LLM Metacognition: Shared and Shallow? Across 19 frontier models, metacognitive confidence on question and answer tasks tracks a shared difficulty heuristic with only a weak relationship to actual performance.

My team at Pareto.ai has been trying to understand LLM's potential for metacognition. As it turns out, they don't tend to have it beyond what they appear to have memorized awareness of from pre-training.

Check out the details here: pareto.ai/blog/llm-met...

Excited to see our work coming out + @joshnguyen.bsky.social & @duncanjwatts.bsky.social

After establishing a means to study common sense in humans (and finding it rather limited — common sense is not so common) in a prior paper, we wondered if the same challenge faced language models.

It does!

The Resonant Computing Manifesto Technology should bring out the best in humanity, not the worst—a manifesto for resonant computing built on five principles that reject hyper-scale extraction for human flourishing.

What if technology didn’t feel so… hollow?

Some friends and I just released a manifesto about a world where tech leaves us feeling nourished (along with an evolving list of theses about how we can build it)

resonantcomputing.org

Looking forward to reading this new PNAS paper "A framework for quantifying individual and collective common sense"

doi.org/10.1073/pnas...

Reminds of something I've always wondered...if there is a quantitative way to think about how Sherlock Holmes patches together his observations

Common sense is much more complicated than most people realize. Fascinating new research in PNAS provides the first empirical measure of the concept, showing that it varies considerably across groups, and there is much less collective common sense than most realize: www.pnas.org/doi/abs/10.1...

How common is commonsense? Turns out that except for statements about physical reality... not that common.

https://www.pnas.org/doi/10.1073/pnas.2309535121