つ目は、開発者がこれらの実験を迅速に構築するためのツールを手に入れることです。そして2つ目は、より後の段階での課題ですが、これらのdApp、特にDeFi dAppで製品市場適合性（PMF）を見つけることです。現在、私たちは最初の摩擦を解決しようとしており、開発者に実験を構築するためのすべてのツールを提供し、日々、週ごと、月ごとに立ち上げられる実験の量を増やし、実験の中からPMFを見つけ出し、それが長期的に成功する可能性を高めようとしています。
English: And Io over to you same question.
Japanese: ではIOにも同じ質問です。
English: Yeah I can I can jump in in terms of sort of how we think about reducing friction points right
Japanese: はい、摩擦ポイントをどのように減らすかについてお話しします。
English: especially from a decentralized compute perspective.
Japanese: 特に分散型コンピュートの観点からです。
English: So I kind of really like to start to think about these problems like from the end user perspective right and what and the developer perspective in terms of what they're actually trying to optimize for when you're trying to build sort of you know the most sophisticated DefAI agent that's going to you know attract the most capital and offer the best returns and it kind of almost takes me back to sort of how hedge funds think about optimizing their returns from a technical perspective as well.
Japanese: そこで私は、これらの問題についてエンドユーザーの視点から、また開発者の視点から考え始めるのが良いと思います。具体的には、最も洗練されたDefAIエージェントを構築し、最も多くの資本を引き付け、最高の収益を提供しようとする際に、何を最適化しようとしているのかという点です。これは、ヘッジファンドが技術的観点から収益を最適化しようとする方法を思い出させます。
English: So I think there's three things that I kind of think about in terms of producing friction enabling DefAI developers to be able to sort of build the best agents they can(...) on a decentralized compute network. So I think in terms of what we offer
Japanese: DefAI開発者が分散型コンピュートネットワーク上で最良のエージェントを構築できるようにするための摩擦点について、3つのことを考えています。私たちが提供するものに関して言えば、
English: I think first and foremost is sort of the permissionless access to compute right(...) through IONEC. I think this one is fairly underappreciated in this context because you know theoretically sure an AI agent could go get compute from an AWS or Google Cloud or something like that too but I think having that permissionless access sort of really offers up a benefit in use cases where you know crypto finance and DeFi is a 24-7 market that moves very very quickly and the last thing you want is for your DeFi agent to get liquidated because you know it got stuck in a support queue at AWS right because it's going to get access to compute or had a billing error or something like that.
Japanese: まず第一に、IONEで提供される「許可不要のコンピュートアクセス」が挙げられます。これは、この文脈ではあまり評価されていないかもしれません。理論的にはAIエージェントがAWSやGoogle Cloudなどから計算リソースを取得することも可能ですが、この許可不要のアクセスは、24時間365日で非常に速く動く仮想通貨金融やDeFiの市場において大きな利点をもたらします。最後に望むのは、DeFiエージェントがAWSのサポートキューで立ち往生し、コンピュートへのアクセスが遅れたり、請求エラーが発生したりして清算されてしまうことです。
English: So I think having access to compute through a permissionless network like IONEC is going to be sort of one of the de facto things that any DeFi agent is going to need in order to be able to ensure that it can accomplish its directive in the most effective manner.
Japanese: そのため、IONEのような許可不要のネットワークを通じてコンピュートにアクセスできることは、DeFiエージェントがその目的を最も効果的に達成するために必要不可欠な要素の一つになると思います。
English: The other thing I think about and you know kind of connecting to the hedge fund example again is sort of the idea of latency right like I talked about sort of all these DeFi agents needing to have needed to perform a constant inference to analyze different market trends, different market data, prices, volumes, things like that to go to make the best investment decisions and one of the things you know like the amount of optimization around latency in the traditional finance world and high frequency trading is such an important issue that a lot of like hedge funds actually have their ID infrastructure located like close to where the exchanges are so that they have the data that they need for their investments you know with an advantage of milliseconds above other competitors and that can make the difference at high frequency trading.
Japanese: 次に考えるのはレイテンシーの問題です。これもヘッジファンドの例と関連しています。DeFiエージェントは、市場の動向、異なる市場データ、価格、取引量などを分析し、最適な投資決定を行うために継続的に推論を実行する必要があります。伝統的な金融世界や高頻度取引におけるレイテンシー最適化の重要性は非常に高く、多くのヘッジファンドは取引所に近い場所にインフラを設置して、他の競合他社よりミリ秒単位で有利なデータアクセスを確保しています。これが高頻度取引における成否を分ける要因となり得ます。
English: So very similarly the optimization around latency time will be really important here and I think with our distributed compute network on IONEC we're able to actually offer, we're actually able to offer incredibly reduced times when it comes to latency for inference because we're able to provide the node closest to where the inferences happen right like whether the user is in Indonesia or Brazil or the US or Finland we have GPUs available everywhere so that time for someone to perform it for an agent to perform inference and get a response back after it's been computed is massively reduced compared to that same inference workload or every inference workload having to go to US East or wherever you know the massive data center from AWS is located to be able to give you an answer so I think that's an area where we're reducing a lot of friction too and I think lastly it's also just cost reduction right like as these AI agents look to optimize the returns that they offer for their DeFi users every user is actually is kind of want to look at which AI agent is able to have the least overhead as well right so that they can provide more of the returns back to the users providing them capital rather than having to spend it on compute so to take it back to the hedge fund analogy again it's hedge funds often charge the the two and twenty model for management correct like two percent management fee twenty percent on returns so in a very similar manner I think and you know different hedge funds compete on sort of like what date rate they they take on on the capital allocate but then so very similarly I think with AI agents they're able to reduce the cost of this compute usage by using a decentralized network like IO and producing that friction point they're able to sort of provide greater returns back to the users who allocate capital to them.
Japanese: 同様に、レイテンシー時間の最適化もここでは非常に重要です。IONEの分散型コンピュートネットワークでは、推論に関するレイテンシーを大幅に削減できます。推論が行われる場所に最も近いノードを提供できるためです。ユーザーがインドネシア、ブラジル、米国、フィンランドのどこにいても、GPUを利用可能な場所に配置しており、推論ワークロードがAWSのデータセンターに送られる場合と比較して、応答時間を大幅に短縮できます。この点でも多くの摩擦を解消しています。そして最後に、コスト削減についても触れておきます。AIエージェントがDeFiユーザーに提供する収益を最適化しようとする場合、ユーザーはエージェントがどれだけの間接費を抑えられるかを重視します。その結果、計算コストを削減し、ユーザーに還元できる収益を増やすことができます。この点でも、分散型ネットワークであるIOは大きな利点を提供しています。
English: and speaking of AI agents how does Injectives iAgent framework enable the seamless integration of AI agents could you give some details about what exactly is under the hood
Japanese: ところで、AIエージェントに関連して、InjectiveのiAgentフレームワークがAIエージェントのシームレスな統合をどのように可能にしているのか教えてください。その仕組みについて詳しく説明していただけますか？
English: yeah so I can't speak too much about the technical aspect of it but the TLDR is that(...) it kind of the framework itself provides seamless integration onto the Injectives blockchain and network as well so whenever you're using the framework you can easily build that and kind of add that agent let's say for example of trade on helix which is a club it's a perpetual and spot club that's built and embedded within the Injective network itself so it gives you that seamless compatibility that a lot of other supposed frameworks might not offer with their base pain
Japanese: そうですね、技術的な詳細についてはあまり話せませんが、要点をまとめると、このフレームワーク自体がInjectiveのブロックチェーンおよびネットワークへのシームレスな統合を提供します。このフレームワークを使用することで、簡単にエージェントを構築して統合できます。例えば、Helix（永続的および現物の取引クラブで、Injectiveネットワーク内に組み込まれている）での取引にエージェントを追加する場合、他のフレームワークが抱える基本的な痛点を解消し、シームレスな互換性を提供します。
English: and you know a follow-up question I have is Tasha was talking about in two levels friction but also what you had mentioned after experimentation comes scalability what have you noticed in terms of scalability so after the experimentation is done or and a builder has created a successful dApp where do you see the next stages of friction of work that needs to be done in our space
Japanese: 追って質問ですが、Tashaが言及した摩擦の2つのレベルについて、また実験の後にスケーラビリティが来るという話がありましたが、スケーラビリティに関してどのようなことに気づいていますか？実験が終わり、ビルダーが成功したdAppを作成した後、この分野で取り組むべき次の摩擦点はどこにあると考えていますか？
English: yeah so I think that look to be honest I think we're still very very early we're still in the experimentation part of this I mean it's like a month or it's only a couple months old that we've seen a lot of AI agents pop up I still do not believe we've had like any agent find product market fit and move on to the second part of the stage but I still think that we're still there everyone's still experimenting there hasn't been too many other problems outside of just experimenting getting deep and continuous access compute which is being solved by IONet but I haven't seen the second part of that friction really come out to date but we're I also believe we're very very early in this
Japanese: 正直なところ、私たちはまだ非常に初期段階にいると思います。まだ実験の段階であり、AIエージェントが登場し始めて1、2ヶ月ほどしか経っていません。まだどのエージェントも製品市場適合性を見つけ、それを超えて次の段階に進んだとは思っていません。しかし、まだそこにいます。皆がまだ実験を続けており、深く継続的なコンピュートアクセスの取得（これはIONetが解決している問題）以外にはそれほど多くの問題はありません。ただ、摩擦の次の段階が本格化しているとはまだ感じていません。しかし、この分野は非常に初期段階にあると信じています。
English: so so I think I'll take a different stab at it right I think(...) there are some fundamental problems which always exist in system and if you basically solve that fundamental problem your your rate of growth is exponential post that right so for example like beat any company right normally when they like when the company starts up and they start to become bigger and bigger the first thing they try to optimize on is actually on experimenting that how as a company I can change let's say my pipeline of development so that I can change more changes quicker and less time right how many developers at the same time can make changes so all the web two companies or web three companies basically optimize their tech stack and operations in a way that more experimentation could be done because that enables more development in parallel more changes to go in the outside world in parallel see what matters faster and what is basically client's liking and then build on top of it so these are like one of the fundamental being able to experiment experiment fast with minimal of minimal need of changes is one of the fundamental things if you can do it well you have exponential growth right
Japanese: それでは、別の視点から考えてみます。システムには常にいくつかの基本的な問題が存在しますが、その基本的な問題を解決すれば、その後の成長率は指数関数的に向上します。例えば、どの企業もそうですが、スタートアップから規模が大きくなるにつれて、最初に最適化を試みるのは「どうすればより短時間で変更を加えられるか」という点です。開発パイプラインを変更して、より迅速に変更を行い、複数の開発者が同時に変更を加えられるようにすることが目標です。Web2企業やWeb3企業は基本的に技術スタックや運用を最適化して、より多くの実験を行えるようにします。これにより、並行してより多くの開発を進め、何が重要で、クライアントが何を好むのかを迅速に見極め、それを基に構築することができます。迅速かつ最小限の変更で実験を行う能力は、成長を指数関数的に加速させる重要な要素です。
English: I think that's where this particular thing which the Injective team is doing I think as in like it's a very very fresh concept they have taken a first stab at it when they become matured more and more people know about it I think there'll be a huge adoption and people will see a value and it'll just keep like becoming more and more bigger and important in the industry you'll see right
Japanese: Injectiveチームが行っていることは非常に新しいコンセプトであり、彼らが最初に手を付けたものです。これが成熟していき、より多くの人々がその存在を知るようになれば、大規模な採用が進み、その価値が認識されることで、この分野でますます重要な存在になっていくと考えています。
English: and the other thing which I feel is a fundamental change also is that if you take a look at the centralized players right for example let's take about AWS itself like why many players use AWS right they can just go and take a bunch of bare metal machines from somewhere and build on top of it why do they go to AWS because they have the ML kits there which their teams can use they have a deployment pipelines on AWS which they can use right whenever they want to scale their CPU and GPU they have that software stack
Japanese: もう一つの基本的な変化として感じるのは、中央集権型プレイヤーの例です。例えば、AWSを考えてみましょう。なぜ多くのプレイヤーがAWSを使用するのでしょうか？単にどこかからベアメタルマシンを購入してそれを基盤に構築することもできますが、それでもAWSを選ぶ理由は、MLキットや展開パイプラインがあり、それをチームが利用できる点にあります。また、CPUやGPUをスケールさせたいときに必要なソフトウェアスタックも備わっています。
English: so what happens in in that case is that whenever startup comes and let's say if they only have eight ten people those ten developers or ML scientists(...) will predominantly focus on the problem statement not on the periphery problems of creating the infrastructure and all that we just focus on the business problem and those kind of companies succeed faster right
Japanese: このような場合、例えばスタートアップが8～10人の小規模チームで始める場合、その10人の開発者やML科学者たちは、主に問題の本質に集中することができます。インフラ構築やその周辺問題に時間を取られることなく、ビジネス問題に専念できるため、そのような企業はより早く成功するのです。
English: and that's where the ROI of each developers matter right now what has happened with this partnership is that injective is solving the AI agent kit side of the problems that the developers when they come they have less amount of friction when they build their businesses they don't have to they don't have to focus about other integration and all those things with the SDK which these guys have given right where I.O comes in picture is that I.O is giving that cheap compute that when a startup even have a slow low amount of funding they can still build on top of it right there still will be censorship averse right they still can scale at a at a very high level and they'll have access to the strongest of GPUs at the cheap price
Japanese: これが、開発者一人ひとりのROIが重要になる理由です。このパートナーシップによって、InjectiveはAIエージェントキット側の問題を解決しています。その結果、開発者がビジネスを構築する際の摩擦が少なくなり、統合やその他の問題に気を取られる必要がありません。ここでIOが登場します。IOは安価なコンピュートリソースを提供しており、スタートアップが少ない資金であってもその上に構築できます。また、検閲に耐性があり、非常に高いレベルでスケールできるだけでなく、最も強力なGPUを安価で利用できます。
English: right so this partnership will enable developers to build confidently iterate many times faster and can experiment at a very cheap price which is the fundamental values any like a startup would look out for I think that's the way I take a look at this partnership and what both these teams are bringing to the table.
