Rosy-faced Lovebirds on Saguaro cactus 🌵 https://t.co/hIrjuaFbfp

A shepherd in abandoned Sacred Grove, Italy, ca. 1952 - by Herbert List (1903 – 1975), German https://t.co/O6DXE1EXAg

@mov_axbx If you’re a professional developer, you’re not “vibe coding”, you’re using AI tools. Reviewing the output is part of the job. If you’re not checking quality, that’s not AI, it’s just being unprofessional.

@traits_reality I can relate, but there’s a narrow band of value between AI slop and perfection. The trick is riding that wave without falling to either side.

@ProofofMaro Exactly. Some people can get away with that for their own use. But if you’re a professional whose work depends on AI, understanding how it actually works is the difference between safety and liability, and potentially much worse. Don’t wait for reality to catch up with you.

scrolling through this site feels like watching a movie → https://t.co/a41OXXNDFB https://t.co/jvN7Z1LVvM

this is beautifully done

@Srirachachau the competition for this podcast thing is really heating up

imagine what could’ve happened if tyra banks was born earlier and they recruited her to consult on the mk ultra experiments…. what they could’ve discovered

This is a real gold mine! 😎

@polynoamial It did not trigger my mental AI detector

"If Trump can depose any leader he likes, what is to stop Russia or China acting similarly against their foes? (Russia, of course, already has tried, in its war on Ukraine.) Is that the world we want to live in?" https://t.co/lvjGKwuJmv

I can't help thinking that the AI community moved the bulk of the resource and efforts on getting as much as possible from the GPT architecture through scaling, prompting, and agent-swarming, even though the said architecture is missing key elements. 1/2

@francoisfleuret Obviously

today we're launching struct - the on-call agent every team deserves. and we're making it free. no sales call. 5 min setup. 80% faster on-call. https://t.co/vuX7iATL2l

Super excited to present MetaClaw 🦞 — our self-evolving agent that learns just by talking to it. 💬 Conversations → training trajectories 🧠 Failures → new reusable skills ⚡ Weights update live (no interruption) ☁️ Powered 100% via API — no GPU cluster needed Try it here 👇 https://t.co/p5yqdWmwzR Huge kudos to the team @richardxp888, Jianwen Chen, @Xinyu2ML, @lillianwei423, @StephenQS0710, Zeyu Zheng, @HuaxiuYaoML

Wideframe just launched on @ycombinator! Do hours of video work in minutes with @wideframeai. Check out our launch: https://t.co/rPU7Pqu7vE https://t.co/r2yFYFgIZz

Wideframe just launched on @ycombinator! Do hours of video work in minutes with @wideframeai. Check out our launch: https://t.co/rPU7Pqu7vE https://t.co/r2yFYFgIZz

@Docker @NVIDIAGTC See you there!

@Yuchenj_UW Codex is a know issue :( It basically don't work with autoresearch sadly, in the way it's set up atm: https://t.co/4xjdf4wQyX I pung a friend at OpenAI to see if something can be done, e.g. need a /loop equivalent or something like that. More generally, I really dislike the -p + ralph loop pattern of running agents "headless". I want nice, interactive sessions running in tmux so that I can see what they are doing, pitch in, etc.

Andrew Ng just solved one of the biggest problem with Agents. He released Context Hub, a CLI tool to fetch live API documentation. One command. The agent gets exactly what it needs before writing a single line of code. Agents trained months ago are flying blind. They invent parameter names. They call functions that no longer exist. They confidently write code against a spec that changed in the last release. > No more hallucinated parameters > Docs pulled fresh before each call > Agents log useful discoveries > Notes persist between sessions The agent runs a CLI command before touching the code. Instead of relying on stale data, it reads the actual spec. Fast-moving APIs used to mean maintaining a doc dump in every prompt. Now the agent does that work itself. When it finds a workaround, it saves a note for next time.

https://t.co/e7xglk9wXC

Wait, did Agent mode just disappear from ChatGPT?

🚨 New: Integrating Harbor (@harborframework) for end-to-end Computer-Use evaluation(for Windows and Linux) at scale with @thinkymachines' Tinker, OSWorld, @daytonaio, and bare-metal servers. We just added support for Computer Use, @tinkerapi, and OSWorld to Harbor - a framework for evaluating agents and generating RL training data by running large-scale rollouts across parallel sandboxed environments and collecting trajectories for SFT and RL. Repo and blogpost below 👇