AGI is closer and closer: "After 80 iterations of the DGM, the coding agent’s performance increases from 20.0% to 50.0% on SWE-bench, and from 14.0% to 38.0% on Polyglot 😱"
This is a great paper arxiv.org/pdf/2505.22954
Podcast 👇with a report using Gemini 2.5 Pro +original
Muchas gracias McGee por llevarme de tour por Mendoza ayer 🙌🏼
Muy cebado con lo que están construyendo con chrom.ar, agent orchestration layer sobre Base 🤖
notebooklm.google.com/notebook/3b9dd… Sharing this conversation with a bit of some of were we see AI OnChain automation is going! We R cooking great things ☺️
I can't believe Elon Musk has laid out a concrete plan to get us to Mars and it's not getting the attention it deserves! We're literally on the verge of becoming a multi-planetary species. Hey anon, we are going to MARS!!!
We just automated an on-chain Swap using Nathan Apple. Drag, drop, deploy - that's it. This is what Web3 accessibility looks like.
Built with LI.FI on top of Base
New Market Opens up on Polymarket=> LLM Strategist makes the analysis and defines investment => Onchain wallet makes the investment. Simple automation coming soon to your door for no coders. Information gets created.
HTTP APIs evolved to serve human developers and browser-based applications, not AI agents. MCP addresses AI-specific requirements from the ground up: runtime discovery, deterministic execution, and bidirectional communication.
For AI-first applications, MCP provides structural
AI as a First-Class Citizen in Web3.
What if....??? every DeFi protocol, and decentralized application (DApp) could function as an MCP server or client😜🧐