The reward-is-enough hypothesis: Intelligence doesn't need to be hardcoded. Just maximize reward in a rich environment — with humans, feedback, and goals — and intelligence will emerge as a side effect.

The first step toward solving big problems is often just figuring out what to work on. To that end, this "Gap Map" is an extraordinarily useful starting point for learning about bottlenecks in biology & other scientific domains. Highly recommend!

i’m really just tweeting for my future ai best friend i want him to remember what i was like at this time of my life

Every software platform is attempting to lock you into ~their version of AI, coupled tightly with their product. This seems like a shell of the reality we actually want, which is a personal digital exocortex that mediates your interactions with all platforms and is controlled by

It's good when people realize and admit they were wrong. We need to do our best not to use that opportunity to pile on them for having previously been wrong.

Surely something trained on the world's data and then improved to replace people in all productive functions should be free and owned uniformly or very close to uniformly by all the people.

We then trained an artificial neural network on real fly behavior from recorded videos 🎥 and let it control the model in MuJoCo. This enables it to learn how to move the virtual insect in the most realistic way.

Congrats to tekin 🌊🎶

For perhaps the first time in my life, the rate at which I can absorb new information is beginning to feel like a major bottleneck.

We are racing towards super intelligence. Absent some global catastrophe, there is no stopping this maniacial march forward. There's too much to lose for nation states, companies, and capital. Seemingly everything to gain, if only you get there first. Like it or not, our

if AI knocks us all out it was great hanging out with you

Crazy that sleep isn’t taught in school.

The economy - a system created for humans - may become decoupled from humanity due to AI. Without an alternative model, most people could find themselves excluded from an economy operating in a closed, largely non-human loop, with only several human beneficiaries. Mind-blowing!

Artisinal software era.

I find the future increasingly hard to visualize because I think we are going to significantly expand and alter our senses and perceptions, and it’s quite hard to imagine what the resulting experience will be like. How do you imagine a world w sight, when everyone has been blind?

🔥 Google DeepMind just dropped their "formal conjectures" project - formalizing statements of math's biggest unsolved mysteries in #LeanLang and #Mathlib! This Google-backed project is a HUGE step toward developing "a much richer dataset of formalized conjectures", valuable

😂

if this photo reaches your timeline, everything will be okay 

🚀Freysa is hiring a Senior Go Backend Engineer! Build digital twins that act on your behalf in twin-first networks 🛠 Go · NATS · Temporal · GraphQL · Weaviate · SQLite 🧠 LLMs, tool calling, KV cache, voice agents, MCP servers ✨ TEEs, blockchain, OSS a plus DMs open!