Japanese: このように、このパートナーシップにより、開発者は自信を持って構築し、はるかに高速に反復し、非常に低価格で実験を行うことができるようになります。これはどのスタートアップでも求める基本的な価値です。このパートナーシップがもたらすものを私はそのように見ています。
English: Awesome awesome what's up guys I'm Zach from the I.O team really excited about being a part of the injective agent kit and I guess just being that this is our weekly spaces and we have a lot of people who don't have background on injective and what you guys have done and kind of the impact that you guys have had in the space do you mind just giving us some background for for listeners that are less familiar on the history of injective and how we've gotten to this moment and how you guys stand out in this crowded DeFi L1 landscape.
Japanese: 素晴らしいですね。こんにちは、IOチームのZachです。Injectiveエージェントキットの一部となることにとてもワクワクしています。そして、これは私たちの毎週のスペースですが、多くの人々がInjectiveについてのバックグラウンドや、これまでの実績、そしてこの分野での影響をあまり知らないと思います。Injectiveの歴史と、これまでの道のり、そしてこの競争の激しいDeFi L1環境でどのように際立っているのかについて簡単に教えていただけますか？
English: Yeah for sure so some background on injective launched a couple of years ago there's two good founders they launched initially as a simple Perpetex using the cause of wasm staff and then shortly after they realized that the biggest benefit of launching or using crypto rails to launch and expand is the compatibility aspect that crypto offers that crowdfi doesn't offer so then they launch injective which is an L1 and then helix is the perpetual and spot DEX embed embedded within injective itself so you can think of it as any DeFi protocol any protocol any token that is built on injective is seamlessly can be used on helix itself and this lends out to very cool um very cool and innovative ways to add more capital efficiency into finance
Japanese: もちろんです。Injectiveの背景について説明します。数年前に設立され、2人の優れた創業者によって立ち上げられました。当初はCosmos Wasmを使用したシンプルなPerpetexとしてスタートしましたが、すぐに仮想通貨のレールを利用して立ち上げと拡張を行う最大の利点が、従来の金融（CrowdFi）では提供されない互換性にあることに気付きました。そのため、InjectiveをL1として立ち上げ、HelixはInjectiveに組み込まれた永続的および現物のDEXとして提供されています。これにより、Injective上に構築された任意のDeFiプロトコル、任意のプロトコル、任意のトークンがHelix内でシームレスに利用できるようになります。これによって、非常にクールで革新的な方法で金融にさらなる資本効率を追加することが可能になります。
English: so I guess another thing is that injective is made for financial applications and it and one example this could be suppose if there was a DeFi agent that is building out a(...) F1 strategy where it allows anyone to deposit capital and it trades it beyond it trades it on behalf of its users one thing that you seem to see on injective is that suppose you put in ten dollars of your money into this DeFi agent then you could take that Lpster that you got from that ten dollars and deposit that on the helix to trade perps yourself so the end goal for injective is that it's built with composability in mind and with capital efficiency in mind is to bring not only crypto native DeFi and prop that up but also work on bringing traditional capital whether the rural assets or other stuff on chain um and at least grow finance in that sense
Japanese: Injectiveが金融アプリケーション向けに作られているもう一つの理由は、例えばF1戦略を構築しているDeFiエージェントがあったとします。このエージェントでは、誰でも資本を預けることができ、エージェントがユーザーの代わりに取引を行います。Injectiveで見られる一例として、あなたがそのDeFiエージェントに10ドルを預けた場合、その10ドルで得たLPトークンをHelixに預けて、自分自身でパーペチュアル取引を行うこともできます。Injectiveの最終的な目標は、コンポーザビリティ（相互運用性）と資本効率を念頭に置いて構築されており、暗号ネイティブのDeFiを支えるだけでなく、伝統的な資本（農村資産やその他のもの）をオンチェーンに持ち込み、金融を成長させることです。
English: Cool um and then I guess like what are some exciting uh real world use cases uh that you've seen I mean I personally uh like to call it AI-Fi uh but uh DeFi uh is is the is the popular colloquial term um yeah so what are some exciting uh projects that you guys have gotten some inspiration from um on this SDK and I guess like yeah what do you find most exciting going on in this AI-Fi space at this moment for injection yeah definitely
Japanese: なるほど。それでは、現実世界で見られるエキサイティングなユースケースについて教えてください。個人的にはこれを「AI-Fi」と呼びたいですが、一般的には「DeFi」がよく使われる用語ですよね。このSDKで得たインスピレーションを受けたエキサイティングなプロジェクトについて、またInjectiveにとって現在このAI-Fi分野で最もエキサイティングだと感じることについて教えてください。
English: so there's the example I just gave um about the so I think one thing it's also like it's like DeFi is like DFAI or A5 but yeah um the naming is really it's hard to say but the example I just gave is there's a developer right now using the Injective SDK to build out a authentic finance um hedge fund where people can kind of give this agent agent money and then the agent will trade on behalf of will trade on behalf of them on Helix itself and then kind of distribute more distributed earnings on a monthly basis but I think that's very cool where instead of getting suppose going back to the hedge fund analogy you see a prior instead of giving a human your capital to invest on behalf of you now I think shortly will certainly soon we'll be seeing people getting agents their money to herd yield and profit you know on top of
Japanese: 先ほどの例についてですが、「DeFi」や「DFAI（AI-Fi）」という名前はまだ確立されていないものの、現時点でInjective SDKを使用して本格的な金融ヘッジファンドを構築している開発者がいます。このエージェントでは、人々が資金を預けると、その資金をHelixで取引し、月単位で分配利益を配布する仕組みを提供しています。これが非常に面白いと思うのは、従来のヘッジファンドのアナロジーに戻ると、人間に資本を預けて投資してもらう代わりに、今後はエージェントに資本を預けて利回りや利益を得る場面が増えることです。
English: and I think one long-term difference this is going to make into market structure is that it's going to be much more capital it's going to be much more efficient where there isn't going to be a lot of different price discrepancies between different sexes and different sexes for Bitcoin let's say because of many more players whether they be agents kind of neutralizing the price discrepancies
Japanese: そして、これが市場構造に与える長期的な影響の一つは、資本の効率性が大幅に向上することです。例えば、ビットコインについて言えば、エージェントが価格差を中和することで、異なる取引所間や異なる市場間での価格差が大幅に減少するでしょう。
English: cool um yeah so rip to the uh to the people who are trading on those on those opportunities um and then I guess uh what are some current limitations uh within either DeFi or AI whether that be um I guess interchain operability um I guess like what what limitations do you currently see uh in the way for uh for this space specifically um and hurdles um that we'll need to get through in order to have like a fully uh functioning AI-Fi environment um yeah
Japanese: なるほど、それでは、DeFiやAIの現状の制限について教えてください。例えば、チェーン間の相互運用性や、具体的にこの分野で直面している課題について、また完全に機能するAI-Fi環境を実現するために克服すべきハードルについて、どのようなものがあると考えていますか？
English: yeah I can take a start right here I think like go ahead the indicating I think you were about to say something
Japanese: はい、ここでお答えします。どうぞ、お話される予定だったことがあれば続けてください。
English: oh no go ahead sorry go ahead go ahead
Japanese: いえ、どうぞ。申し訳ありません、続けてください。
English: so I think the first thing is uh we need to have a enhanced usability and accessibility uh in the first place so I still believe(...) there are like several key developments which need to occur which is across your technology your user experience or your regulatory side and overall the other ecosystem building side as well for this whole DeFi space to grow right
Japanese: まず最初に必要なのは、ユーザビリティとアクセシビリティの向上だと思います。現時点では、このDeFi分野全体が成長するためには、技術面、ユーザー体験、規制面、さらにはエコシステム全体の構築において、いくつかの重要な進展が必要だと考えています。
English: and it starts from the first one it is like your enhanced usability and accessibility right I believe the AI agents must become more intuitive and user-friendly uh so that it can becomes more seamless to use both fucked up to natives and the newcomers like I genuinely haven't seen till now uh basically an agent where a complex DeFi uh agent uh uses NLP and you can see simple things like where can I earn the highest yield on my eat and the AI agent like could just handle this research do the transaction staking automatically and end-to-end do everything right
Japanese: 最初に挙げられるのは、ユーザビリティとアクセシビリティの向上です。AIエージェントはより直感的で使いやすいものにならなければならないと考えています。これにより、暗号ネイティブな人々と新規参入者の両方がシームレスに利用できるようになります。現時点では、NLPを使用して「私のETHで最高の利回りを得られるのはどこか」といったシンプルな質問を処理し、自動的にリサーチを行い、トランザクションを実行し、ステーキングを完了し、エンドツーエンドで対応できる複雑なDeFiエージェントはまだ見られません。
English: I'm surprised nobody has done a multilingual support uh till now right like basically like any good AI agent should provide services to a global audience right by offering support in multiple languages we don't have that
Japanese: 現時点で誰も多言語対応をしていないのには驚いています。本来、優れたAIエージェントは複数の言語でサポートを提供し、グローバルなオーディエンスにサービスを提供するべきですが、それがまだ実現していません。
English: One of the things I think uh which I feel in many areas in crypto including the deep end or DeFi is that we I don't know why we compete a lot between ourselves and we don't work together as a vertical industry to become better right like I've seen that there's no unified standard API right the collaborations like Injective and I have done there not many collaborations like this which basically creates synergies and the(...) better value in the lesser amount of work as protocols and giving more services and value services to the customer I don't see that
Japanese: 私が感じるのは、暗号業界全体、特にDeFiなどの分野で、なぜかお互いに競争することが多く、垂直的な業界として連携して改善する動きが少ないという点です。統一された標準APIが存在しないのもその一例です。Injectiveと私たちのようなコラボレーションがもっとあれば、少ない労力でより良いシナジーや価値を提供でき、プロトコルとして顧客により多くのサービスと価値を提供できるのに、それがまだ見られない状況です。
English: on the security and press side there have not been many innovation being done right like many of the projects still on the DeFi side I don't have the audible code right transparency isn't there they they're their models which they're deploying is still not there visible to everyone right
Japanese: セキュリティや透明性の面で、DeFiのプロジェクトにはまだあまりイノベーションが見られません。多くのプロジェクトは、コードが完全にオープンではなく、透明性が欠けています。また、デプロイされているモデルもすべての人に見える状態にはなっていません。
English: similarly there are more models itself which should be created where like you just throw those models on any smart contract which comes and it audits it and transparently show the results to everyone in the audience and then if it's good then deploy it right so that everyone knows what's happening what were the different(...) issues which came in this code which they were with the respective protocol was thinking of deploying and if they have done the fixes what the second version of the review was and so and so forth
