The Next Generation of AI Models
This blog article reflects thoughts on what the author imagines the next generation of AI models will be like. What we need to progress to more reliable and efficient AI is a paradigm shift in model design, one which is fundamentally logical rather than statistical.
Lego sets vs Sandcastles
I decided to publish this article in its draft state in case it might be useful to someone regardless. It’s roughly edited, but I have no idea if I’m going to have the time or motivation to finish it later. If you found it interesting, please let me know your own thoughts in the comments. Thanks!
There’s been a lot of talk lately about “reasoning models.” As I understand them, a reasoning model is simply a language model with some clever tricks to get it to review and refine its output before returning it to a user. But under the hood, the technology hasn’t fundamentally changed.
