grohan
Joined 6 years ago
Karma
42
Submissions
19
Member since
Nov 16
Recent submissions
Comment · 1 month ago
I quite enjoyed Tony Hoare's essay - thanks for sharing. A lot of what he wrote has stood the test of time!
Comment · 1 month ago
Quite the load-bearing phrase!
Comment · 1 month ago
how are people currently evaluating between the different spec writing frameworks out there?
Comment · 1 month ago
"load-bearing" - lol, nice easter egg. no one likes the AI-isms in writing!
Comment · 1 month ago
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.
Comment · 8 months ago
Here's a reverse engineer of the Claude token counter API as well as the model's vocabulary. Fast, offline, 96% accurate.
Comment · 11 months ago
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.