Ross Taylor

@rosstaylor90

Co-founder and CEO @GenReasoning. Previously lots of other things like: reasoning lead Meta AI, Llama 3/2, Galactica, Papers with Code.

following
followers

A call to allow AI pre-training on UK soil

A commenter argues UK laws should change to enable domestic AI pre-training, rather than restricting local expertise to testing.

1 Source, 6d ago, first seen 6d ago

Potential AI gains in hard-to-verify reasoning may be hard to spot

One post argues that possible conceptual or strategic insights are harder to assess than mathematical claims checked with formal proofs.

3 Sources, 12d ago, updated 12d ago

BackSearch gets an update for internal reinforcement-learning training

A BackSearch team member highlights the backdated arXiv index and says broader uses, such as a cached internet, are expected as the index expands.

1 Source, 33d ago, first seen 33d ago

MiMo-V2.6 reportedly scales reinforcement learning to about 2 billion tokens per step

In a September 16 update, the team behind MiMo-V2.6 said it was scaling compute, agent environments and grading, with plans to open-source details in stages.

38 Sources, 27d ago, updated 27d ago

Inherent Labs Introduces Faraday 27B AI Scientist

The model adds scientific intuition to coding agents for paper replication.

43 Sources, 67d ago, first seen 67d ago

Qwen Releases Open Weights for 27B Multimodal Model

The 27B model is now available on Ollama for agentic coding and professional tasks.

33 Sources, 67d ago, first seen 67d ago

Junyang Lin Founds Pragmatik Labs to Research AI Agents

Former Qwen lead Junyang Lin announces Shanghai startup Pragmatik Labs focused on next-generation agents.

15 Sources, 70d ago, first seen 70d ago

Meta Releases Weights for Muse Glimmer 30B Model

Mark Zuckerberg posted that Meta opened weights for a local model with another release planned soon.

178 Sources, 71d ago, first seen 71d ago

Featured in

A call to allow AI pre-training on UK soil

Potential AI gains in hard-to-verify reasoning may be hard to spot

BackSearch gets an update for internal reinforcement-learning training

MiMo-V2.6 reportedly scales reinforcement learning to about 2 billion tokens per step

Inherent Labs Introduces Faraday 27B AI Scientist

Qwen Releases Open Weights for 27B Multimodal Model

Junyang Lin Founds Pragmatik Labs to Research AI Agents

Meta Releases Weights for Muse Glimmer 30B Model

Featured in

A call to allow AI pre-training on UK soil

Potential AI gains in hard-to-verify reasoning may be hard to spot

BackSearch gets an update for internal reinforcement-learning training

MiMo-V2.6 reportedly scales reinforcement learning to about 2 billion tokens per step

Inherent Labs Introduces Faraday 27B AI Scientist

Qwen Releases Open Weights for 27B Multimodal Model

Junyang Lin Founds Pragmatik Labs to Research AI Agents

Meta Releases Weights for Muse Glimmer 30B Model