Japanese: 同様に、より多くのモデルが必要です。例えば、スマートコントラクトにモデルを適用し、それを監査して結果を透明に公開し、問題がない場合にのみデプロイする仕組みが必要です。これにより、何が行われているのか、コードで発生した問題、そのプロトコルがどのように対応したのか、修正後のレビューの結果など、すべてを誰でも確認できるようになります。
English: so people actually have confidence in the industry overall that that what the teams are doing is is real is audible they can read understand it and even if they was wrong how it was fixed and how soon it was fixed and what was the final outcome
Japanese: こうすることで、業界全体に対する信頼が高まり、チームが行っていることが本物であること、誰もがその内容を読んで理解できること、そして問題があった場合、それがどのように修正され、どれくらい早く修正され、最終的な結果がどうだったのかを確認できるようになります。
English: there's nothing around right the interoperability across the protocols and change is still very bad right there's not much innovation being done on the AI agent side where it's seamless to work like across multiple DeFi protocols blockchains as a holistic solution
Japanese: 現時点では、そのような取り組みはほとんど見られません。プロトコル間やチェーン間の相互運用性は依然として非常に悪く、AIエージェントが複数のDeFiプロトコルやブロックチェーンをまたいでシームレスに機能するような包括的なソリューションにおいて、あまりイノベーションが進んでいません。
English: I think this is one area where as industry we can become better right there was always a there's a part of cost I think that with a lot of deepens has been solved there's some issue the real-time data also I think these are few friction points which exists in the industry technically which could be taken a stab at and we'll be better as industry
Japanese: これは業界全体が改善できる分野だと思います。コストに関しては多くのDePINプロジェクトで解決されつつありますが、リアルタイムデータの課題など、技術的な摩擦点がいくつか存在しています。これらの問題に取り組むことで、業界全体がより良くなるでしょう。
English: and then obviously there's there's the whole education piece also that still there's a lot of friction from web to giants and these great builders who know how to create these(...) distributed system they're still not coming to to the web tree side because we as industry are not educating them enough about the benefits of the decentralization the tokenomics and what they can take advantage of this whole blockchain ecosystem and crypto ecosystem
Japanese: また、教育の側面も重要です。分散型システムを構築する技術を持った優れたビルダーやWeb2の大企業が、Web3側に進出してこない理由の一つは、分散化やトークンエコノミクス、そしてブロックチェーンや暗号エコシステム全体の利点について、業界全体として十分に教育を行っていないからです。
English: I think that education is is also not being spread out and obviously the last piece which is a problem is the regulatory clarity right the regulatory uncertainty around the DeFi and the AI could really hinder the adoption right a clear framework will create confidence in the ecosystem
Japanese: 教育が十分に普及していない点も課題だと思います。そして、最後の課題としては規制の明確さが挙げられます。DeFiやAIに関する規制の不透明性は、採用の妨げとなり得ます。明確なフレームワークがエコシステムへの信頼を生むでしょう。
English: and I think the new government will just come in everyone's really hopeful that like the best days are coming pretty soon right so I think that that's all which I had if this these are things of just all like we'll be much better than the strict
Japanese: そして、新しい政府が登場し、みんながこれから良い時代がすぐに来ることを期待しています。以上が私の考えです。これらの課題に取り組むことで、業界はより良い方向に進むと思います。
English: Cool awesome and then I guess you guys that injective if you want to answer the question about other hurdles that you guys see in the way and then I guess yeah another side question
Japanese: 素晴らしいですね。それでは、Injectiveの皆さんに、直面している他のハードルについて教えていただけますか？また、関連する別の質問があります。
English: for for injective is us at Ironet where we're happy and excited to be to be also partnering with the likes of creator bids the rebrow and AI 16z in the space and then I guess outside of things going on at injective what do you guys find most exciting within the crypto AI space or the AI 5 space that's kind of inspiring to you guys
Japanese: Injectiveに関連してですが、私たちIronetは、クリエイタービッド、Rebrow、AI16zなどとパートナーシップを結んでいることを嬉しく思います。それでは、Injective以外で暗号AI分野やAI-Fi分野で最もエキサイティングでインスピレーションを与えるものは何ですか？
English: yeah okay so that's a lot let me break it down for the first question I think it's best to view like what was like the it's best to view the current challenges as going back to the analogy of who's the user so we can split up the users as crypto-dgens crypto-natives with that are more risk-averse like funds or whales and the last one being will who are not creative so we'll split split up two ways
Japanese: そうですね。質問がたくさんあるので分解してみます。まず最初の質問については、課題を「ユーザーとは誰か」というアナロジーに戻して考えるのが良いと思います。ユーザーを、暗号Degen、リスク回避型の暗号ネイティブ（ファンドやホエールのような人々）、そして最後に非暗号ネイティブの人々に分けることができます。それをさらに2つの方法で分けてみましょう。
English: the first one I think thus far we've seen the most activity for these crypto agents or dia dia 5 among the main coin traders last dgens because they tend to be more risk-averse and a lot of the tooling that's being built out or a lot of the tools that have a lot of usage thus far have been crackers of um like deep trenches is this like what's the probability that this tokens are run or not so I think that has found really strong product market fit
Japanese: 最初に、これまで見られた暗号エージェントやDeFiの活動の多くは、主にミームコイントレーダーやDegenに集中しています。彼らはリスクを回避する傾向があるため、これまで多く使われてきたツールは、トークンが「ランする可能性」などを計算するクラックツールが中心でした。この分野では非常に強いプロダクト・マーケット・フィットを見つけたと言えます。
English: but going to the crypto-native whales last funds I think a big problem is you guys mentioned that is that security is still unclear and it's difficult to give a AI agent let's say a large sum of money if there hasn't been any will in the effect that it is safe and he will not lose my money or there won't be a smart contract risk where I kind of lose all my money
Japanese: しかし、暗号ネイティブのホエールやファンドに話を移すと、大きな問題はセキュリティがまだ不明瞭であることです。安全であるという保証がない場合、AIエージェントに大金を預けることは難しいです。資金を失う可能性がない、またはスマートコントラクトのリスクで資金を全て失うことがないという信頼性が必要です。
English: kind of going back to the early d5 days where d5 days where there's a lot of different borrow lending protocols on ethereum but over time there only a lot of the large crypto players feel comfortable in allocating capital onto maker dower obage because they survived months and years worth of um usage and while whereas a lot of the other borrow lending particles suddenly died and got robbed a couple of times and the smart contracts were not safe
Japanese: 初期のDeFi時代を振り返ると、Ethereum上には多くの借入・貸付プロトコルが存在していましたが、長期間にわたり利用されて信頼性を確立したMakerDAOやAaveに大規模な資本が配分されるようになりました。一方で、他の多くの借入・貸付プロトコルは突然消滅したり、数回ハッキングされたりしてスマートコントラクトが安全ではありませんでした。
English: so I think that's a similar thing that's going to happen with d5 where it's going to take some time for the products to prove that these market products themselves are safe and that they are
Japanese: これはDeFiでも同様のことが起こると思います。製品が安全であり、信頼できることを証明するには時間がかかるでしょう。
English: the returns that they promise or whatever the product might be is that's sustainable over time and the last part regarding people that are not native I think that comes down to again you guys mentioned this as well is just accessibility where it's still not at the level that if you want to create a snapchat account or a instagram account or tech account it's much easier to do that than create a account where like a d5 agent that manages your money
Japanese: 製品が約束するリターンやその他の機能が時間をかけて持続可能であることを証明する必要があります。そして、最後に非ネイティブな人々に関してですが、ここでもアクセスのしやすさが問題になります。SnapchatやInstagramのアカウントを作成するのは簡単ですが、お金を管理するDeFiエージェントのアカウントを作成するのはまだそれほど簡単ではありません。
English: so I think it's the hurdle that's going to pass over time but that's probably the biggest friction onboarding new users into these products now
Japanese: これは時間が経てば克服される課題だと思いますが、現時点では新規ユーザーをこれらの製品に導入する際の最大の摩擦点になっています。
English: my second question well first of all I just like I just like to acknowledge that yeah I think like that's the the total end dream that these consumer um the the regular consumer can come on to into the crypto world through one of these bots and not have to worry about bridging to Solana bridging to ETH just that it all works seamlessly and they just see the fee alongside it
Japanese: 次の質問に入る前に、まず認めたいのは、一般消費者がこれらのボットを通じて仮想通貨の世界に参加でき、SolanaやEthereumへのブリッジングを気にすることなく、すべてがシームレスに機能し、手数料も一緒に表示されるというのは、まさに最終的な理想だと思います。
English: and then the second question was just outside of things going on at Injective what do you guys find most exciting and inspiring right now at the moment in the crypto AI space or the AI 5 space yeah yeah yeah
Japanese: そして2つ目の質問ですが、Injectiveに関連しない話題で、現在、暗号AI分野やAI-Fi分野で最もエキサイティングでインスピレーションを受けるものは何だと思いますか？
English: definitely so I think the thing that excites me and excites us the most is the amount of net new talent and net new developers coming into crypto for the first time so we haven't really seen this kind of bloom in talent in crypto for a year or so it's been a long time but the AI crypto AI or DeFi has really encouraged a lot of my web2 friends kind of work in traditional tech or work in the AI not the world kind of give a stab at crypto kind of play around um nights and weekends kind of see what they can build in crypto
Japanese: 確かに、私や私たちが最もワクワクしているのは、新しいタレントや開発者が仮想通貨の分野に初めて参入していることです。このようなタレントの急増は、ここ1年ほどでは見られなかった現象です。AIや暗号AI、DeFiは、多くのWeb2の友人たち、つまり従来の技術分野やAIの世界で働いていた人々に、仮想通貨に挑戦し、夜や週末にいろいろ試して、何が作れるかを探る機会を与えています。
English: so I think that's the most exciting part where we're seeing a lot of new talent come into crypto industry because of crypto AI and DeFi
Japanese: これが最もエキサイティングな部分だと思います。暗号AIやDeFiのおかげで、多くの新しいタレントが仮想通貨業界に参入しているのです。
English: Let me pass the question of new talent to the IOT team you all have talked about you know and mentioned on this call about friction for new developers people coming into web3 and the issues facing them do you see with DefAI any other unique challenges or do you think that it follows the traditional developer breaking into web3 challenges?
