I often see a lot of people with outdated understanding of modern LLMs.

This is probably the best interpretability research to date, by the leading interpretability research team.

It's worth a read if you want a peek behind the curtain on modern models.

I've been saying this for about a year since seeing the Othello GPT research, but it's nice to see more minds changing as the research builds up.

Edit: Because people aren't actually reading and just commenting based on the headline, a relevant part of the article:

New research may have intimations of an answer. A theory developed by Sanjeev Arora of Princeton University and Anirudh Goyal, a research scientist at Google DeepMind, suggests that the largest of today’s LLMs are not stochastic parrots. The authors argue that as these models get bigger and are trained on more data, they improve on individual language-related abilities and also develop new ones by combining skills in a manner that hints at understanding — combinations that were unlikely to exist in the training data.

This theoretical approach, which provides a mathematically provable argument for how and why an LLM can develop so many abilities, has convinced experts like Hinton, and others. And when Arora and his team tested some of its predictions, they found that these models behaved almost exactly as expected. From all accounts, they’ve made a strong case that the largest LLMs are not just parroting what they’ve seen before.

“[They] cannot be just mimicking what has been seen in the training data,” said Sébastien Bubeck, a mathematician and computer scientist at Microsoft Research who was not part of the work. “That’s the basic insight.”

I've been saying this for about a year, since seeing the Othello GPT research, but it's great to see more minds changing on the subject.

I'd been predicting this would happen a few months ago with friends and old colleagues (you can have a smart AI or a conservative AI but not both), but it's so much funnier than I thought it would be when it finally arrived.

Pretty cool thinking and promising early results.

I've suspected for a few years now that optoelectronics is where this is all headed. It's exciting to watch as important foundations are set on that path, and this was one of them.

I've had my eyes on optoelectronics as the future hardware foundation for ML compute (add not just interconnect) for a few years now, and it's exciting to watch the leaps and bounds occurring at such a rapid pace.

The Minoan style headbands from Egypt during the 18th dynasty is particularly interesting.