grohan

spot on :)

I quite enjoyed Tony Hoare's essay - thanks for sharing. A lot of what he wrote has stood the test of time!

Quite the load-bearing phrase!

how are people currently evaluating between the different spec writing frameworks out there?

"load-bearing" - lol, nice easter egg. no one likes the AI-isms in writing!

Pretty good point - I would actually argue that the imperative nature of prompting takes some of the blame here too. Your initial goal gets "steered" by the agent system and follow-up messages take precedence.

Here's a reverse engineer of the Claude token counter API as well as the model's vocabulary. Fast, offline, 96% accurate.

Bellard has trained various models, so it may not be the specific 169M parameter LLM, but his Transformer-based `nncp` is indeed #1 on the "Large Text Compression Benchmark" [1], which correctly accounts for both the total size of compressed enwik9 + decompresser size (zipped).

There is no unfair advantage here. This was also achieved in the 2019-2021 period; it feels safe to say that Bellard could have likely pushed the frontier far further with modern compute/techniques.

[1] https://www.mattmahoney.net/dc/text.html

They appear to have Python bindings which seems reasonable from an API / usability perspective? https://github.com/deepseek-ai/smallpond

In terms of fast FUSE - also my first question, appears to be`io_uring` + FUSE :)

https://github.com/deepseek-ai/3FS/blob/main/src/lib/api/Usr...

Impressive numbers. Does anyone have any read or anecdotes on how much a small/mid/large company loses from low quality of software/bad practices (or conversely profits from the opposite)?

Seems like a challenging metric to measure, but always been curious on what the numbers look like.