Japanese: 新しいタレントについての質問をIOTチームにお聞きします。この通話で、新しい開発者がWeb3に参入する際の摩擦や課題について触れていましたが、DefAIにおいて特有の課題があると感じますか？それとも、Web3に参入する際の従来の課題に従うものだと思いますか？
English: I could take a stab at it. I think
Japanese: 私が答えましょう。私の考えでは、
English: you know in terms of new talent entering the market I think sort of like this AI agent boom has really been kind of a boon for the overall sort of crypto space right and I think it sort of like really inspired a lot of net new developers with sort of broader skill sets to enter the market kind of explore sort of the various like toolkits and SDKs and you know sort of like programming languages that are available in the blockchain space and really try to come at it with a fresh lens and build on top of it from this lens of AI agents
Japanese: 新しいタレントが市場に参入するという点では、このAIエージェントのブームが仮想通貨業界全体にとって非常に大きな恩恵となっていると思います。幅広いスキルを持つ多くの新しい開発者が市場に参入し、ブロックチェーン分野で利用可能なさまざまなツールキットやSDK、プログラミング言語を探求し、新たな視点でそれらを活用して構築を試みるようになっています。
English: right so I think as as Abhi mentioned I think it's been like a really exciting time of seeing all the people pour into crypto again you know after sort of like the bear market of 2022 when a lot of people left the space I think people are kind of coming back and you know in sort of a larger you know almost an order of attitude more than what was what was before because not only are we attracting sort of you know classic like web3 developers but a lot of folks who have never dutch web3 have never thought about crypto before have been you know sort of primarily in the AI ML space or experimenting with AI agents
Japanese: Abhiが言ったように、2022年の弱気市場で多くの人がこの分野を去った後、再び多くの人々が仮想通貨に戻ってきており、以前よりもさらに多くの人々が参入しています。従来のWeb3開発者だけでなく、これまでWeb3に関わったことのない、あるいは仮想通貨について考えたことのないAIや機械学習分野の人々やAIエージェントを実験している人々も多く参入しています。
English: I think they've started to discover you know a lot of the benefits of being able to build AI agents on crypto rails like the benefits of like permissionless finance like access to sort of like decentralized compute data oracles things like that like a lot of tools I think they're trying to sort of have this aha moment of realizing that a lot of what we've been building in crypto and web3 for several years are a natural fit with what AI agents actually need and how they'll likely operate in the future
Japanese: 彼らは、仮想通貨のレール上でAIエージェントを構築する利点、たとえば許可不要な金融や分散型コンピュート、データオラクルへのアクセスなど、多くの利点を発見し始めています。これまで仮想通貨やWeb3で構築してきた多くのものが、AIエージェントが実際に必要とするものや将来の運用方法と自然に適合していることに気付き、いわゆる「aha」モーメントを迎えていると思います。
English: so I think this marriage of AI and crypto and deep DefAI I think has sort of like really created a new paradigm of development and sort of really opened up kind of the design space right
Japanese: このAIと仮想通貨、そしてDefAIの融合が、新しい開発のパラダイムを生み出し、設計スペースを大きく広げたと思います。
English: and Zach you mentioned a few of the folks we've we've partnered with you know so Shaw, the i16z, like Jeffy and Augustin and Zarebro are you know great examples of that like there were folks who were primarily on the AI agent side are sort of you know fully in the AI development side before came into crypto did a lot of interesting experimentation and have now built sort of like billion dollar projects right and in a short amount of time and have also sort of like enabled a lot of other developers to build on top of sort of the foundational open source stuff that they've done
Japanese: Zach、あなたが言及したように、Shaw、ai16z、Jeffy、Augustin、Zarebroなどのパートナーはその良い例です。これらの人々は主にAIエージェント側で活動していましたが、仮想通貨の分野に入り、多くの興味深い実験を行い、短期間で数十億ドル規模のプロジェクトを構築しました。また、彼らが行った基盤的なオープンソースの取り組みの上に、多くの開発者が構築できるようになりました。
English: so I think it's it's a really exciting time for you know the future of AI and crypto and for talent coming into this space
Japanese: ですから、AIと仮想通貨の未来、そしてこの分野に参入してくるタレントにとって、本当にエキサイティングな時期だと思います。
English: Let me ask everyone on the panel hey what do you say to people who say great this is just another bubble we've heard all of this before about onboarding new people that things are going to become easier that we're getting interest from outside the space that's happening what makes this sound different what are you all seeing in use cases technology growth(...) from conversations you're having and just generally looking at the space.
Japanese: パネルの皆さんにお聞きします。「これはただのバブルだ」と言う人々に対して、皆さんはどう答えますか？新しい人々が参加しやすくなり、業界外からの関心が高まっているという話はこれまでにも聞いてきましたが、今回は何が違うのでしょうか？皆さんが見ているユースケースや技術の成長、会話の中で感じていること、そして全体的な業界の様子について教えてください。
English: Yeah I think it can be both right it can be both a bubble that's kind of blown up over the next couple of months or years but also have really strong fundamentals where the team's building out some of the projects will sustain over the next decades to come so I kind of agree with Bill takes that there are fundamental use cases for example right now a lot of blowing fruit spin being solved in the trenches and what that how kind of new tech usually evolves is that kind of it starts as the first adopters tend to be the most risk on people and then over a longer time rising it gets onto more risk-averse people kind of start using similar products using the underlying technology
Japanese: そうですね、どちらの側面もあると思います。次の数ヶ月から数年でバブルとして膨らむ可能性もありますが、同時に、チームが構築しているプロジェクトのいくつかが次の数十年にわたって持続可能な強固な基盤を持つことも確かです。基本的なユースケースが存在することには同意します。例えば、現在、多くの簡単に解決できる課題が取り組まれており、新しい技術が進化するプロセスとしては、最初の採用者が最もリスクを取る傾向があり、その後、よりリスク回避型の人々が基盤技術を使用して類似の製品を利用し始めるという流れです。
English: so I think the underlying technology of authentic finance makes a ton of sense given how it can increase composability in the financial market and increase capital efficiency
Japanese: ですから、金融市場でのコンポーザビリティ（相互運用性）を高め、資本効率を向上させる方法を考えると、本格的な金融の基盤技術は非常に理にかなっていると思います。
English: and finance is kind of built on top of capital efficiency so fundamentally it's sound but that being said I can I also see it kind of being a bubble over the next couple of months to years giving you out of interest so I think both are true.
Japanese: そして、金融は資本効率の上に構築されているので、基本的には健全です。ただし、それでも次の数ヶ月から数年の間に興味の対象としてバブルになる可能性もあると思います。どちらの側面も正しいと感じます。
English: I think for me personally it has not been hard in a previous experience and now in IO also to basically convince people to come to the crypto side because if you break it down right I don't think in crypto companies there is any more risk compared to a web2 startup also right like it's the same risk in in a crypto startup as you have in a web2 startup right
Japanese: 個人的な経験として、これまでの経験や現在のIOでの活動において、人々を仮想通貨の分野に引き込むのはそれほど難しくありませんでした。なぜなら、よく考えてみると、仮想通貨企業にはWeb2スタートアップと比べてリスクがそれほど多いわけではないからです。仮想通貨スタートアップとWeb2スタートアップのリスクは同程度だと思います。
English: the percentage is more or so will look similar right in terms of the problem statement which you solve I think like crypto has more interesting problem to solve compared a lot of web2 giants to be honest right and practically speaking the velocity at which the crypto companies iterate and build is far faster compared to any web2 company
Japanese: 解決する問題の性質に関しては、仮想通貨の方がWeb2の大手企業と比べて、より興味深い問題を解決する必要があると思います。実際のところ、仮想通貨企業が反復して構築するスピードは、どのWeb2企業よりもはるかに速いです。
English: right and at the end engineers are builders and they like to work on interesting stuff ship it in front of people get their feedback and build on top of it right that's what excites them the most right
Japanese: 結局のところ、エンジニアはビルダーであり、興味深いものを作り、人々の前にリリースし、フィードバックを得て、それを基にさらに構築していくことが最も彼らをワクワクさせます。
English: so that is practically there in the crypto company like imagine like in IO personally we're building cloud platform which took your Amazon years to build right we're literally solving the same problem which Amazon did so for a builder who has seen those products and they say oh we'll get the same product we built here in the next one and a half year there's so many interesting problems solved they're super excited right
Japanese: 仮想通貨企業では、それが現実のものとなっています。例えば、IOでは、Amazonが何年もかけて構築したクラウドプラットフォームを構築しています。つまり、Amazonが解決したのと同じ問題を解決しているのです。そのような製品を見たビルダーは、「次の1年半でここで同じ製品を作ることができる」と言って、多くの興味深い問題を解決し、大いに興奮しています。
English: and practically speaking I think the founders and the senior guys were there in crypto they'll practically tell you the salaries in the crypto companies are far better in the web2 site right so it's it's you don't have to compete in that side also because we do pay well in our industry anyhow
Japanese: また、仮想通貨業界の創業者やシニア層の人々は、仮想通貨企業の給与がWeb2の企業よりもはるかに良いと教えてくれるでしょう。このため、給与面で競争する必要もありません。私たちの業界では、もともと十分な報酬が提供されています。
English: so attracting people in terms of the problem statement and this thing is comparatively easy like at least for me it has been historically easy the only thing I think as as tech people we have to do is that when new people come in in the industry they get overwhelmed right
Japanese: 問題の性質やこの分野への魅力という点では、人々を引き付けるのは比較的簡単です。少なくとも私にとっては、歴史的に簡単でした。ただし、技術者として私たちが行うべき唯一のことは、新しい人々がこの業界に入るときに、彼らが圧倒されないようにすることです。
English: for a new person who's coming in the whole industry itself they're normally risk on they can come and join any industry web2 or web3 right the problem comes in if somebody is a 12 year experience guy he's basically one of the senior most person in a web2 company and for him to leave a web2 company a comfortable job where he's proven his value in everything to come in a web3 site where everything is new for them that's overwhelming right
Japanese: 業界全体に新しく入ってくる人は通常、リスクを取る準備ができており、Web2でもWeb3でもどちらの業界にも参加できます。しかし、12年の経験を持ち、Web2企業の最もシニアな人物の一人であるような人が、居心地の良い仕事や自分の価値を証明した環境を離れて、すべてが新しいWeb3分野に移るのは、非常に大きな負担です。
English: and over there for those guys for make them an easy transition as leaders we have to break it down that we are basically solving a problem which we'll cater to a web2 side also and web3 side also and here is a plan for you that in the coming six months this is the problem statement you will solve which will predominantly web2 focus distributed systems and slowly and steadily we will train you on web3
Japanese: そのような人々にとって、スムーズな移行を実現するためには、リーダーとして、私たちがWeb2とWeb3の両方に対応する問題を解決していることを説明し、次の6ヶ月間で解決する問題を示す必要があります。この問題は主にWeb2に焦点を当てた分散システムに関連しており、徐々にWeb3についても訓練していきます。
English: and as long as you show that transition plan to them those senior guys are comfortable but if you just throw all these learnings to them in one shot which is blockchain distributed systems decentralized system and all these other tokenomics then it's becoming overwhelming and those kind of people don't transition well are afraid to make that jump from web2 to web3 I think that's where this transition and explaining things helps a lot
Japanese: そして、この移行計画を彼らに示す限り、そういったシニアな人々も安心します。しかし、ブロックチェーン、分散システム、非中央集権システム、その他のトークノミクスなどのすべてを一度に投げ込むと、圧倒されてしまい、そのような人々はスムーズに移行できず、Web2からWeb3への移行を恐れてしまいます。この移行計画を説明することが非常に役立つと考えています。
English: just to add a few things there too um actually i'm going to go back to these point on bubbles um that he made earlier too and kind of how that how that plays out right so
Japanese: ここにいくつか追加すると、実際に彼が先ほど述べたバブルについてのポイントに戻り、その展開について話したいと思います。
English: kind of as he mentioned um there's you know a couple different ways to look at it and I think often people kind of like look at bubbles as sort of like this evil thing um you know in financial markets but i kind of really think of it as sort of being a natural sort of part of technological innovation and progress right
Japanese: 彼が言ったように、バブルにはいくつか異なる見方があります。多くの人が金融市場でのバブルを悪いものと見なす傾向がありますが、私はこれを技術革新や進歩の自然な一部と考えています。
English: and that's kind of really what we're doing here is with dfa and ai agents we're sort of at the bleeding edge of technological progress and what is really possible to do in finance and crypto and ai
Japanese: 私たちがここで行っていることは、DeFiやAIエージェントを使って、金融、暗号通貨、AIにおける技術進歩の最前線に立っているということです。
English: and you know i think historically with bubbles like whether it's related to you know the internet or railroads or crypto um i think what it really sort of you know tangentially is related to is this idea of experimentation right which we talked about earlier as well
Japanese: 歴史的に見ても、バブルはインターネットや鉄道、仮想通貨などに関連しており、その本質的な要素は「実験」という考え方にあります。これは先ほどもお話ししました。
English: I think bubbles typically sort of are a breeding ground for experimentation and it sort of like attracts sort of like this initial capital injection really attracts a lot of talent into the space sort of breeds a lot of new ideas gets people excited about trying out a lot of different things trying to build and solve for a variety of different problems using this sort of like new technology
Japanese: バブルは通常、実験の温床となり、初期の資本投入を引き付け、多くのタレントをその分野に呼び込みます。そして、多くの新しいアイデアを生み出し、人々がさまざまなことを試し、新しい技術を使用してさまざまな問題を解決しようとするきっかけを提供します。
English: and you know in some ways a lot of those experiments don't end up working out and you know whether it's a company or a project or something like that it can end up going to zero because maybe the technology just wasn't quite there yet or like the adoption curve wasn't quite there yet
Japanese: もちろん、多くの実験がうまくいかず、企業やプロジェクトなどがゼロになってしまうこともあります。それは、技術がまだ成熟していなかったり、採用の曲線が追いついていなかったりする場合に起こります。
English: but what it really does to the experimentation is it shines a light on gaps and things that need to be solved for the overall movement to be successful
Japanese: しかし、実験がもたらすものは、全体的な成功のために解決すべきギャップや課題に光を当てることです。
English: like yeah as I mentioned whether it's internet or railroads or crypto or AI agents(...) and what that kind of really does is sort of it helps the next generation of builders really sort of tighten focus and solve for those gaps and kind of identify what kind of things need to happen to be to enable that business to be sustainable in the long term right
Japanese: インターネットや鉄道、暗号通貨、AIエージェントなど、これらが次世代のビルダーに焦点を絞らせ、ギャップを解決し、ビジネスを長期的に持続可能にするために必要なことを特定する助けとなります。
English: like to take sort of like the bubble example one of the you know biggest stories of that time or companies at that time was pets.com right like getting pet food over the internet which was sort of often used as this example of being a ridiculous business that you know described the hubris of the dot-com bubble
Japanese: バブルの例を挙げると、当時の最大の話題の一つがPets.comでした。これは、ペットフードをインターネットで販売するというビジネスで、ドットコムバブルの典型的な過剰さを象徴する例としてよく挙げられました。
English: but now is you know a really major sustainable profitable business in the form of chewy.com
Japanese: しかし、現在ではChewy.comという形で、非常に大きな持続可能で収益性の高いビジネスとなっています。
English: um and you know and I think you can kind of make that argument about a variety of different things like whether it's like airbnb or uber or instacart a lot of those ideas were tried out in like back then in 99 didn't work out but a lot of the builders stuck around on the internet stuck around building apps figured out what gaps needed to be solved and then now they're you know multi-billion dollar businesses
Japanese: これは、AirbnbやUber、Instacartなどにも同じことが言えると思います。これらのアイデアは1999年当時に試みられたものの、当時はうまくいきませんでした。しかし、多くのビルダーがインターネットに留まり、アプリを構築し、解決すべきギャップを特定し、現在では数十億ドル規模のビジネスになっています。
English: so I think we'll go through a very similar cycle where you know we'll have the amazon's and google's who'll we're being built now in you know an equivalent space and we'll survive whatever sort of bear market comes in the future
Japanese: 私たちは非常に似たサイクルを経験すると思います。現在、この分野で構築されているAmazonやGoogleのような企業が、将来の弱気市場を乗り越えて生き残るでしょう。
English: but I think the experimentation that will come from this will really drive the overall space forward and ensure that the movement is sustainable for the future.
