Hacker News: pizza

New comment by pizza in "Nvidia agrees to acquire Hugging Face for $13B"

pizza — Thu, 27 Aug 2026 16:47:29 +0000

They’ve provided insane value to the ai community. Imo is there really any obvious way to standardize model/preprocessing/inference/training/rl recipes across the space of frankenstein model family/architecture/modality combinations? It’s not even clear to me that the current frontier would be where it is today without HF doing what they did how they did it. Where else were people widely sharing datasets w that ease of reuse/redistribution, or sharing their retuned models etc? Would there even be as much interest/tinkerers today?

New comment by pizza in "Compression is prediction"

pizza — Wed, 12 Aug 2026 04:32:24 +0000

Compression is just counting. Probability, also, pretty much, just counting. For these reasons I think the role of information theory in describing the process of the development of reasoning and the gain of understanding has been overstated.

New comment by pizza in "OpenAI and Government of Malta partner to roll out ChatGPT Plus to all citizens"

pizza — Sat, 16 May 2026 21:05:23 +0000

you can say the same thing of the watts in a person too

New comment by pizza in "Compression & Decompression w/ FHE via Err Correcting Codes and Copy-and-Recurse"

pizza — Thu, 14 May 2026 19:31:03 +0000

Had to butcher the title slightly to get it under the limit- original:

Compression And Decompression Under FHE Using Error-Correcting Codes and Copy-And-Recurse

Compression & Decompression w/ FHE via Err Correcting Codes and Copy-and-Recurse

pizza — Thu, 14 May 2026 19:29:13 +0000

Article URL: https://eprint.iacr.org/2026/504

Comments URL: https://news.ycombinator.com/item?id=48140087

Points: 1

# Comments: 1

Learning Pseudorandom Numbers with Transformers

pizza — Sun, 03 May 2026 07:21:50 +0000

Article URL: https://arxiv.org/abs/2510.26792

Comments URL: https://news.ycombinator.com/item?id=47994307

Points: 11

# Comments: 3

Objects of Categories as Complex Numbers (2002)

pizza — Sat, 25 Apr 2026 22:29:03 +0000

Article URL: https://arxiv.org/abs/math/0212377

Comments URL: https://news.ycombinator.com/item?id=47905221

Points: 2

# Comments: 0

To Dissect a Mockingbird: A Graphical Notation for the Lambda Calculus (1996)

pizza — Sun, 19 Apr 2026 23:03:20 +0000

Article URL: https://dkeenan.com/Lambda/index.htm

Comments URL: https://news.ycombinator.com/item?id=47828492

Points: 11

# Comments: 1

New comment by pizza in "Measuring Claude 4.7's tokenizer costs"

pizza — Sat, 18 Apr 2026 06:42:16 +0000

For most tasks, at some future date, isn't there going to be some ambient baseline of capabilities you can get per $/tok, starting at ~0 for OSS models, such that eventually all tooling gets trivially transferable?

All elementary functions from a single binary operator

pizza — Mon, 13 Apr 2026 01:49:32 +0000

Article URL: https://arxiv.org/abs/2603.21852

Comments URL: https://news.ycombinator.com/item?id=47746610

Points: 858

# Comments: 298

New comment by pizza in "Entropy as a Measure of Surprise"

pizza — Sun, 12 Apr 2026 09:51:13 +0000

OP is correct; surprisal is outcome-dependent and entropy is distribution-dependent

- entropy is E_p[informativeness of measuring outcome x]

- take n outcomes, then a distribution over them lives on the simplex \delta ^ (n - 1). you can lift this to R^n via the log odds map p_k -> x_k = log p_k -- now x \in R^n can describe a histogram with n-1 degrees of freedom

- in log odds space, measurement is literally a linear functional from vector space of log probability onto the index of the outcome k.

- imo surprisal of some p(x) is best understood as "the length of a pointer", entropy "the rarity-weighted average length of a pointer", and collision entropy "how specific you would have to be to describe witnessing a specific outcome"

and in the same way, a single molecule of water, you might get by, calling dry

New comment by pizza in "GPT‑5.4 Mini and Nano"

pizza — Tue, 17 Mar 2026 21:22:02 +0000

I meant in the sense of - you have benchmarkers and trainers. If you publicize your evaluation, trainers may likely have their models 'consume' it, even if only indirectly: another person creating their own benchmark from scratch may be influenced by yours, even if the new question sets are clean-room. That, and the rule of thumb that benchmark value dissipates like sqrt(age) [0]

So there is a definite advantage to never publicizing your internal benchmark. But then, no one else can replicate your findings. You should assume that the space of benchmarks that are actually decent at evaluating model performance is much larger and most of the good ones, the ones that were costliest to produce, are hidden, and might not even correspond very well with the public ones. And that the public expensive benchmarks are selective and have a bias towards marketing purposes.

[0] https://www.offconvex.org/2021/04/07/ripvanwinkle/

New comment by pizza in "GPT‑5.4 Mini and Nano"

pizza — Tue, 17 Mar 2026 19:38:24 +0000

There’s a Dark Forest problem for evals. As soon as they’re made public they start running out of time to be useful. It’s also not clear how to predict how the model will perform on a task based on an eval. Or even whether, given two skills that the model can individually do well on in the evals, it still does well on their composition. It might at this point be better to be scientific in unscientific approaches, than to attribute more power to relatively weakly predictive evals than they actually have

New comment by pizza in "UBI as a productivity dividend"

pizza — Sat, 14 Mar 2026 21:34:41 +0000

In Singapore it seems 80% of people live in public housing https://en.wikipedia.org/wiki/Public_housing_in_Singapore though I can't speak as to what the effect is on its housing market

New comment by pizza in "Tell HN: I'm 60 years old. Claude Code has re-ignited a passion"

pizza — Sat, 07 Mar 2026 04:28:57 +0000

I mean. Sounds like the guy had existing long term goals, needed to overcome an activation threshold, and used AI as a catalyst to just get started. Seems like, behaviorally, AI was pivotal for him to learn things, even if the things he learned came from elsewhere / his own effort.