Hacker News: nutrientharvest

New comment by nutrientharvest in "Wine 11 rewrites how Linux runs Windows games at kernel with massive speed gains"

nutrientharvest — Tue, 24 Mar 2026 23:31:55 +0000

In many cases for game devs/publishers "supporting Linux" now means making sure the Windows build runs well under Proton.

New comment by nutrientharvest in "Watching o3 guess a photo's location is surreal, dystopian and entertaining"

nutrientharvest — Sat, 26 Apr 2025 19:30:14 +0000

It's cropping the original image then tokenizing it again with less downsampling, not cropping its internal representation.

New comment by nutrientharvest in "Falsehoods programmers believe about languages (localization)"

nutrientharvest — Sun, 02 Mar 2025 18:41:16 +0000

I live in Poland and speak Polish but tend to have my devices set to English because of how low quality UI translations tend to be. The endless confusion this causes to software is amusing. Worst/best of all is Google Maps on Android, in which random parts of the interface, including voice directions, can independently be in Polish or English, and which is which changes on its own on a daily basis.

New comment by nutrientharvest in "Twitch limiting uploads to 100 hours, deleting the rest starting April 19th"

nutrientharvest — Sat, 22 Feb 2025 13:16:15 +0000

It's the latter. Some people abuse the system by highlighting the full length of every broadcast, turning their highlights section into a complete archive of their streams, which is not something Twitch ever wanted to offer.

New comment by nutrientharvest in "Why do War Thunder players leak classified information? [video]"

nutrientharvest — Sun, 02 Feb 2025 22:05:26 +0000

Very true. The most recent "leak" was of a document that is available on archive.org and has been for years[1]. But the bottoms of the pages have "NATO RESTRICTED" printed, so it was declined, and news sites once again exploded with "War Thunder player leaks secret military documents on forum!!"

[1] https://archive.org/details/da-7-flight-manual-2003

New comment by nutrientharvest in "Starship Flight 7"

nutrientharvest — Fri, 17 Jan 2025 11:46:55 +0000

The requirements of orbital launch are unyielding. If you make a car 50% heavier, it will have worse mileage and handling, but it will still get you where you need to go. If you make a spacecraft 50% heavier, it will never reach orbit.

New comment by nutrientharvest in "[dead]"

nutrientharvest — Sun, 20 Oct 2024 04:55:23 +0000

> All LLM runs under the complexity of O(n), where n is the length of the output.

Wouldn't n be the length of the complete sequence including input?

New comment by nutrientharvest in "Run the strongest open-source LLM model: Llama3 70B with just a single 4GB GPU"

nutrientharvest — Wed, 26 Jun 2024 18:41:51 +0000

Ollama can already run Llama-3 70B with a 4GB GPU, or no GPU at all, it'll just be slow.

Considering this says it's "not designed for real-time interactive scenarios" it's probably also really slow

New comment by nutrientharvest in "Creativity has left the chat: The price of debiasing language models"

nutrientharvest — Mon, 17 Jun 2024 08:21:18 +0000

Usually "uncensored" models have been made by instruction tuning a model from scratch (i.e. starting from a pretrained-only model) on a dataset which doesn't contain refusals, so it's hard to compare directly to a "censored" model - it's a whole different thing, not an "uncensored" version of one.

More recently a technique called "orthogonal activation steering" aka "abliteration" has emerged which claims to edit refusals out of a model without affecting it otherwise. But I don't know how well that works, it's only been around for a few weeks.

New comment by nutrientharvest in "Creativity has left the chat: The price of debiasing language models"

nutrientharvest — Mon, 17 Jun 2024 08:01:38 +0000

Funnily enough, of all that I've tried, the model by the best at writing porn has been not one of ones uncensored and tuned exactly for that purpose, but stock Command R - whose landing page lists such exciting uses as "suggest example press releases" and "assign a category to a document".

New comment by nutrientharvest in "Creativity has left the chat: The price of debiasing language models"

nutrientharvest — Mon, 17 Jun 2024 07:51:44 +0000

That might be part of it, but I think the bigger factor is cost optimization. OpenAI in particular keeps replacing their models with with versions that are much faster (and therefore cheaper to run) which are supposed to be of equivalent quality but aren't really. GPT-4 -> GPT-4-Turbo -> GPT-4o have all been big upgrades to cost and latency but arguably downgrades to "intelligence" (or whatever you want to call it)