shchegrikovich

I have another example - just a few people believed that you can apply 'a simple next token prediction algorithm' and achieve what we know as LLM. From my perspective, in the past few years, we've tried a lot of different approaches to improve LLM reasoning; some of them were good, others not so good. We need to keep trying and researching. 'Prolog + LLM' is not the answer to all questions, but it looks like a good step to move us forward.

The draft version of this blog was called - 'Renaissance of Programming Languages'. With so much hidden gems we are re-discovering the full potential of formal languages.

I agree; we can do much more with LLMs if we focus on declarative-style languages.

I discussed this a few weeks back. The idea is to take a Python dataset, as Python is the most popular language, and write a transpiler to Prolog with the help of llms. So, creating this synthetic dataset is not a huge problem.

I think another problem which we can try to solve with this approach is factual verification.