Hacker News: pet_the_bird

New comment by pet_the_bird in "Did my old job only exist because of fraud?"

pet_the_bird — Mon, 22 Jun 2026 00:00:04 +0000

The Dutch University of Delft systematically 'maximized' grants and shuffled the money between projects, according to investigative journalists of the NRC[1].

1 https://www.nrc.nl/nieuws/2024/02/15/het-subsidiepotje-moet-...

New comment by pet_the_bird in "Two Qwen3 models on one DGX Spark: the residency math"

pet_the_bird — Sun, 21 Jun 2026 13:52:19 +0000

Highy anecdotal: I have tried various self-hosted models using both vllm and llama.cpp. I am in a situation where I have access to large amount of memory (~320 GB).

While experimenting with quantization I found that there is a non-trivial tradeoff between quality and memory footprint. Overall my experience follows the reported pattern of "2-bit is mwah, 4-bit half decent and 6-bit required for programming. Still, although MiniMax-m2.7 is useable with the 6-bit quantizations that unsloth provides, it felt like such a breath of fresh air when I used the reference full-size model.

I find it difficult to say why. I had mostly the same setup as before (parsing had to be slightly adjusted in Zed). Aside from not experiencing the thinking loops (where minimax would get stuck generating the same sentences over and over) there is little evidence of any real improvement (although the average thinking time felt shorter).

I would recommend against very low quantizations of GLM 5.0/5.1/5.2 or Kimi 2.5/2.6. Smaller models were more reliable, and therefore more useful.

Robot Dogs Are a Security Nightmare [video]

pet_the_bird — Mon, 11 May 2026 10:40:15 +0000

Article URL: https://www.youtube.com/watch?v=lA8WuXDXfcI

Comments URL: https://news.ycombinator.com/item?id=48093271

Points: 5

# Comments: 0

New comment by pet_the_bird in "Why are neural networks and cryptographic ciphers so similar? (2025)"

pet_the_bird — Mon, 04 May 2026 12:26:02 +0000

"Cryptography Made Simple" By Nigel Smart and "A Graduate Course in Applied Cryptography" by Dan Boneh and Victor Shoup are excellent resources for people that have affinity with Math and CS. The second resource can be a tough read, and I would strongly recommend not skipping the first few chapters.

New comment by pet_the_bird in "Russia Poisons Wikipedia"

pet_the_bird — Sat, 02 May 2026 14:10:26 +0000

I think the article tried to refer to this link https://arxiv.org/abs/2504.10663 As I understand from scanning the paper, the authors attempt to determine differences between the Russian wikipedia articles and the articles on the Russian fork. They show that articles on the fork that were that differ from RU wikipedia have a significantly higher number of edits on RU wikipedia. The authors suggest that these may be signs of manipulations, however, it may not have affected the quality negatively (as stated in the discussion).

I do not find state sponsored activity on Wikipedia unlikely, but I am not convinced there is clear evidence that Russia poisoned wikipedia succesfully.