Stephen McAleer (@McaleerStephen) / X

Stephen McAleer

792 posts

Stephen McAleer

@McaleerStephen

AI researcher at Anthropic

Joined July 2014

Pinned
Stephen McAleer
@McaleerStephen
Nov 22, 2023
"Toward General Virtual Agents" I recently gave a talk at MIT. I argued that we should use tools from reinforcement learning and search to improve the capability and alignment of LLM agents. Slides: drive.google.com/file/d/1kDvmrm… Video:
00:00
164K
Stephen McAleer
@McaleerStephen
Nov 23, 2023
We invented Q* first Glad openai is building on top of our idea
758K
Stephen McAleer
@McaleerStephen
Sep 29, 2025
Having done RL at OpenAI and Anthropic, here's what I can say about GRPO:
518K
Stephen McAleer
@McaleerStephen
Jan 4, 2025
I kinda miss doing AI research back when we didn't know how to create superintelligence.
172K
Stephen McAleer
@McaleerStephen
Jan 15, 2025
Honest question: how are we supposed to control a scheming superintelligence? Even with a perfect monitor won't it just convince us to let it out of the sandbox?
658K
Stephen McAleer
@McaleerStephen
Jan 9, 2025
It's hard to convey my views without sounding like an AI grifter 😅 I will say this: many researchers at frontier labs are taking the prospect of short timelines very seriously, and virtually nobody outside the labs is talking enough about the safety implications.
112K
Stephen McAleer
@McaleerStephen
Sep 11, 2025
Today was my last day at OpenAI. It's tough to leave given all the great safety work going on but I decided to deepen my focus on scalable oversight and frontier AI safety. Extremely grateful for my time at such a legendary company!
155K
Stephen McAleer
@McaleerStephen
Jan 21, 2025
Compute is not a bottleneck
OpenAI
@OpenAI
Jan 21, 2025
Announcing The Stargate Project The Stargate Project is a new company which intends to invest $500 billion over the next four years building new AI infrastructure for OpenAI in the United States. We will begin deploying $100 billion immediately. This infrastructure will secure
102K
Stephen McAleer
@McaleerStephen
Sep 23, 2025
I joined Anthropic! I've wanted to work with @EthanJPerez ever since he safety pilled me a few years ago. It's been amazing working with him, Jan, Jared, and everyone so far. I can't remember a time in my life being this excited to wake up and get to work every day 🙏
72K
Stephen McAleer
@McaleerStephen
Dec 30, 2024
Not enough people are thinking about fully-automated AI R&D.
73K
Stephen McAleer
@McaleerStephen
Jan 12, 2025
Controlling superintelligence is a short-term research agenda.
174K
Stephen McAleer
@McaleerStephen
Feb 20, 2025
The smarter AI becomes, the harder it is to make it do what we want.
195K
Stephen McAleer
@McaleerStephen
Dec 31, 2024
In AlphaGo, just training on human data only got to amateur level. But once they cracked high-compute RL, then superintelligence was inevitable.
67K
Stephen McAleer
@McaleerStephen
Jan 27, 2025
The real takeaway from DeepSeek is that with reasoning models you can achieve great performance with a small amount of compute. Now imagine what you can do with a large amount of compute.
38K