What if you could not only watch a generated video, but explore it too? 🌐 Genie 3 is our groundbreaking world model that creates interactive, playable environments from a single text prompt. From photorealistic landscapes to fantasy realms, the possibilities are endless. 🧵

Genie 3 feels like a watershed moment for world models 🌐: we can now generate multi-minute, real-time interactive simulations of any imaginable world. This could be the key missing piece for embodied AGI… and it can also create beautiful beaches with my dog, playable real time

what a time we live in

This is legitimately the craziest thing I have ever seen in my entire life Google deepmind is the best ai lab on planet earth right now

I'd like to formally apologize for making every button in Tailwind UI `bg-indigo-500` five years ago, leading to every AI generated UI on earth also being indigo.

When startup founders go to therapy:

such a terrible take

« The outer loop Era » It’s very hard to predict the future even a few months out in terms of models capabilities. End of 2024, word on the street was that pre-training was saturated (which is mostly true since model sizes have stopped increasing dramatically for years now)

Introducing DeepSeek-V3.1: our first step toward the agent era! 🚀 🧠 Hybrid inference: Think & Non-Think — one model, two modes ⚡️ Faster thinking: DeepSeek-V3.1-Think reaches answers in less time vs. DeepSeek-R1-0528 🛠️ Stronger agent skills: Post-training boosts tool use and

The Proton VPN servers for Afghanistan running at over 80% load was not something that we expected to see when we set them up last year.

I was talking with my bro about how he uses Claude Code, and realized he was operating the agents without a radar view. That's just silly

Si l'enfant d'ouvrier ne pense pas à HEC, ce n'est pas parce qu'il se sent illégitime, c'est parce qu'il ne connaît littéralement pas son l'existence. J'ai entendu parler de Sciences Po pour la première fois en terminale quand 4-5 mecs du lycée ont participé au concours.

1. Human thought is non-linear 2. Interfaces that augment human intelligence should preserve the same structure 3. Linear chat does not Therefore, the conversations must BRANCH. QED

Both DSPy and (especially) GEPA are currently severely under hyped in the AI context engineering world

We’re open-sourcing everything. All the projects built during the Nano Banana Hackathon are now released as open-source. Since AI Studio supports remixing, you can easily adapt and reuse them. They may not have been huge projects, but it was a great opportunity to truly

i swear Chinese AI labs are just having fun turning LLM finetuning into a game of pathfinding

EOY will be so cracked for me & my team can't wait anymore /o/

This is an insanely large world created using our 3D world generation model. It blew my mind!

> be Google > understand how important machine learning and deep learning is and create Google Brain in 2011 when everyone in the industry is screaming "big data". > acquire hinton's company called DNNresearch in 2013, which is basically Alex Krizhevsky and Ilya Sutskever and

Keep thinking.