Japanese: しかし、この実験から生まれるものは、全体の分野を大きく前進させ、このムーブメントが未来に向けて持続可能であることを保証するでしょう。
English: So as we close out the first question i have of two ones left is(...) what's in the pipeline that people should be paying attention to i'll start with Injective.
Japanese: 最後の質問に移る前に、2つ残った質問のうち最初のものをお聞きします。注目すべきInjectiveのパイプラインにはどのようなものがありますか？
English: Yeah so um on our end is this particular to DeFi or Injective in general as Injective in general? Okay cool yeah so on the pipeline i think one thing that internally the team has been grinding on the past couple of months is our EVM mainnet launch so that is probably the biggest thing that we will be soon going live with over the coming months and that again that's again decreases the friction for developers to launch OnInjective because we realize that there's a fixed amount of developers comfortable in writing on custom awesome code so with this development and with the compatibility between our EVM mainnet and our mainnet i think that's gonna
Japanese: そうですね。これはDeFiに特化したものですか？それともInjective全般に関するものですか？了解しました。パイプラインについてですが、チームがここ数ヶ月間取り組んできた最大のプロジェクトは、EVMメインネットの立ち上げです。これはおそらく、今後数ヶ月以内に公開される予定で、開発者がInjective上での開発を開始する際の摩擦を減少させます。カスタムWasmコードの記述に慣れている開発者の数には限りがあるため、この開発により、EVMメインネットとメインネット間の互換性が向上します。
English: enable a lot more devs and a lot more action on Injective mainnet.
Japanese: これにより、より多くの開発者が参加し、Injectiveメインネット上での活動が大幅に増加するでしょう。
English: And I know there's a bunch of things coming up what in particular is something that people should be paying attention to?
Japanese: これからたくさんのことが予定されていますが、特に人々が注目すべきものは何ですか？
English: I think it basically attaches to the same question which you asked earlier so when we actually(...) plan for any quarter we basically do the research and like you asked like whether it's a bubble we don't see it's a bubble our research said that the whole AI agents it's not just a web3 phenomena it's basically web2 and traditional enterprises are equally invested and which shows basically isn't a bubble but a universal shift towards smarter more efficient systems across the industries like InShot
Japanese: これは先ほどの質問とほぼ同じ内容に関連します。私たちが四半期ごとの計画を立てる際、研究を行い、例えばバブルかどうかと問われた場合、私たちはバブルとは考えていません。研究によれば、AIエージェント全体はWeb3だけの現象ではなく、Web2や従来の企業も同様に投資しており、これはバブルではなく、業界全体でよりスマートで効率的なシステムへの普遍的な移行を示しています。
English: and what we are doing is we are creating the platform where all these AI agents can build seamlessly so we are basically launching a IO intelligence platform where we'll have hundreds of models which exist already by multiple builders all throughout the industry open source and we basically take our strong inventory put these models on those(...) inventories and we'll give this access of powerful models for free to the users so they can actually build the next versions and the future of the models on our platform right
Japanese: 私たちが行っているのは、すべてのAIエージェントがシームレスに構築できるプラットフォームを作ることです。そのために、「IOインテリジェンスプラットフォーム」を立ち上げる予定です。このプラットフォームには、業界全体の多くのビルダーが既に作成した数百ものオープンソースモデルが含まれます。そして、それらのモデルを私たちの強力なインベントリに統合し、ユーザーに無料で提供します。これにより、ユーザーは次世代のモデルや未来のモデルを私たちのプラットフォーム上で構築できるようになります。
English: that's the first phase which we call IO intelligence and we'll also promote more and more new companies of AI agents to basically build on top of it provide a space on our platform to deploy and be accessed by like thousands and thousands of people going forward right
Japanese: これが「IOインテリジェンス」と呼ばれる最初のフェーズであり、より多くの新しいAIエージェント企業がこのプラットフォーム上で構築できるよう促進します。これにより、何千人もの人々がアクセスできる空間を提供していきます。
English: that's just the first phase and the second phase of the same product we're building a platform where we'll like the developers will have access to the apis to the models and to the access of data where they can build new powerful models where initially we will fund our GPUs for free to a certain limit(...) for these builders and beyond that we'll charge a minimal amount
Japanese: これは最初のフェーズに過ぎません。同じプロダクトの第二フェーズでは、開発者がモデルやデータにアクセスできるAPIを利用できるプラットフォームを構築しています。ここでは、開発者が新しい強力なモデルを構築できるようにします。最初は、GPUを一定の制限内で無料で提供し、それを超える部分については最小限の料金を課す予定です。
English: and if these models are really good and will be used by a lot of people then going forward will be incentivized we'll be giving incentivization to these builders to come deploy a build on a platform we basically grow a whole economy for these web 2 and web 3 builders to come to the platform and build the future of the AI and AI agents at a very very cheap price on the most stable and the cost effective platform of the computer in the industry and
Japanese: そして、これらのモデルが非常に優れており、多くの人々に使用される場合、今後はこれらのビルダーに対してインセンティブを提供します。これにより、Web2およびWeb3のビルダーがプラットフォームに参加し、非常に低コストでAIとAIエージェントの未来を構築できる経済圏を構築します。このプラットフォームは業界で最も安定して費用対効果の高いコンピューティングプラットフォームとなるでしょう。
English: then final question uh let's say i'm a dev i'm a builder i'm excited about what i'm hearing what are the next steps i'll be with injective what do i do
Japanese: 最後の質問です。私は開発者でありビルダーで、この話を聞いて興奮しています。次に何をすればいいでしょうか？Injectiveではどのように進めればよいですか？
English: yeah the next step so you can like reach out to me or anyone on the executive team whether it be like telegram or on twitter and then we're going to get you situated with some members of the team if you have any questions on developing on the engineering side or on the marketing and economic side and but yeah it just kind of it's just kind of like we're all open to any dms and all feedback so we chat to any one of us or join our public um builders telegram and you'll have other people help you there as well
Japanese: 次のステップについてですが、私やエグゼクティブチームの誰かに連絡してください。TelegramでもTwitterでも構いません。その後、エンジニアリング面やマーケティング・経済面での開発に関する質問があれば、チームメンバーがサポートします。また、私たちはDMやフィードバックにオープンなので、誰にでも気軽に連絡してください。また、パブリックなビルダー向けTelegramに参加すれば、他の人たちもサポートしてくれます。
English: then toss if interested in learning more about our compute network what's the best way to reach out
Japanese: それでは、コンピュートネットワークについてもっと知りたい場合、どのように連絡を取るのが最適ですか？
English: yeah a few different ways um so of course you know feel free to reach out to me or go over anyone else on the ionette team here on twitter feel free to email me at tossup.io.net um also of course our self-service platform is available to anyone to go on and try out and deploy some clusters and start playing around with it to see what you need um if you feel you have more sort of customized needs or the platform doesn't quite meet the use case that you're looking for there's also a contact us form directly on our iocloud platform um where you can reach out with more specifics and someone from my team will get back to you to to help you get exactly the gpu's and the clusters you need to so yeah
Japanese: 方法はいくつかあります。まず、私やIonetチームの誰かにTwitterで気軽に連絡してください。また、メールでも構いません。私のメールアドレスはtossup.io.netです。また、セルフサービスプラットフォームを誰でも利用できるので、クラスターを試しにデプロイして、自分に必要なものを確認できます。さらに、特定のニーズがある場合やプラットフォームが求めるユースケースに完全に合致しない場合は、Iocloudプラットフォーム上の「お問い合わせフォーム」から連絡してください。私のチームの誰かが具体的な要件に対応し、必要なGPUやクラスターを提供します。
English: awesome well i'd like to thank uh avi from objective for joining us the io team as always for speaking of course all of our listeners excited to continue to collaborate and see what the future holds um and that's all folks thank you very much
Japanese: 素晴らしいです。ObjectiveのAviさん、そしてIOチームの皆さん、いつもながら素晴らしいお話をありがとうございました。また、リスナーの皆さんも、引き続きコラボレーションし、未来がどうなるかを見ることを楽しみにしています。それでは、以上です。ありがとうございました。
English: thank you thank you for your time and invitation thanks everyone thanks guys
Japanese: ありがとうございました。お時間とご招待に感謝します。皆さん、ありがとうございました。
英語原文
hello
everyone how goes it
doing well doing well it's pretty cold here in New York it's snowing a lot but pretty good pretty wild the weekend as well for crypto where all the Trump coin launches yeah
I don't think we'll get into much about the new meme coins being launched by the administration on this space but plenty of time to talk about that continued part as we go on I'd like to thank all of our listeners and of course our panel today we're gonna be not doing community questions but we'll be hitting them the ground hard again as always if you have community questions please join our discord and ask away and the team will answer them very special space today we have injective and our IOT team let's start with a brief overview from each of our panelists objective let's start with you
yeah so names I'll be I'm on the business development team and objective and what injective is for the listeners who might not know is we are in cosmo was in based L1 right now developing our EBM side as well so every dollar of unit would be a stable coin or a dap token that you have on either end of the cosmo wasm or EBM they'll be interoperable you use them side by side but we're effectively built for finance and we look at finance by the means that anything of value can be transferred globally for three 24 seven but yeah thank you for having having us on
we were just talking a moment about you know Trump coin but one of the other big narratives that's been happening in our space over the last couple of weeks is defi AI and it is been increasing for those in the audience that don't have a full understanding or maybe have just heard this word in passing could someone give a description of what that is and how both IO and injective are positioned for it
(...)
