Hacker News: segh

New comment by segh in "Making Video Games in 2025 (without an engine)"

segh — Mon, 02 Mar 2026 09:43:21 +0000

I can do long division manually but I still reach for a calculator.

New comment by segh in "We tasked Opus 4.6 using agent teams to build a C Compiler"

segh — Fri, 06 Feb 2026 15:42:30 +0000

People skills :/

New comment by segh in "We tasked Opus 4.6 using agent teams to build a C Compiler"

segh — Fri, 06 Feb 2026 15:41:23 +0000

This is an experiment to see the current limit of AI capabilities. The end result isn't useful, but the fact is established that in Feb 2026, you can spend $20k on AI to get a inefficient but working C complier.

New comment by segh in "How AI assistance impacts the formation of coding skills"

segh — Fri, 30 Jan 2026 10:20:33 +0000

Far far more people use ChatGPT than Claude.ai

New comment by segh in "The Jolla Phone Proved We've Been Using Smartphones Wrong All Along"

segh — Mon, 19 Jan 2026 10:52:55 +0000

I started skimming and instantly thought AI, then came to the comments to see if it was just me.

New comment by segh in "Lights and Shadows (2020)"

segh — Thu, 08 Jan 2026 14:38:53 +0000

School incentives are not really aligned around maximizing learning rate for every student. (E.g. that is why there is/was debate around teaching phonetics)

New comment by segh in "Can LLMs do randomness?"

segh — Wed, 30 Apr 2025 15:24:32 +0000

Cool experiment! My intuition suggests you would get a better result if you let the LLM generate tokens for a while before giving you an answer. Could be another experiment idea to see what kind of instructions lead to better randomness. (And to extend this, whether these instructions help humans better generate random numbers too.)

New comment by segh in "Widespread power outage in Spain and Portugal"

segh — Tue, 29 Apr 2025 11:09:33 +0000

ChatGPT's tone is slowly taking over the entire internet

New comment by segh in "AI Horseless Carriages"

segh — Thu, 24 Apr 2025 16:19:29 +0000

People still play chess, even though now AI is far superior to any human. In the future you will still be able to hand-write code for fun, but you might not be able to earn a living by doing it.

New comment by segh in "AI Horseless Carriages"

segh — Thu, 24 Apr 2025 16:18:14 +0000

The live demos are using a very cheap and not very smart model. Do not update your opinion on AI capabilities based on the poor performance of gpt-4o-mini

New comment by segh in "AI Horseless Carriages"

segh — Thu, 24 Apr 2025 16:15:56 +0000

The system prompt can include examples. That is often a good idea.

New comment by segh in "AI Agents: Less Capability, More Reliability, Please"

segh — Mon, 31 Mar 2025 16:32:22 +0000

For me, they have come from the AI labs themselves. I have been impressed with Claude Code and OpenAI's Deep Research.

New comment by segh in "AI agents: Less capability, more reliability, please"

segh — Mon, 31 Mar 2025 15:27:12 +0000

Lots of people are building on the edge of current AI capabilities, where things don't quite work, because in 6 months when the AI labs release a more capable model, you will just be able to plug it in and have it work consistently.

New comment by segh in "Why Anthropic's Claude still hasn't beaten Pokémon"

segh — Thu, 27 Mar 2025 17:12:42 +0000

This is the crux of the issue. Whether you think this is like extending a ladder to the moon, or more like we figured out how to get to the moon and are now aiming at Jupiter.

New comment by segh in "Why Anthropic's Claude still hasn't beaten Pokémon"

segh — Thu, 27 Mar 2025 17:05:01 +0000

Claude Plays Pokemon is one person's side project to see how well Sonnet can play pokemon. It is a neat LLM benchmark; it's not a serious attempt at making Pokemon-playing AI.

New comment by segh in "Why Anthropic's Claude still hasn't beaten Pokémon"

segh — Thu, 27 Mar 2025 12:56:23 +0000

Do you disagree that AI will ever reach the level of a "high-income knowledge worker", or do you disagree that it will happen in a year or two?

New comment by segh in "The model is the product"

segh — Tue, 18 Mar 2025 13:19:49 +0000

If you have not been reading every OpenAI blog post, you can't be blamed for thinking the model picker affects Deep Research, since the UI heavily implies that.

New comment by segh in "Show HN: I convert videos to printed flipbooks for living"

segh — Tue, 04 Feb 2025 12:29:29 +0000

I don't know how literal you are being, but it is not "literally anyone". If you know what FFmpeg is, you are a very small minority of the population.

New comment by segh in "AI Is Bad/Good Actually"

segh — Tue, 05 Mar 2024 11:11:49 +0000

Improve in intelligence and capability.

New comment by segh in "How to run a small social network (2019)"

segh — Sun, 31 Dec 2023 18:33:29 +0000

Enforce a no politics rule.