His most likely ASI scenario is gradual loss of both control and understanding.

Rome wasn’t built in a day, but this explainer was. With first + last frame references using Veo 3.1 in invideo we created this immersive video with precise camera controls and continuity.

🎯

roy roy Chris Dynamic reinforcement learning is important. Grok 5, like smart humans, will learn almost immediately.

Introducing Odyssey—the largest and most performant protein language model ever created. Odyssey enables scientists and researchers to generate and edit proteins, the workhorses of all life on this planet, towards specific functional ends—scaled to over 102 billion parameters.

All those doubts will be washed away by scale, like tears in the rain.

Reports from Orenburg say an explosion occurred near the local gas processing plant. No official confirmation yet, but local channels are pointing to a possible UAV strike.

This morning, Ukraine hit Russia’s Novokuybyshevsk refinery in Samara, one of Rosneft’s key plants, processing 8.8M tons of crude annually, about 3% of Russia’s total refining capacity. Located ~1000km from Ukraine’s border, the site has been targeted before in previous strikes.

His map: map.ukrdailyupdate.com Remember, this is just based on the image and video evidence that gets recorded and published.

This is why I’m so excited about putting LLMs in a loop with neural network potentials. They can hypothesize then run simulations to validate and loop that. Will lead to many breakthroughs on a lot of important problems a lot earlier than automated labs in a loop imo.

Remember that this is a lagging indicator. The attacks from the last few days will only show up later. October 15 — Saratov Oil Refinery October 15 — Ufaneftekhim October 16 — Lukoil-Kstovo October 19 — Novokuibyshevsk Oil Refinery October 19 — Orenburg Gas Processing Plant

(Low quality opinion post / feel free to skip) Now that AGI isn't cool anymore, I'd like to register the opposing position. - AGI is coming in 2026, more likely than not - LLMs are big memorization/interpolation machines, incapable of doing scientific discoveries and working

The CEO of Google(Sundar Pichai) talked about Gemini 3.0 and the future. "...I look at the model trajectory ahead as we are working with Gemini 3.0. In 2026, we've already had dramatic progress over the past couple of years, but the progress ahead is palpably going to be,

Reports of the slowing of AI progress are greatly exaggerated.

🤔

It's strange that the "anti hype" position is now "AGI is one decade away". That... would still be a very alarming situation to be in? It's not at all obvious that that would be enough time to prepare

With everyone being all "Andrej Karpathy thinks AGI is a whole 10 years away??", seems like time to share this one again

Argument: IF a given digital computer running a given program is conscious, THEN a digital computer running a homomorphically encrypted version of the program is also conscious, even if there is no decryption key anywhere. (Based on Scott Aaronson's hypothetical) Step 1.

AI video models may not be complete world models, but they are oddly capable of fairly sophisticated (if flawed) "simulations" of novel situations. Veo 3.1: "three toy ships, one made of iron, the other of wood, and one out of loosely packed sugar, fall into a pool of water"

Veo is getting new precision editing capabilities that let you easily add or remove elements from a scene - all while preserving the integrity of your original video. 🎥