yeah sure so what defi AI is it's effectively you have a lot of these agents that are coming on blockchain and in the real world and web to business so what defi AI does is that you it allows these agents to do financial transactions on behalf of you yourself so that could be a great a perp trade on a perp exchange like we have helix or that could be you have an agent then you tell them to I have a hundred dollars of USDT how can I maximize the yield and it kind of takes that money and puts it into different yield farms across crypto but there's no one exact use case I would say I think it's very experimental right now but there's a lot of untapped uses that defi AI can't handle for everyone
(...)
and before we start talking about some of those use cases to the IO team we've been talking about defi AI but for those that haven't tuned in how is IO connected to defi AI and why is this an exciting time for us
sure I can jump in on that can you guys hear me that's a little bit of trouble with the mic loud and clear awesome awesome so in terms of how IO in particular relates to defi AI right I think in terms of how to describe IO itself we're sort of like the foundational layer of decentralized compute underlying all AI use cases and applications right like whether it's model training inference fine-tuning things like that so when it comes to defi AI I like how it was described before in terms of you know we don't necessarily know what the use cases are out there I think people understand the possibilities of what AI agents can do in defi AI today but I think the actual design space is massive and I think what the wider adoption of this will look like is still anyone's guess right so I think in terms of where IO fits into it I think actually enabling a lot of these AI agents to conduct these defi transactions on chain where every time they make a decision to sort of optimize yield across one protocol or the other or do sort of financial analyses to decide where capital is best allocated on chain that's a significant amount of inference time that each of these AI agents need to constantly do in the face of changing financial data and prices that happen constantly on chain so I think with each of those actions that an AI agent takes that's an inference workload that can be run on IO net and IO can power it using our decentralized GPU network right so I think that's an area where I think IO fits in really nicely with how defi AI is enabled I think the other aspect that's really intriguing is one sort of the actual use case royals of AI agents being able to perform actions on chain and being able to be sort of like defi AI agents sort of takes hold the next aspect of evolution will really be in terms of how these AI agents differentiate amongst themselves right because in the same way today people select where is you know like wealth managers or hedge funds or mutual funds where they choose to deploy capital based on their investment returns people will start allocating their capital to specific AI agents who are able to provide outsize returns compared to other agents or the broader market in general and the developers for buildings were these more sophisticated defi AI agents will likely have to do a significant amount of fine-tuning or sort of like data structure optimization or or retrieval log-menter generation and things like this to make sure that their AI agents are constantly able to beat the market and maintain an edge and are able to sort of absorb more capital from capital alligators and thus have an edge in the market right so I think with each of those AI agent fine-tuning and sort of like optimizations that need to be done there's a significant amount of sort of like fine-tuning training compute that also needs to be used which can come from IO in a decentralized manner so I think there's a lot of foundational ways that IO can sort of support the overall defy AI ecosystem and movement
let me ask it was mentioned and it's been touched upon about use cases we've seen a lot of potential and existing use cases let's define that between your degens your heavy defy users and those that might be new onboarding to defy again going back to the attention that's been happening from individuals that don't know as much about crypto that are now getting into it what are the use cases that exist for both of these groups
yeah so I can take that so for let's start with degens the use cases for DJ as it is right now it could be suppose there are 20 different mean points that launched on crypto on crypto lines the defy AI could be an entry that gives you metrics whether that be number of holders number of sniper bots that control supply number on the amount of volume for five minute volume and a volume whatever volume kind of gathers up all those metrics and then out of those 20 mean coins it gives you what it believes has the highest probability of going higher or the next set time period so I think that's a pretty cool use case for more of the future engine space one use case for more so traditional defy users and or more risk-averse users that are crypto native if a what I've mentioned prior is yield farming optimization suppose you have X amount of money and you want to earn risk-free yield or not risk-free but you want to earn yield on that capital without taking a lot of negative drawdown on your principle so it could be a defy agent that kind of scours through all of the yield that's available on chain and tells you which one is the best for you to park your money in given a certain time period and the last one you mentioned non crypto users so I think one of the biggest problems in crypto that's persisted is that it's still difficult the onboarding is so difficult for onboarding someone who has $0 in crypto kind of first buy his or her first crypto token open up a wallet and do all that so all this can be abstracted away if there is suppose a defy agent front-end
(...)
wallet where you can put in like you can wire in or we can put in $100 and cash and then you can tell him or you can tell the agent I want to in I want an index of the best AI token for you without having to go across and do all that kind of stuff that is really difficult for non crypto natives to do
(...)
and to the IOT team you've all talked about what to look for on the dev and builder side when thinking about a compute network thinking about those case studies is there anyone in particular that seems more if you're building on any of those that should be something you should be thinking about decentralized compute network so if you mentioned some use cases do any of those stand out as priorities in it at all
to be honest like the way the industry is headed I think in the near future you will have companies or industry let me defy or others who use AI or basically visit out over a period of time right and when we start talk about defy and the use cases right so or even the AI industry overall I think any vertical industry that is data intensive which requires transparency or can benefit from the decentralized infrastructure basically we be the adopters of the early adopters of defy technology because like basically when you are combining the AI intelligence with the defy openness and efficiency then the sectors can unlock the new value streams and can basically drive innovation at scale right and it will be in every sector whether you talk about institutional finance asset management let's say you take an example right like hedge funds what do they do they basically can use AI today to analyze liquidity pools basically across multiple chains identify arbitrage opportunities and execute trade basically autonomously you think smart contracts right now the whole decentralized concept itself will help this defy industry to bring trust so basically more adoption right similarly in the gaming and metaverse side same thing could be said you can come to the real estate and the corresponding tokenization of the asset which is happening there same case you can talk about supply chain you can talk about renewable energy I think any market you touch right now you will see the decentralization penetrating that particular market and AI penetrating that market more and more and I think where Injective and IONET will succeed is that the IONET like the Injective team has created this whole like AI kit which will help the builders greatly to focus on just the business problem which people want to focus on and all the other problems statements and abstractions which they normally have to build for that will go away and where IONET helps them overall as a player for both Injective and the builders on top of Injective is that whenever the startup comes in they have to really think of the scale the resilience and the cost right and IONET brings those values to the table that are inventories decentralized so whenever there's a spike of requests for that AI model which is getting built on Injective they don't have to worry about that right the startups are coming in they don't have to worry about the cost of the infrastructure on which they're building right they don't have to think about there's any centralized players which can bring the whole company down in future if it starts to get traction and if there's anything gray area right because it's it's basically really resilient to censorship so overall you'll see all these industries benefiting in in in time going forward and I think Injective and the ION partnership which has happened that will play a major role in the months to come you'll see.
(...)
Injective let me hand it over to you we're talking about the partnership where do you see it on your side of the equation?
(...)
Yeah so we see it in a very similar manner so the way we see it is that the Injective SDK that we built out well I think Hildiard just comes to right now we believe that we're in the very early experimentation stage of DeFi AI and how do you most benefit from the early experimentation days is that you give developers you decrease the friction to experiment and to build for developers so in that case what Injective did was we built out an SDK which allows developers kind of plug and play and easily build out DeFi apps on Injective and what we really loved about this ION partnership was that so okay now they have the SDK to build out these dApps but they also need compute and like to make that easier for them so they don't have to worry about where do we source compute for inference they can just come to IONet and use that(...) in a very similar plug and play manner where it's all kind of available to them really easily so all they have to focus on is experimenting and trying new stuff out to see what works and what doesn't work.
(...)
And speaking of friction first I'll ask you Injective where do you see the friction happening now and where do you see that evolving as the space continues to evolve?
(...)
Friction in what sense? In what part of stack?
Yeah the friction that you mentioned that you're solving with your SDK and the partnership with ION in building compute network.
Yeah so I think there's two key points where the friction lies so the first key point is just getting developers the tools to build out these experiments really fast and then I think the second friction point which is much later in the kind of curve is actually finding product market fit with these dApps, DeFi dApps. So I think right now we're trying to solve the first part of the friction where we're giving developers all the tools kind of build out an experiment and increase the amount of experiments launched every day every week on an every month basis to kind of increase the probability that one of the experiments going to find PMF and kind of like block subs over time.
(...)
And Io over to you same question.
(...)
Yeah I can I can jump in in terms of sort of how we think about reducing friction points right
(...)
especially from a decentralized compute perspective.
(...)
So I kind of really like to start to think about these problems like from the end user perspective right and what and the developer perspective in terms of what they're actually trying to optimize for when you're trying to build sort of you know the most sophisticated DeFi AI agent that's going to you know attract the most capital and offer the best returns and it kind of almost takes me back to sort of how hedge funds think about optimizing their returns from a technical perspective as well. So I think there's three things that I kind of think about in terms of producing friction enabling DeFi AI developers to be able to sort of build the best agents they can(...) on a decentralized compute network. So I think in terms of what we offer
(...)
I think first and foremost is sort of the permissionless access to compute right(...) through IONEC. I think this one is fairly underappreciated in this context because you know theoretically sure an AI agent could go get compute from an AWS or Google Cloud or something like that too but I think having that permissionless access sort of really offers up a benefit in use cases where you know crypto finance and DeFi is a 24-7 market that moves very very quickly and the last thing you want is for your DeFi agent to get liquidated because you know it got stuck in a support queue at AWS right because it's going to get access to compute or had a billing error or something like that. So I think having access to compute through a permissionless network like IONEC is going to be sort of one of the de facto things that any DeFi agent is going to need in order to be able to ensure that it can accomplish its directive in the most effective manner.
(...)
