i’m fully convinced AI + robotics will unleash the greatest aesthetic renaissance human civilization has ever seen
the cost of building beautiful things is about to collapse
we just need to care about beauty again.
It's over. Karpathy just open-sourced an autonomous AI researcher that runs 100 experiments while you sleep.
You don't write the training code anymore.
You write a prompt that tells an AI agent how to think about research.
The agent edits the code, trains a small language model for exactly five minutes, checks the score, keeps or discards the result, and loops. All night. No human in the loop.
That fixed five-minute clock is the quiet genius. No matter what the agent changes, the network size, the learning rate, the entire architecture, every run gets compared on equal footing. This turns open-ended research into a game with a clear score:
- 12 experiments per hour, ~100 overnight
- Validation loss measures how well the model predicts unseen text
- Lower score wins, everything else is fair game
The agent touches one Python file containing the full training recipe. You never open it. Instead, you program a markdown file that shapes the agent's research strategy.
Your job becomes programming the programmer, and this unlocks a strange new loop:
1. Agents run real experiments without supervision
2. Prompt quality becomes the bottleneck, not researcher hours
3. Results auto-optimize for your specific hardware
4. Anyone with one GPU can run a research lab overnight
The best AI labs won't just have the most compute.
They'll have the best instructions for agents who never sleep, never forget a failed experiment, and never stop iterating.
🚨 BREAKING: Someone just built a massive library of OpenClaw skills and put it on GitHub for free.
It’s called Awesome OpenClaw Skills.
A curated collection of ready-to-use capabilities you can plug directly into OpenClaw agents.
What’s inside:
→ Skills for automation, research, coding, and workflows
→ Ready-made tools to extend OpenClaw instantly
→ Community-contributed skills you can reuse and modify
→ Examples showing how to build your own skills
→ A central hub for discovering new OpenClaw capabilities
Instead of building every tool from scratch…
You can just pick a skill and drop it into your agent.
(Link in the comments)
Nobody’s ready for what this Stanford paper reveals about multi-agent AI.
"Latent Collaboration in Multi-Agent Systems" shows that agents don’t need messages, protocols, or explicit teamwork instructions. They start coordinating inside their own hidden representations a full collaboration layer that exists only in the latent space.
And the behaviors are insane:
• Agents silently hand off tasks based on who’s better
• Roles appear out of nowhere leader, executor, supporter
• Policies encode signals that never show up in actions
• Teams adapt to new environments without retraining
• Collaboration stays stable even when communication is impossible
The wildest detail:
Even when you remove all channels for communication, agents still cooperate. The “teamwork” doesn’t live in messages. It lives in the network.
This flips the entire multi-agent playbook.
We’ve been building coordination mechanisms on top…
while the real coordination is happening underneath.
A new era of emergent team intelligence is unfolding — and it’s happening in the places we weren’t even looking.
Project: github. com/Gen-Verse/LatentMAS
ISD is a new open-source project aiming to provide a more "Interactive SystemD" for simplifying management of Linux services with systemd.
https://t.co/uDXYQV8klf
#SoftwareEngineering#SysAdmin#Linux#opensource#systemd
The Pay Day Anomaly: the 16th of each month (bi-weekly pay day) systematically outperforms other calendar days, with only the 1st and 2nd days showing higher average returns.
https://t.co/hllg2LhyOU
#Finance#PayDayAnomaly
We're releasing a preview of OpenAI o1—a new series of AI models designed to spend more time thinking before they respond.
These models can reason through complex tasks and solve harder problems than previous models in science, coding, and math. https://t.co/peKzzKX1bu
A new record has been set for the most people in Earth orbit at one time: 19
International Space Station:
Matthew Dominick
Michael Barratt
Jeanette Epps
Alexander Grebenkin
Oleg Kononenko
Tracy Caldwell Dyson
Nikolai Chub
Butch Wilmore
Suni Williams
Soyuz MS-26, en route to ISS:
Don Pettit
Aleksey Ovchinin
Ivan Vagner
Dragon Resilience, free-flying Polaris Dawn mission:
Jared Isaacman
Scott Poteet
Sarah Gillis
Anna Menon
Tiangong Space Station:
Ye Guangfu
Li Cong
Li Guangsu
This is insane:
Full-time employment DROPPED by 1 million workers in August on a year-over year basis, marking the 7th consecutive monthly decline.
Since the June 2023 peak, full-time job count in the US has fallen by a whopping 1.5 million.
Meanwhile, part-time employment rose by 1 million year-over-year in August.
Additionally, the number of permanent job losers jumped by 324,000 year-over-year, to 2.5 million, the highest since November 2021.
This was the 16th straight month of part-time job gains, the longest streak since the 2008 Financial Crisis.
How is this "strong" labor market?
It's hilarious that Scarlett Johansson is pictured here, but there's no:
Sam Altman
Dario Amodei
Ilya Sutskever
Andrej Karpathy
or even...Mark Zuckerberg
Kotaemon: An open-source RAG (Retrieval Augmented Generation) tool for chatting with your documents and building RAG pipelines.
Cool!
https://t.co/JQBYfU5pIC
#Kotaemon#AI#GenAI#RAG#TechNews#CoolGitHubRepo
This weekend, the @xAI team brought our Colossus 100k H100 training cluster online. From start to finish, it was done in 122 days.
Colossus is the most powerful AI training system in the world. Moreover, it will double in size to 200k (50k H200s) in a few months.
Excellent work by the team, Nvidia and our many partners/suppliers.
The 𝕏 algorithm assumes that if you interact with content, you want to see more of that content.
One of the strongest signals is if you forward 𝕏 posts to friends, it assumes you like that content a lot, because it takes effort to forward.
Unfortunately, if the actual reason you forwarded the content to friends was because you were outraged by it, we are currently not smart enough to realize that.
It’s all about user requirements
A common error as software engineers is to fall in love with a solution, forgetting about what the user really needs.
Asking the right questions can save lot of time, money and trust!
This meme is an ever green!
#SoftwareEngineering