Hacker News: artursapek

New comment by artursapek in "Gait Game"

artursapek — Tue, 28 Jul 2026 10:53:36 +0000

It’s an attempt to build a locomotion model. It uses a physics engine (Avian in Bevy) to animate a walking biped from first principles.

An earlier version can be seen here https://x.com/tmuxvim/status/2081822899149963422

New comment by artursapek in "Gait Game"

artursapek — Tue, 28 Jul 2026 10:50:25 +0000

WASD or just touch on mobile

Gait Game

artursapek — Tue, 28 Jul 2026 10:45:52 +0000

Article URL: https://gait.game

Comments URL: https://news.ycombinator.com/item?id=49081968

Points: 2

# Comments: 4

New comment by artursapek in "Claude Opus 5"

artursapek — Fri, 24 Jul 2026 20:52:53 +0000

It's definitely not cheaper than Sonnet on my benchmark, but it's cheaper than Fable and outperforms it. Which is big IMO. https://revise.io/errata-bench

ErrataBench

artursapek — Fri, 24 Jul 2026 18:08:24 +0000

Article URL: https://revise.io/errata-bench

Comments URL: https://news.ycombinator.com/item?id=49039503

Points: 1

# Comments: 0

New comment by artursapek in "Claude Opus 5"

artursapek — Fri, 24 Jul 2026 17:22:16 +0000

HN users are world champions are trivializing difficult things with snarky comments

Kimi K3 first open model in ErrataBench Top 10

artursapek — Wed, 22 Jul 2026 01:34:31 +0000

Article URL: https://revise.io/blog/07-22-2026/kimi-k3-broke-90-on-erratabench

Comments URL: https://news.ycombinator.com/item?id=49000711

Points: 3

# Comments: 0

New comment by artursapek in "Sleep regularity is a stronger predictor of mortality risk than sleep duration (2023)"

artursapek — Wed, 15 Jul 2026 14:16:05 +0000

Yep, I've been taking glycine and magnesium for years. I am not as consistent as I should be but it makes a big difference when I use them.

GPT 5.6 sets new record on proofreading benchmark

artursapek — Mon, 13 Jul 2026 15:40:27 +0000

Article URL: https://twitter.com/tmuxvim/status/2076692926059499646

Comments URL: https://news.ycombinator.com/item?id=48894384

Points: 1

# Comments: 0

Mantissa, a distributed workload orchestration system

artursapek — Sat, 04 Jul 2026 23:08:22 +0000

Article URL: https://mantissa.io/blog/init/

Comments URL: https://news.ycombinator.com/item?id=48789896

Points: 4

# Comments: 0

New comment by artursapek in "Claude Sonnet 5"

artursapek — Tue, 30 Jun 2026 20:41:59 +0000

I run a proofreading benchmark that tests how well models can find and fix errors in English text. They get several passes in a simple agent loop. Sonnet 5 is definitely better than Sonnet 4.6, but inferior on both quality and cost to GLM 5.1, GLM 5.2, Gemini 3.1 Flash, and Gemini 3.1 Pro. https://revise.io/errata-bench

New comment by artursapek in ".self: A new top-level domain designed to support self-hosting"

artursapek — Tue, 30 Jun 2026 01:39:40 +0000

haha yeah I've bet the last 12 months of my career on a .io

New comment by artursapek in ".self: A new top-level domain designed to support self-hosting"

artursapek — Mon, 29 Jun 2026 23:32:02 +0000

The .ai TLD is some tiny island with a few thousand people

New comment by artursapek in ""No, I swear I wrote this.""

artursapek — Sat, 27 Jun 2026 22:15:20 +0000

Trivial to simulate basic keystrokes. But I don't think it's trivial to simulate the natural process of drafting something. There's no concrete heuristic or algorithm (yet) for judging these types of replays, but I'd be impressed if someone can actually make a program that reliably simulates a natural keystroke-by-keystroke thought process which appears human when replayed in this way.

"No, I swear I wrote this."

artursapek — Sat, 27 Jun 2026 17:17:05 +0000

Article URL: https://revise.io/blog/06-27-2026/no-i-swear-i-wrote-this

Comments URL: https://news.ycombinator.com/item?id=48699969

Points: 17

# Comments: 46

New comment by artursapek in "Previewing GPT‑5.6 Sol: a next-generation model"

artursapek — Fri, 26 Jun 2026 18:02:47 +0000

They claim extreme performance on ExploitBench, which Mythos was touted as being incredible at. https://x.com/OpenAI/status/2070555278576439306

Which LLM is the best proofreader?

artursapek — Thu, 11 Jun 2026 14:00:52 +0000

Article URL: https://revise.io/errata-bench

Comments URL: https://news.ycombinator.com/item?id=48490478

Points: 1

# Comments: 0

New comment by artursapek in "Claude Fable 5"

artursapek — Tue, 09 Jun 2026 21:01:03 +0000

Fable 5 beats GPT 5.5 in my proofreading benchmark. And it does so at approximately the same total cost; it used significantly fewer turns than 5.5

https://x.com/tmuxvim/status/2064452096800198930

New comment by artursapek in "Siri AI"

artursapek — Tue, 09 Jun 2026 13:32:53 +0000

I would expect Apple to hedge their bet on Gemini and build everything so that the model can be swapped out in the future.

New comment by artursapek in "Apple reveals new AI architecture built around Google Gemini models"

artursapek — Mon, 08 Jun 2026 22:22:22 +0000

I use Carplay all the time and I didn't even realize it has voice control. I just set things up on my phone and drive.