We asked @_seanli of @MagicNewton about the concept of a trusted agent and untrusted agent.
"I like to describe trusted agents through analogies"
"It's like a self-driving vehicle where you set a destination and parameters like speed, and the car follows through to get you from point A to point B."
"In the middle, you want a secure computer that does the task based on your parameters, and no one should be able to tamper with that"
"Everything is recorded in the verifiable part, every turn and action is recorded like in a flight black box with a notarized log where you can check if they did the correct actions afterwards." 🦣
"That's the ZK proof part, and underneath we use @SuccinctLabs and @RiscZero for the infrastructure."
NEW EP: Sean Li on Application Abstraction, Verifiable Agents, and AI’s Trust Problem
A new UX for onchain interactions is being built before our eyes.
In today's episode of Mammoth May 🦣, @ayyyeandy sits down with @_seanli of @MagicNewton to explore:
> Why 80% of DApp Interfaces Will Disappear
> How To Build AI Agents That Can't Go Rogue
> Cryptographic Proofs vs. Traditional Trust Models
> From Embedded Wallets to Verifiable Agents
> The Future of Natural Language Crypto UX
> Why Web3 Infra Ousts Web2 for AI
Full episode links below.
Timestamps:
00:00 Intro
01:13 Mantle Ad
01:54 Sean’s Music to Tech Journey
07:16 Embedded Wallets & Starting Magic
14:54 AI x Crypto Landscape
19:11 Trusting AI Agents
24:46 Agent to Agent Interactions
28:49 Omni Ad
29:17 UX Abstraction Evolution
32:51 Newton’s Value Accrual Model
37:04 Trusting Tools With Your Capital
40:25 Enforcing Agent Guardrails
44:20 AI Ethics
48:32 Newton’s Upcoming Roadmap
53:32 Privacy in AI
58:19 Future of Newton
7,317
0
本页面内容由第三方提供。除非另有说明，欧易不是所引用文章的作者，也不对此类材料主张任何版权。该内容仅供参考，并不代表欧易观点，不作为任何形式的认可，也不应被视为投资建议或购买或出售数字资产的招揽。在使用生成式人工智能提供摘要或其他信息的情况下，此类人工智能生成的内容可能不准确或不一致。请阅读链接文章，了解更多详情和信息。欧易不对第三方网站上的内容负责。包含稳定币、NFTs 等在内的数字资产涉及较高程度的风险，其价值可能会产生较大波动。请根据自身财务状况，仔细考虑交易或持有数字资产是否适合您。