Thoughts on AI progress (Dec 2025) - by Dwarkesh Patel
substack.com · 2,375 words · saved by 1 readers
Why I'm moderately bearish in the short term, and explosively bullish in the long term
I’m confused why some people have short timelines and at the same time are bullish on the current scale up of reinforcement learning atop LLMs. If we’re actually close to a human-like learner, this whole approach of training on verifiable outcomes is doomed. Currently the labs are trying to bake in a bunch of skills into these models through “mid-training” - there’s an entire supply chain of companies building RL environments which teach the model how to navigate a web browser or use Excel to write financial models. Either these models will soon learn on the job in a self directed way -…
related reading
- Thoughts on AI progress (Dec 2025)dwarkesh.com
- Thoughts on AI progress (Dec 2025)substack.com
- Thoughts on AI progress (Dec 2025) - by Dwarkesh Pateldwarkesh.com
- Andrej Karpathy — AGI is still a decade awaydwarkesh.com
- Questions about the Future of AI - by Dwarkesh Pateldwarkesh.com
- AI in 2025: gestalt — LessWronglesswrong.com
- Why I don’t think AGI is right around the cornerdwarkesh.com
- Will scaling work?dwarkeshpatel.com
- The Extreme Inefficiency of RL for Frontier Models - Toby Ordtobyord.com
- Dario Amodei — "We are near the end of the exponential"dwarkesh.com
- Why do people disagree about when powerful AI will arrive?blog.bluedot.org
- AI Timelines — LessWronglesswrong.com