The other thing I think about and you know kind of connecting to the hedge fund example again is sort of the idea of latency right like I talked about sort of all these DeFi agents needing to have needed to perform a constant inference to analyze different market trends, different market data, prices, volumes, things like that to go to make the best investment decisions and one of the things you know like the amount of optimization around latency in the traditional finance world and high frequency trading is such an important issue that a lot of like hedge funds actually have their ID infrastructure located like close to where the exchanges are so that they have the data that they need for their investments you know with an advantage of milliseconds above other competitors and that can make the difference at high frequency trading. So very similarly the optimization around latency time will be really important here and I think with our distributed compute network on IONEC we're able to actually offer, we're actually able to offer incredibly reduced times when it comes to latency for inference because we're able to provide the node closest to where the inferences happen right like whether the user is in Indonesia or Brazil or the US or Finland we have GPUs available everywhere so that time for someone to perform it for an agent to perform inference and get a response back after it's been computed is massively reduced compared to that same inference workload or every inference workload having to go to US East or wherever you know the massive data center from AWS is located to be able to give you an answer so I think that's an area where we're reducing a lot of friction too and I think lastly it's also just cost reduction right like as these AI agents look to optimize the returns that they offer for their DeFi users every user is actually is kind of want to look at which AI agent is able to have the least overhead as well right so that they can provide more of the returns back to the users providing them capital rather than having to spend it on compute so to take it back to the hedge fund analogy again it's hedge funds often charge the the two and twenty model for management correct like two percent management fee twenty percent on returns so in a very similar manner I think and you know different hedge funds compete on sort of like what date rate they they take on on the capital allocate but then so very similarly I think with AI agents they're able to reduce the cost of this compute usage by using a decentralized network like IO and producing that friction point they're able to sort of provide greater returns back to the users who allocate capital to them
(...)
and speaking of AI agents how does Injectives iAgent framework enable the seamless integration of AI agents could you give some details about what exactly is under the hood
(...)
yeah so I can't speak too much about the technical aspect of it but the TLDR is that(...) it kind of the framework itself provides seamless integration onto the Injectives blockchain and network as well so whenever you're using the framework you can easily build that and kind of add that agent let's say for example of trade on helix which is a club it's a perpetual and spot club that's built and embedded within the Injective network itself so it gives you that seamless compatibility that a lot of other supposed frameworks might not offer with their base pain
(...)
and you know a follow-up question I have is Tasha was talking about in two levels friction but also what you had mentioned after experimentation comes scalability what have you noticed in terms of scalability so after the experimentation is done or and a builder has created a successful doubt where do you see the next stages of friction of work that needs to be done in our space
(...)
yeah so I think that look to be honest I think we're still very very early we're still in the experimentation part of this I mean it's like a month or it's only a couple months old that we've seen a lot of AI agents pop up I still do not believe we've had like any agent find product market fit and move on to the second part of the stage but I still think that we're still there everyone's still experimenting there hasn't been too many other problems outside of just experimenting getting deep and continuous access compute which is being solved by IONet but I haven't seen the second part of that friction really come out to date but we're I also believe we're very very early in this
so so I think I'll take a different stab at it right I think(...) there are some fundamental problems which always exist in system and if you basically solve that fundamental problem your your rate of growth is exponential post that right so for example like beat any company right normally when they like when the company starts up and they start to become bigger and bigger the first thing they try to optimize on is actually on experimenting that how as a company I can change let's say my pipeline of development so that I can change more changes quicker and less time right how many developers at the same time can make changes so all the web two companies or web three companies basically optimize their tech stack and operations in a way that more experimentation could be done because that enables more development in parallel more changes to go in the outside world in parallel see what matters faster and what is basically client's liking and then build on top of it so these are like one of the fundamental being able to experiment experiment fast with minimal of minimal need of changes is one of the fundamental things if you can do it well you have exponential growth right that's one I think that's where this particular thing which the injective team is doing I think as in like it's a very very fresh concept they have taken a first tab at it when they become matured more and more people know about it I think there'll be a huge adoption and people will see a value and it'll just keep like becoming more and more bigger and important in the industry you'll see right and the other thing which I feel is a fundamental change also is that if you take a look at the centralized players right for example let's take about AWS itself like why many players use AWS right they can just go and take a bunch of bare metal machines from somewhere and build on top of it why do they go to AWS because they have the ML kits there which their teams can use they have a deployment pipelines on AWS which they can use right whenever they want to scale their CPU and GPU they have that software stack so what happens in in that case is that whenever startup comes and let's say if they only have eight ten people those ten developers or ML scientists(...) will predominantly focus on the problem statement not on the periphery problems of creating the infrastructure and all that we just focus on the business problem and those kind of companies succeed faster right and that's where the ROI of each developers matter right now what has happened with this partnership is that injective is solving the AI agent kit side of the problems that the developers when they come they have less amount of friction when they build their businesses they don't have to they don't have to focus about other integration and all those things with the SDK which these guys have given right where I.O comes in picture is that I.O is giving that cheap compute that when a startup even have a slow low amount of funding they can still build on top of it right there still will be censorship averse right they still can scale at a at a very high level and they'll have access to the strongest of GPUs at the cheap price right so this partnership will enable developers to build confidently iterate many times faster and can experiment at a very cheap price which is the fundamental values any like a startup would look out for I think that's the way I take a look at this partnership and what both these teams are bringing to the table.
(...)
Awesome awesome what's up guys I'm Zach from the I.O team really excited about being a part of the injective agent kit and I guess just being that this is our weekly spaces and we have a lot of people who don't have background on injective and what you guys have done and kind of the impact that you guys have had in the space do you mind just giving us some background for for listeners that are less familiar on the history of injective and how we've gotten to this moment and how you guys stand out in this crowded DeFi L1 landscape.
(...)
Yeah for sure so some background on injective launched a couple of years ago there's two good founders they launched initially as a simple Perpetex using the cause of wasm staff and then shortly after they realized that the biggest benefit of launching or using crypto rails to launch and expand is the compatibility aspect that crypto offers that crowdfi doesn't offer so then they launch injective which is an L1 and then helix is the perpetual and spot DEX embed embedded within injective itself so you can think of it as any DeFi protocol any protocol any token that is built on injective is seamlessly can be used on helix itself and this lends out to very cool um very cool and innovative ways to add more capital efficiency into finance so I guess another thing is that injective is made for financial applications and it and one example this could be suppose if there was a DeFi agent that is building out a(...) F1 strategy where it allows anyone to deposit capital and it trades it beyond it trades it on behalf of its users one thing that you seem to see on injective is that suppose you put in ten dollars of your money into this DeFi agent then you could take that Lpster that you got from that ten dollars and deposit that on the helix to trade perps yourself so the end goal for injective is that it's built with composability in mind and with capital efficiency in mind is to bring not only crypto native DeFi and prop that up but also work on bringing traditional capital whether the rural assets or other stuff on chain um and at least grow finance in that sense
Cool um and then I guess like what are some exciting uh real world use cases uh that you've seen I mean I personally uh like to call it AI-Fi uh but uh DeFi uh is is the is the popular colloquial term um yeah so what are some exciting uh projects that you guys have gotten some inspiration from um on this SDK and I guess like yeah what do you find most exciting going on in this AI-Fi space at this moment for injection yeah definitely
so there's the example I just gave um about the so I think one thing it's also like it's like DeFi is like DFAI or A5 but yeah um the naming is really it's hard to say but the example I just gave is there's a developer right now using the Injunctive SDK to build out a authentic finance um hedge fund where people can kind of give this agent agent money and then the agent will trade on behalf of will trade on behalf of them on Helix itself and then kind of distribute more distributed earnings on a monthly basis but I think that's very cool where instead of getting suppose going back to the hedge fund analogy you see a prior instead of giving a human your capital to invest on behalf of you now I think shortly will certainly soon we'll be seeing people getting agents their money to herd yield and profit you know on top of and I think one long-term difference this is going to make into market structure is that it's going to be much more capital it's going to be much more efficient where there isn't going to be a lot of different price discrepancies between different sexes and different sexes for Bitcoin let's say because of many more players whether they be agents kind of neutralizing the price discrepancies
cool um yeah so rip to the uh to the people who are trading on those on those opportunities um and then I guess uh what are some current limitations uh within either DeFi or AI whether that be um I guess interchain operability um I guess like what what limitations do you currently see uh in the way for uh for this space specifically um and hurdles um that we'll need to get through in order to have like a fully uh functioning AI-Fi environment um yeah
(...)
yeah I can take a start right here I think like go ahead the indicating I think you were about to say something
(...)
oh no go ahead sorry go ahead go ahead
so I think the first thing is uh we need to have a enhanced usability and accessibility uh in the first place so I still believe(...) there are like several key developments which need to occur which is across your technology your user experience or your regulatory side and overall the other ecosystem building side as well for this whole DeFi space to grow right and it starts from the first one it is like your enhanced usability and accessibility right I believe the AI agents must become more intuitive and user-friendly uh so that it can becomes more seamless to use both fucked up to natives and the newcomers like I genuinely haven't seen till now uh basically an agent where a complex DeFi uh agent uh uses NLP and you can see simple things like where can I earn the highest yield on my eat and the AI agent like could just handle this research do the transaction staking automatically and end-to-end do everything right I'm surprised nobody has done a multilingual support uh till now right like basically like any good AI agent should provide services to a global audience right by offering support in multiple languages we don't have that one of the things I think uh which I feel in many areas in crypto including the deep end or DeFi is that we I don't know why we compete a lot between ourselves and we don't work together as a vertical industry to become better right like I've seen that there's no unified standard API right the the collaborations like Injective and I have done there not many collaborations like this which basically creates synergies and the(...) better value in the lesser amount of work as protocols and giving more services and value services to the customer I don't see that right on the security and press side there have not been many innovation being done right like many of the projects still on the DeFi side I don't have the audible code right transparency isn't there they they're their models which they're deploying is still not there visible to everyone right similarly there are more models itself which should be created where like you just throw those models on any smart contract which comes and it audits it and transparently show the results to everyone in the audience and then if it's good then deploy it right so that everyone knows what's happening what were the different(...) issues which came in this code which they were with the respective protocol was thinking of deploying and if they have done the fixes what the second version of the review was and so and so forth right so people actually have confidence in the industry overall that that what the teams are doing is is real is audible they can read understand it and even if they was wrong how it was fixed and how soon it was fixed and what was the final outcome there's there's nothing around right the interoperability across the protocols and change is still very bad right there's not much innovation being done on the AI agent side where it's seamless to work like across multiple DeFi protocols blockchains as a holistic solution I think this is one area where as industry we can become better right there was always a there's a part of cost I think that with a lot of deepens has been solved there's some issue the real-time data also I think these are few friction points which exists in the industry technically which could be taken a stab at and we'll be better as industry and then obviously there's there's the whole education piece also that still there's a lot of friction from web to giants and these great builders who know how to create these(...) distributed system they're still not coming to to the web tree side because we as industry are not educating them enough about the benefits of the decentralization the tokenomics and what they can take advantage of this whole blockchain ecosystem and crypto ecosystem I think that education is is also not being spread out and obviously the last piece which is a problem is the regulatory clarity right the regulatory uncertainty around the DeFi and the AI could really hinder the adoption right a clear framework will create confidence in the ecosystem and I think the new government will just come in everyone's really hopeful that like the best days are coming pretty soon right so I think that that's all which I had if this these are things of just all like we'll be much better than the strict
(Music)
(...)
Cool awesome and then I guess you guys that injective if you want to answer the question about other hurdles that you guys see in the way and then I guess yeah another side question
(...)
for for injective is us at Ironet where we're happy and excited to be to be also partnering with the likes of creator bids the rebrow and AI 16z in the space and then I guess outside of things going on at injective what do you guys find most exciting within the crypto AI space or the AI 5 space that's kind of inspiring to you guys
yeah okay so that's a lot let me break it down for the first question I think it's best to view like what was like the it's best to view the current challenges as going back to the analogy of who's the user so we can split up the users as crypto-dgens crypto-natives with that are more risk-averse like funds or whales and the last one being will who are not creative so we'll split split up two ways the first one I think thus far we've seen the most activity for these crypto agents or dia dia 5 among the main coin traders last dgens because they tend to be more risk-averse and a lot of the tooling that's being built out or a lot of the tools that have a lot of usage thus far have been crackers of um like deep trenches is this like what's the probability that this tokens are run or not so I think that has found really strong product market fit but going to the crypto-native whales last funds I think a big problem is you guys mentioned that is that security is still unclear and it's difficult to give a AI agent let's say a large sum of money if there hasn't been any will in the effect that it is safe and he will not lose my money or there won't be a smart contract risk where I kind of lose all my money kind of going back to the early d5 days where d5 days where there's a lot of different borrow lending protocols on ethereum but over time there only a lot of the large crypto players feel comfortable in allocating capital onto maker dower obage because they survived months and years worth of um usage and while whereas a lot of the other borrow lending particles suddenly died and got robbed a couple of times and the smart contracts were not safe so I think that's a similar thing that's going to happen with d5 where it's going to take some time for the products to prove that these market products themselves are safe and that they are
(...)
the returns that they promise or whatever the product might be is that's sustainable over time and the last part regarding people that are not native I think that comes down to again you guys mentioned this as well is just accessibility where it's still not at the level that if you want to create a snapchat account or a instagram account or tech account it's much easier to do that than create a account where like a d5 agent that manages your money so I think it's the hurdle that's going to pass over time but that's probably the biggest friction onboarding new users into these products now and sorry what was your second question
my second question well first of all I just like I just like to acknowledge that yeah I think like that's the the total end dream that these consumer um the the regular consumer can come on to into the crypto world through one of these bots and not have to worry about bridging to Solana bridging to ETH just that it all works seamlessly and they just see the fee alongside it and then the second question was just outside of things going on at Injective what do you guys find most exciting and inspiring right now at the moment in the crypto AI space or the AI 5 space yeah yeah yeah
definitely so I think the thing that excites me and excites us the most is the amount of net new talent and net new developers coming into crypto for the first time so we haven't really seen this kind of bloom in talent in crypto for a year or so it's been a long time but the AI crypto AI or DeFi has really encouraged a lot of my web2 friends kind of work in traditional tech or work in the AI not the world kind of give a stab at crypto kind of play around um nights and weekends kind of see what they can build in crypto so I think that's the most exciting part where we're seeing a lot of new talent come into crypto industry because of crypto AI and DeFi.
