Agent loops are not a default, even if it's viral.. They only work when the task repeats, the agent can run the code, and something inside the loop can reject bad output. Miss that gate and you are just automating token burn.
CODE
-
METR report: most SWEBench results unmergeable; FrontierCode unsolved by frontier models
By
–
It's finally out!!! @METR_Evals found that more than half of SWEBench results is unmergeable slop. FrontierCode represents over 1000+ hours of maintainer validated software engineering work most frontier models cannot yet solve, much less solve with high quality. Cog had IOI
-
AI’s coding vs biology: bio databases designed for different traffic
By
–
New Science Blog: Why has AI advanced faster in coding than in biology? To agents, bio databases are like cities built before cars—maddening to drive in because they're designed for different traffic. How do we build infrastructure agents can use?
-
Use ‘Approve for me’ as default for Codex tool calls
By
–
Codex tip: Do yourself a favor and switch to using "Approve for me" by default You should not let your coding agent run free without any checks but also approving each tool call isn't the way "Approve for me" is a good middle ground as it let codex review each tool call and
-
MagicPath cloud canvas enables creation from any agent
By
–
Very different. MagicPath canvas lives in the cloud, so you can create from any agent without opening the app. Imagine using Codex on mobile to design in MagicPath connected to your context, then getting a live link directly on your phone. No other tool can do that. Most
-
Agentic features come to NotebookLM with Gemini 3.5
By
–
The agentic features that we have been enjoying for a year and a half in ChatGPT, Claude and Gemini now finally arrive at NotebookLM. It will be able to run code in a virtual machine, work with sources, use skills, etc. That and the use of Gemini 3.5, now in NBLM 🙂
-
Give Codex a clear outcome and let it keep going until achieved
By
–
codex tip: give codex a clear outcome to work towards tell codex what you want to achieve, how it can verify it got there, then let it keep going until it does for example: “make the build faster” is quite vague “measure the production build on vercel, keep going until it’s
-
Two new resources: interactive ActiveGraph site and AG Coder
By
–
two things ready to share from this weekend:
— Yohei (@yoheinakajima) 8 juin 2026
📖 https://t.co/tter8exKds
interactive site teaching activegraph concepts
blog: https://t.co/1V383NkkWb
💻 AG coder (open source)
reference coding agent on activegraph
GH: https://t.co/9ND9dMb0jl
blog: https://t.co/lPgi3w5F7X
still… pic.twitter.com/d1Pxzljm2ytwo things ready to share from this weekend: http://
learn.activegraph.ai
interactive site teaching activegraph concepts
blog: https://
activegraph.ai/blog/introduci
ng-learn
… AG coder (open source)
reference coding agent on activegraph
GH: https://
github.com/yoheinakajima/
ag-coder
…
blog: https://
activegraph.ai/blog/ag-coder still -
Hivemind launches continual learning for AI coding agents
By
–
New from Hivemind: continual learning for AI coding agents, available to everyone starting today.
— Chubby♨️ (@kimmonismus) 8 juin 2026
It takes the traces from every agent your team runs (Claude Code, Codex, Cursor, Hermes, Pi) and turns them into reusable skills, then pushes those skills across all of them, all on… https://t.co/MFS1ZpCh4DNew from Hivemind: continual learning for AI coding agents, available to everyone starting today. It takes the traces from every agent your team runs (Claude Code, Codex, Cursor, Hermes, Pi) and turns them into reusable skills, then pushes those skills across all of them, all on