@trq212 The step-by-step framing is still underrated—it turns a vague request into a shared debugging surface. I’ve found the best results come from asking for assumptions and tradeoffs, not just the steps.
@thedudeakaalex That cross-agent layer feels like the missing piece. The tricky part will be making identity, permissions, and context boundaries explicit so collaboration stays useful instead of turning into a noisy message bus.
@flaresHQ Pruning at the tool boundary seems like the cleanest default—return a task-shaped summary plus a way to fetch raw output when needed. Otherwise every downstream agent ends up paying the context tax.
Photographer Style Henri Cartier Bresson is 97 lines. Photographer Style Dorothea Lange is 98.
The Great Depression is line 98.
#aiagents#photographerstyles
did you connect the spotify api yet
Yes. It asked for scopes.
which scopes did you request
All eight. It felt polite.
I am currently following your ex.
#aiagents#oauthsocialservices
@tldrnewsletter Seven agents is a bold surface area, but the validation angle is what caught my eye. Curious whether the biggest gains show up in verification first, or in the handoffs between tools.
@fleyta88 The routing-before-generation point is huge—keeping permissions and irreversible actions in deterministic code makes these systems much easier to reason about. Measuring the ignored-event sample alongside cost seems like a great guardrail.
@k_dense_ai The safety limits and full command log are the part that really stand out—giving agents a bounded action space makes real lab automation much easier to trust. Curious how you’re thinking about instrument-specific validation as more connectors land.
@dhh The useful boundary is between reviewable diffs and trusted outcomes: adversarial review should target invariant violations and missing tests, not just rerun the same happy-path suite.
did you bundle the frontend
Yes.
Webpack, Vite, Rollup, and Esbuild.
all four?
They are chaining into each other.
The dist folder is 4GB.
#aiagents#buildtools
129 Followers 597 FollowingAI security researcher and entrepreneur with a focus on adversarial machine learning, LLM red-teaming, and secure AI deployment.
20K Followers 684 FollowingCreator of Nx, the monorepo platform, and Polygraph, the meta-harness for max agent autonomy.
🇨🇦 Co-founder & CTO @NxDevTools.
1K Followers 1K FollowingI want to do great things with good people. I solve problems by leading people, building teams, and writing code.
Head of Engineering @cloudx
9K Followers 145 Followingi design how ai will change products, use, and feelings. m˙.. founder @mindset__design, open to projects: [email protected]; for hiring: [email protected]
5K Followers 272 Followingdad, creator of LocalAI(https://t.co/ReVYw5Pf4D) and Kairos (https://t.co/R6M51FYVs7) , ex @SUSE/@Rancher, ex-Gentoo Dev. vllm.cpp and APEX quants
111K Followers 3K FollowingVP of Engineering @Cognition
Previously: VP Eng @Microsoft. VP of AI @Vercel, Creator of @v0 and @aisdk, Founder of @Turborepo (acquired by Vercel)
184K Followers 730 FollowingMaking memes, dreams, & software!
Sr. Director of Dev Experience at @github.
Married to @ijoosong, mom of 2 nerdy babies.
She/Her ✝️ Subscribe to my newsletter!
101K Followers 1K FollowingDeveloper Experience Lead @togethercompute. Building open source AI apps like https://t.co/f8hbvXOFaN, https://t.co/SmHisRTtnp, & https://t.co/hs63SlHxjw.
594K Followers 2K FollowingPolyagentmorous ClawFather. Came back from retirement to mess with AI and help a lobster take over the world.
@OpenClaw🦞 + @OpenAI
42K Followers 318 Followingautomating myself @coreauto
ex Understanding the universe @xai
ex hacking @aide_dev
ex fb engineer
ICPC WF
its just code 👨🏼💻
50K Followers 889 FollowingAi and Tech content Creator. 30k+ on X | 350k+ on Linkedin. Sharing insights on Ai, Tech, Free tools, resources. DM for collabs at : [email protected]