Could the (not so perfect but technically simple) solution be to transform the style of content under each tag to the correct expected style for the tag, via a smaller or purpose-built LLM, before the data stream is fed into the main LLM? Perhaps the two LLMs can be co-trained to keep the overall quality of the output stable while role confusion is minimized.
Very interesting. How does this approach work for complex agentic workflows where the LLM is expected to orchestrate across multiple tools (such as when using MCP)? Or is this mainly for simple cases like the ones presented in the blog post?
chessprogramming.org is a treasure trove of knowledge on building chess engines. With its help, I wrote one in C++ a few years ago that got quite good (2100+ rating on FICS but that's nowhere close to the likes of Stockfish). In fact, writing a reasonably strong chess engine is straightforward (and incredibly fun) but at the top end of strength, there's immense depth, and after a point making improvements gets increasingly resource intensive (tuning params, running experiments to verify strength gain all takes a lot of compute).
Chess programming is also extremely addictive. On forums like talkchess.com, you see folks hanging out who have been doing it for decades (most of them are also super helpful to newbies).
Just completed part 1 (used Python instead of Java to spice it up a bit) and it was a great experience! Looking forward to working through part 2, but it will be a while before I can get around to it. I hope more books follow this engaging writing pattern (when it makes sense). Kudos to Bob Nystrom for this great book.
And yet there are hardly any software written in Haskell or Ocaml that are widely used or have any notable positive impact on the modern digital world, compared to those written in languages like Go, C++ or Python.
edit: ps: Big fan of OCaml, but have since moved on to Go and Python for getting things done in the real world.
Bob works for company A which is a vendor for companies B, C and D. Most of company A's revenue stream is dependent on how well B, C and D are doing. Bob has insider knowledge of Company A's finances using which he trades stocks of B, C and D.
But Trump [1] also says the non-white congresswomen known as "the squad" should go back to where they came from. Even riled up his supporters [2] to chant "send her back".
You know "go back to where you came from" is a common racial insult hurled at non-white immigrants decades or generations after they naturalize in the USA?
Tell me this: When Trump riled up his supporters to chant "send them back" when referring to the American congresswomen known as "the squad", was it racist or not?
"Go back to where you came from" is a common racial insult non-white immigrants endure even decades or generations after they naturalize.
"The most devastating argument against the Copernican universe was the star size problem. When we look at a star in the sky, it appears to have a small, fixed width. Knowing this width and the distance to the star, simple geometry reveals how big the star is (right). In geocentric models of the universe, the stars lie just beyond the planets, implying that star sizes are comparable to that of the sun (below). But Copernicus's heliocentric theory demands that the stars be extremely far away. This in turn implies that they should be absurdly large—hundreds of times bigger than the sun (bottom). Copernicans could not explain away the anomalous data without appeals to divine intervention. In reality, the stars are far away, but their apparent width is an illusion, an artifact of the way light behaves as it enters a pupil or telescope—behavior that scientists would not understand for another 200 years."