(...)
Let me pass the question of new talent to the IOT team you all have talked about you know and mentioned on this call about friction for new developers people coming into web3 and the issues facing them do you see with DeFi AI any other unique challenges or do you think that it follows the traditional developer breaking into web3 challenges?
(...)
I could take a stab at it. I think
(...)
you know in terms of new talent entering the market I think sort of like this AI agent boom has really been kind of a boon for the overall sort of crypto space right and I think it sort of like really inspired a lot of net new developers with sort of broader skill sets to enter the market kind of explore sort of the various like toolkits and SDKs and you know sort of like programming languages that are available in the blockchain space and really try to come at it with a fresh lens and build on top of it from this lens of AI agents right so I think as as Abhi mentioned I think it's been like a really exciting time of seeing all the people pour into crypto again you know after sort of like the bear market of 2022 when a lot of people left the space I think people are kind of coming back and you know in sort of a larger you know almost an order of attitude more than what was what was before because not only are we attracting sort of you know classic like web3 developers but a lot of folks who have never dutch web3 have never thought about crypto before have been you know sort of primarily in the AI ML space or experimenting with AI agents I think they've started to discover you know a lot of the benefits of being able to build AI agents on crypto rails like the benefits of like permissionless finance like access to sort of like decentralized compute data oracles things like that like a lot of tools I think they're trying to sort of have this aha moment of realizing that a lot of what we've been building in crypto and web3 for several years are a natural fit with what AI agents actually need and how they'll likely operate in the future so I think this marriage of AI and crypto and deep DFAI I think has sort of like really created a new paradigm of development and sort of really opened up kind of the design space right and and Zach you mentioned a few of the folks we've we've partnered with you know so Shaw, the i16z, like Jeffy and Augustin and Zarebro are you know great examples of that like there were folks who were primarily on the AI agent side are sort of you know fully in the AI development side before came into crypto did a lot of interesting experimentation and have now built sort of like billion dollar projects right and in a short amount of time and have also sort of like enabled a lot of other developers to build on top of sort of the foundational open source stuff that they've done so I think it's it's a really exciting time for you know the future of AI and crypto and for talent coming into this space.
(...)
Let me ask everyone on the panel hey what do you say to people who say great this is just another bubble we've heard all of this before about onboarding new people that things are going to become easier that we're getting interest from outside the space that's happening what makes this sound different what are you all seeing in use cases technology growth(...) from conversations you're having and just generally looking at the space.
(...)
Yeah I think it can be both right it can be both a bubble that's kind of blown up over the next couple of months or years but also have really strong fundamentals where the team's building out some of the projects will sustain over the next decades to come so I kind of agree with Bill takes that there are fundamental use cases for example right now a lot of blowing fruit spin being solved in the trenches and what that how kind of new tech usually evolves is that kind of it starts as the first adopters tend to be the most risk on people and then over a longer time rising it gets onto more risk-averse people kind of start using similar products using the underlying technology so I think the underlying technology of authentic finance makes a ton of sense given how it can increase composability in the financial market and increase capital efficiency
(...)
and finance is kind of built on top of capital efficiency so fundamentally it's sound but that being said I can I also see it kind of being a bubble over the next couple of months to years giving you out of interest so I think both are true.
(...)
I think for me personally it has not been hard in a previous experience and now in IO also to basically convince people to come to the crypto side because if you break it down right I don't think in crypto companies there is any more risk compared to a web2 startup also right like it's the same risk in in a crypto startup as you have in a web2 startup right the percentage is more or so will look similar right in terms of the problem statement which you solve I think like crypto has more interesting problem to solve compared a lot of web2 giants to be honest right and practically speaking the velocity at which the crypto companies iterate and build is far faster compared to any web2 company right and at the end engineers are builders and they like to work on interesting stuff ship it in front of people get their feedback and build on top of it right that's what excites them the most right so that is practically there in the crypto company like imagine like in IO personally we're building cloud platform which took your Amazon years to build right we're literally solving the same problem which Amazon did so for a builder who has seen those products and they say oh we'll get the same product we built here in the next one and a half year there's so many interesting problems solved they're super excited right and practically speaking I think the founders and the senior guys were there in crypto they'll practically tell you the salaries in the crypto companies are far better in the web2 site right so it's it's you don't have to compete in that side also because we do pay well in our industry anyhow so attracting people in terms of the problem statement and this thing is comparatively easy like at least for me it has been historically easy the only thing I think as as tech people we have to do is that when new people come in in the industry they get overwhelmed right for a new person who's coming in the whole industry itself they're normally risk on they can come and join any industry web2 or web3 right the problem comes in if somebody is a 12 year experience guy he's basically one of the senior most person in a web2 company and for him to leave a web2 company a comfortable job where he's proven his value in everything to come in a web3 site where everything is new for them that's overwhelming right and over there for those guys for make them an easy transition as leaders we have to break it down that we are basically solving a problem which we'll cater to a web2 side also and web3 side also and here is a plan for you that in the coming six months this is the problem statement you will solve which will predominantly web2 focus distributed systems and slowly and steadily we will train you on web3 and as long as you show that transition plan to them those senior guys are comfortable but if you just throw all these learnings to them in one shot which is blockchain distributed systems decentralized system and all these other tokenomics then it's becoming overwhelming and those kind of people don't transition well are afraid to make that jump from web2 to web3 i think that's where this transition and explaining things helps a lot
just to add a few things there too um actually i'm going to go back to these point on bubbles um that he made earlier too and kind of how that how that plays out right so
(...)
kind of as he mentioned um there's you know a couple different ways to look at it and i think often people kind of like look at bubbles as sort of like this evil thing um you know in financial markets but i kind of really think of it as sort of being a natural sort of part of technological innovation and progress right and that's kind of really what we're doing here is with dfa and ai agents we're sort of at the bleeding edge of technological progress and what is really possible to do in finance and crypto and ai um and you know i think historically with bubbles like whether it's related to you know the internet or railroads or crypto um i think what it really sort of you know tangentially is related to is this idea of experimentation right which we talked about earlier as well i think bubbles typically sort of are a breeding ground for experimentation and it sort of like attracts sort of like this initial capital injection really attracts a lot of talent into the space sort of breeds a lot of new ideas gets people excited about trying out a lot of different things trying to build and solve for a variety of different problems using this sort of like new technology um and you know in some ways a lot of those experiments don't end up working out and you know whether it's a company or a project or something like that it can end up going to zero because maybe the technology just wasn't quite there yet or like the adoption curve wasn't quite there yet but what it really does to the experimentation is it shines a light on gaps and things that need to be solved for the overall movement to be successful right like yeah as i mentioned whether it's internet or railroads or crypto or AI agents(...) and what that kind of really does is sort of it helps the next generation of builders really sort of tighten focus and solve for those gaps and kind of identify what kind of things need to happen to be to enable that business to be sustainable in the long term right like to take sort of like the bubble example one of the you know biggest stories of that time or companies at that time was pets.com right like getting pet food over the internet which was sort of often used as this example of being a ridiculous business that you know described the hubris of the dot-com bubble but now is you know a really major sustainable profitable business in the form of chewy.com um and you know and i think you can kind of make that argument about a variety of different things like whether it's like airbnb or uber or instacart a lot of those ideas were tried out in like back then in 99 didn't work out but a lot of the builders stuck around on the internet stuck around building apps figured out what gaps needed to be solved and then now they're you know multi-billion dollar businesses so i think we'll go through a very similar cycle where you know we'll have the amazon's and google's who'll we're being built now in you know an equivalent space and we'll survive whatever sort of bear market comes in the future but i think the experimentation that will come from this will really drive the overall space forward and ensure that the movement is sustainable for the future.
(...)
So as we close out the first question i have of two ones left is(...) what's in the pipeline that people should be paying attention to i'll start with Injective.
(...)
Yeah so um on our end is this particular to DeFi or Injective in general as Injective in general? Okay cool yeah so on the pipeline i think one thing that internally the team has been grinding on the past couple of months is our EVM mainnet launch so that is probably the biggest thing that we will be soon going live with over the coming months and that again that's again decreases the friction for developers to launch OnInjective because we realize that there's a fixed amount of developers comfortable in writing on custom awesome code so with this development and with the compatibility between our EVM mainnet and our mainnet i think that's gonna
(...)
enable a lot more devs and a lot more action on Injective mainnet.
(...)
And I know there's a bunch of things coming up what in particular is something that people should be paying attention to?
(...)
I think it basically attaches to the same question which you asked earlier so when we actually(...) plan for any quarter we basically do the research and like you asked like whether it's a bubble we don't see it's a bubble our research said that the whole AI agents it's not just a web3 phenomena it's basically web2 and traditional enterprises are equally invested and which shows basically isn't a bubble but a universal shift towards smarter more efficient systems across the industries like InShot and what we are doing is we are creating the platform where all these AI agents can build seamlessly so we are basically launching a IO intelligence platform where we'll have hundreds of models which exist already by multiple builders all throughout the industry open source and we basically take our strong inventory put these models on those(...) inventories and we'll give this access of powerful models for free to the users so they can actually build the next versions and the future of the models on our platform right that's the first phase which we call IO intelligence and we'll also promote more and more new companies of AI agents to basically build on top of it provide a space on our platform to deploy and be accessed by like thousands and thousands of people going forward right that's just the first phase and the second phase of the same product we're building a platform where we'll like the developers will have access to the apis to the models and to the access of data where they can build new powerful models where initially we will fund our GPUs for free to a certain limit(...) for these builders and beyond that we'll charge a minimal amount and if these models are really good and will be used by a lot of people then going forward will be incentivized we'll be giving incentivization to these builders to come deploy a build on a platform we basically grow a whole economy for these web 2 and web 3 builders to come to the platform and build the future of the AI and AI agents at a very very cheap price on the most stable and the cost effective platform of the computer in the industry and
then final question uh let's say i'm a dev i'm a builder i'm excited about what i'm hearing what are the next steps i'll be with injective what do i do
yeah the next step so you can like reach out to me or anyone on the executive team whether it be like telegram or on twitter and then we're going to get you situated with some members of the team if you have any questions on developing on the engineering side or on the marketing and economic side and but yeah it just kind of it's just kind of like we're all open to any dms and all feedback so we chat to any one of us or join our public um builders telegram and you'll have other people help you there as well
then toss if interested in learning more about our compute network what's the best way to reach out
yeah a few different ways um so of course you know feel free to reach out to me or go over anyone else on the ionette team here on twitter feel free to email me at tossup.io.net um also of course our self-service platform is available to anyone to go on and try out and deploy some clusters and start playing around with it to see what you need um if you feel you have more sort of customized needs or the platform doesn't quite meet the use case that you're looking for there's also a contact us form directly on our iocloud platform um where you can reach out with more specifics and someone from my team will get back to you to to help you get exactly the gpu's and the clusters you need to so yeah
(...)
awesome well i'd like to thank uh avi from objective for joining us the io team as always for speaking of course all of our listeners excited to continue to collaborate and see what the future holds um and that's all folks thank you very much
thank you thank you for your time and invitation thanks everyone thanks guys