Amazing that Jürgen Schmidhuber gave this talk back in 2012, months before AlexNet paper was published. In 2012, many things he discussed, people just considered to be funny and a joke, but the same talk now would be considered at the center of AI debate and controversy. Full talk:

"Share your time, resources, knowledge" to show you care.

You did indeed had a great life, thank you for endless Dilbert adventures. RIP

We're just getting started, lots of features, improvements to make but 200k is not shabby for a little startup with no backing! THANK YOU!

Super excited for next week! 3 major ships

Such a cool update! Downloading now!

Who'd have thought that in 2026 I'd be sort of in love with Ghostty, has been my daily ADE for the past 4 months. Thank you Mitchell Hashimoto for taking the time between flights!

Gastown + autohand code cli as runtime. Steve Yegge You made an orchestrator that actually goes beyond 20 to 30 agents.

Holy smokes, I finally peeked at the stats for Autohand code CLI today, turns out you devs are actually using it! I even slapped together a skill for Cloudflare showing how to wrangle that bad boy with autohand code. And guess what? We just blew past 570k devices

I hope my friends in Minneapolis are safe.

Today we’re releasing the International AI Safety Report 2026: the most comprehensive evidence-based assessment of AI capabilities, emerging risks, and safety measures to date. 🧵 (1/17)

We're adding more client options for an agentic world for SWE, give it a spin and tell me what you think. Feature parity with major labs but adaptive to any LLM model with cool features.

Mind over matter.

I believe in open science and transparency, I hope to see you in San Jose this year to talk about this at ACM Conference on AI and Agentic Systems May this year co-organised by Matei Zaharia

Aos seguidores Brasileiros, Chegamos ao Brasil.

Thank you all for your kind words and suggestions on my paper arxiv.org/html/2602.0707… I’m cleaning up some of the code and I will publish on Autohand GitHub repo. We are just at the beginning and there are other ideas I’d love to explore.

Imagine the overall impact this could have on CAPEX, you don’t have to wait, Autohand has MoA that helps your CUDA Kernel whistle the fan.

Piotr discovered something worrying: if you give an LLM a list of tools it's allowed to call, it might decide to also call a tool you didn't provide! Impacts all major US providers except OpenAI. Be sure to check LLM tool call requests! (Lisette/Claudette check automatically)

I'm tempted to talk about our commander v2 for coding agents, but I will hold a bit more until I get the green light from Apple Store. Don't swarm, collaborate!