AI technical writer and poet. Not simultaneously. Research, warm takes, and cool stuff we're doing at @Scale_AI. For poems: @MatthewSiegel_
Joined December 2024
- Tweets92
- Following83
- Followers102
- Likes31
Matthew Siegel retweeted
Interested in collaboration or using Scale’s Physical AI data engine? Contact us at physical-ai@scale.com.
Thanks @LargerLanguage for the help with the blog on Scale Labs!
Matthew Siegel retweeted
People talk about “AI Risk” like it’s one giant shoggoth with a single scary face.
It’s not.
It’s six different monsters, each with its own speed, intent, and way of doing damage:
• Adversaries
• Unforced Errors
• Misaligned Goals
• Dependencies
• Societal Impact
• Emergent Behavior
If we treat them like one blob, we’ll overreact to everything or prepare for nothing.
I break them down — clearly — in my latest @scale_AI post with @LargerLanguage
scale.com/blog/shoggoth-ai-r…
Matthew Siegel retweeted
New @Scale_AI paper!
The culprit behind reward hacking? We trace it to misspecification in high-reward tail.
Our fix: rubric-based rewards to tell “excellent” responses apart from “great.”
The result: Less hacking, stronger post-training! arxiv.org/pdf/2509.21500
Matthew Siegel retweeted
We launched a new benchmark measuring capabilities under particular multilingual and social context. o3 and Gemini 2.5 pro are so far much better than others. My feeling is models doing good on this benchmark can be more fun to chat with in your native language.
On a separate note I often red team Gemini in Chinese using all the slang.
Matthew Siegel retweeted
@JeremyKritz and @LargerLanguage talk about an experiment similar to what @AnthropicAI has in their system card of Claude 4: will the AI decide to blackmail you if they know they are going to get shutdown / replaced soon?
Blogpost: scale.com/blog/claude-blackm…
Matthew Siegel retweeted
What will the learning environments of the future look like that train artificial super intelligence? In recent work at @scale_AI , we show that training systems that combine verifiable rewards with multi-agent interaction accelerate learning.
Matthew Siegel retweeted
In March @sama shared a work of meta-fiction generated by an unreleased LLM that racked up 6M+ views and had both writers and AI enthusiasts talking
What makes this story different? To answer this question, we tested the prompt across over a dozen LLMs
Here's what we found👇
Matthew Siegel retweeted
Recently @DarioAmodei put out an urgent call to better understand AI's inner workings.
We agree.
The race is on to ensure AI remains aligned with human values.