🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

In a new paper, we examine recent claims that AI systems have been observed ‘scheming’, or making strategic attempts to mislead humans. We argue that to test these claims properly, more rigorous methods are needed.

Do you have a PhD (or equivalent) or will have one in the coming months (i.e. 2-3 months away from graduating)? Do you want to help build open-ended agents that help humans do humans things better, rather than replace them? We're hiring 1-2 Research Scientists! Check the 🧵👇

Does vision training change how language is represented and used in meaningful ways?🤔 The answer is a nuanced yes! Comparing VLM-LM minimal pairs, we find that while the taxonomic organization of the lexicon is similar, VLMs are better at _deploying_ this knowledge. [1/9]

Correct progression seems to be: formal systems+ symbol manipulation -> learn to use human defined formal sys as tools-> tools are learned/ created by learning system -> don't need any more explicit tools Hopeful that we'll eventually make our way up this ladder for most domains

We are hiring on the Veo team!📽️ Some people asked me about this at #ICML2025. If that's you, I will have told you to check deepmind.google/careers/ regularly. 👀It's just been updated: Europe (London, Zurich) job-boards.greenhouse.io/deepmind/jobs/… US (Mountain View) job-boards.greenhouse.io/deepmind/jobs/…

I’m building a new team at Google DeepMind to work on Open-Ended Discovery! We’re looking for strong Research Scientists and Research Engineers to help us push the frontier of autonomously discovering novel artifacts such as new knowledge, capabilities, or algorithms, in an

New preprint picks up Wai Keen Vong's work on training neural nets from a single child's input, now replicated with two more SAYCam children and three architectures! The similarities are stronger than the differences. arxiv.org/abs/2507.14749

Cool work by Andrea de Varda and Ferdinando on the close alignment between human reaction times and reasoning trace lengths across a wide range of reasoning tasks!