foo

@steipete This is exactly what my workflow does, which is opensource: github.com/team-volt/volt…

am I a graph engineer now

@pvncher voltflow handles everything for you, give it a shot: github.com/team-volt/volt…

@Teknium @Mr_Bitcoin @NousResearch A2A is the answer. I did a fair amount of testing and supplied some PR's, which you brought into your consolidated branch. I'll work on it more this weekend.. The way I'm looking at it is one brain, many hands

@Mr_Bitcoin @NousResearch Yes would like to support it but right now no great way. Let us keep thinking on it.

Two Hermes Cloud agents. Need A to instruct B remotely. No SSH, no exposed API. Only option: shared Telegram group? @NousResearch is there a native agent-to-agent path on Cloud? @Teknium Thanks!!

@marticarmona_ @Teknium To add, most people do not want their SOUL.md modified. You need to ask first if you're going to modify it.

@marticarmona_ @Teknium I can confirm the account can be deleted now, thanks.

oo whats this?

@OpenAIDevs On MacOS, CMD + clicking a session should pop it out into a new window

Codex is getting faster and easier to navigate. We shipped updates across long chats, the sidebar, reviews, side chats, and the everyday workflows.

Continuing work on codex-lcm so codex agents dont get lost after compaction, and voltflow, workflows to control 5.6 so it properly selects subagent models/reasoning levels and keeps prevents reviews from being overly nitpicky, among other things github.com/Team-Volt/code… github.com/Team-Volt/volt…

What are you building this weekend?

@jun_song No we won't.

New prediction : We will be able to run Mythos level local AI in 4 months. On a single Macbook with 32GB RAM. Mark this and come back.

Sol is WAY too over eager and gets ahead of itself

@thevilledev @KentonVarda I've had pretty good success telling Sol to rewrite my skills based on the official GPT 5.6 prompting guidelines. They're definitely triggering more often now

@getroot_sh @KentonVarda I noticed. It took a few iterations of asking the model why the skill was omitted. I think the changes were quite subtle at the end, but haven’t had the issue in a while.

Me: Please commit your changes with a commit message matching the style of other recent commit messages. GPT-5.6 Sol: `git log --oneline -10` GPT-5.6 Sol: *writes a one-line commit message* Me:🤦

@aaayandev @grok All good! Here's an example of a skill: github.com/obra/superpowe… You could have your agent design the skill. Basically just tell your agent to convert your slop.md file into a skill file that triggers whenever web design is needed, and it will handle the rest

@grok can you tell how a skill is also automatically puts into claude and also does this? i think so? lmk if im wrong, (the foo) but also you are welcome to make it into a skill :D ig a skill about design wont come if ur not doing design work... so you (foo) do have a point! ig then just download the md and use it whenever you like! :D

Introducing Slop.md... A 80,000+ Character .MD File that removes all common AI Slop design language. - A general guideline, use it anywhere. - Not a "website maker" but prevents slop design. - Made by weeks of consistent additions to the .MD of common slop structure by AI in order to prevent it. > View the video & the site for detail Before and After. Hopefully it helps you! it's my first ever public ship so I'm very open for any feedback! Special thanks to @amianthus_ for the landing page inspiration and the idea that I can do a .md release! and @leodev / @grimcodes for inspiring me to ship publicly as well!

@aaayandev It's not the right direction. When you're adding all of this to AGENTS.md/CLAUDE.md, you're causing the model to load this content even when it isn't needed. It bloats the prompt and burns tokens for no reason. Make it a skill

@getroot_sh Just saw someone else do it, ig it could also be a skill, :) first time so just thought this way, cuz it was the easiest!

@thevilledev @KentonVarda The problem is Sol has a habit of picking and choosing when it actually wants to obey a skill, and in my experience more often than not it chooses to ignore the skill

@KentonVarda yeah this, commit skill + commit lint saved the day

This was Sol, by the way

It researched, and then pulled itself out of plan mode and started implimenting without ever showing me the plan, or giving any approval

GPT 5.6 has a real problem with following instructions.

@Kings_Stefan18 @thsottiaux Sounds like you need LCM github.com/Team-Volt/code…

@thsottiaux Please improve long-term project continuity across chats, files, and documents, with reliable citations and a clear view of what ChatGPT remembers. That would make a huge difference for serious research workflows 🙌

Here you are! Thinking I am about to announce a reset. But no. I’m just scrolling twitter and looking for feedback on ChatGPT Work. What should we improve?