Hacker News: xyzal

New comment by xyzal in "U.S. sanctions against the A/I Collective"

xyzal — Fri, 28 Aug 2026 17:29:03 +0000

You are a good person.

New comment by xyzal in "U.S. sanctions against the A/I Collective"

xyzal — Fri, 28 Aug 2026 17:02:05 +0000

> Since they are now "sanctioned" by the US you cannot legally donate to them even if you wanted to...*[0]

There are many of us who do not live in the US, so many still can. Just not through paypal.

New comment by xyzal in "If I own Claude's outputs why can't I train my own model on them?"

xyzal — Thu, 13 Aug 2026 10:28:50 +0000

Nah, just pass the outputs to a person not related to Anthropic in any way.

Or -- maybe better -- just fsck the TOS.

New comment by xyzal in "Incentives are for losers"

xyzal — Sun, 09 Aug 2026 07:29:22 +0000

Even if you think you cannot escape incentives - there exists a thing for shaping them - it is called "regulation". Some still consider it a dirty word though.

New comment by xyzal in "What else do people draw on gradient.horse?"

xyzal — Fri, 24 Jul 2026 10:12:19 +0000

No need to hide "penis" behind asterisks. It is proper -- even medical -- terminology, in contast to "c*ck", "d*ng" or "schl*ng".

Will AI end the open internet? [video]

xyzal — Wed, 17 Jun 2026 16:23:53 +0000

Article URL: https://www.youtube.com/watch?v=gR2T1uxHG7o

Comments URL: https://news.ycombinator.com/item?id=48572706

Points: 1

# Comments: 0

New comment by xyzal in "AskHN:How do you handle skill atrophy from using coding agents?"

xyzal — Tue, 16 Jun 2026 14:21:38 +0000

I am curious how the workflow of people, who do not write code at all, looks like, or what products do they build. In my experience LLMs are an excavator, but you still have to tweak the fine details with a shovel.

AI Agents Enable Adaptive Computer Worms

xyzal — Tue, 16 Jun 2026 05:49:19 +0000

Article URL: https://cleverhans.io/worm.html

Comments URL: https://news.ycombinator.com/item?id=48551067

Points: 3

# Comments: 0

New comment by xyzal in "If you are asking for human attention, demonstrate human effort"

xyzal — Fri, 12 Jun 2026 05:41:47 +0000

How about reviving key signing parties?

New comment by xyzal in "Policy on the AI Exponential"

xyzal — Thu, 11 Jun 2026 06:40:23 +0000

Given that according to some we can code at 10x speed for at least half a year, I wonder where there are some autocoded softwares with 5 years' worth of equivalent human engineering work.

New comment by xyzal in "OpenAI mulls slashing prices as it competes with Anthropic for users"

xyzal — Thu, 11 Jun 2026 06:27:44 +0000

By socializing the losses of course.

New comment by xyzal in "OpenAI mulls slashing prices as it competes with Anthropic for users"

xyzal — Thu, 11 Jun 2026 06:26:58 +0000

I think it is time to start shorting.

The Jqwik Anti-AI Affair

xyzal — Thu, 11 Jun 2026 06:21:37 +0000

Article URL: https://blog.johanneslink.net/2026/06/09/the-jqwik-anti-ai-affair/

Comments URL: https://news.ycombinator.com/item?id=48486865

Points: 8

# Comments: 1

New comment by xyzal in "The OnlyFans Economy of American AI"

xyzal — Sun, 07 Jun 2026 15:14:02 +0000

I'll root for DeepSeek v4 Flash as well. It surprised me just how "good enough" it is for most of my needs, and also dirt cheap. Everyone should try it at least once.

New comment by xyzal in "LLMs are eroding my software engineering career and I don't know what to do"

xyzal — Sun, 07 Jun 2026 14:32:45 +0000

Which other profession has the same amount of training data freely available for the taking?

New comment by xyzal in "Harness engineering: Leveraging Codex in an agent-first world"

xyzal — Sun, 07 Jun 2026 11:42:02 +0000

Given that we can code at 10x speed for at least half a year, one would expect to see at least some pieces of machine-created software with 5 years' worth of equivalent human engineering work.

Anyone know some?

Ask HN: What are some agent-coded masterpieces?

xyzal — Sun, 07 Jun 2026 08:12:31 +0000

Given that we can code at 10x speed for at least half a year, one would expect to see at least some pieces of machine-created software with 5 years' worth of equivalent human engineering work.

Are there any we know of?

Comments URL: https://news.ycombinator.com/item?id=48432881

Points: 2

# Comments: 1

New comment by xyzal in "Ask HN: What was your "oh shit" moment with GenAI?"

xyzal — Sat, 06 Jun 2026 11:46:37 +0000

YC indisputably has a financial incentive for AI sentiment to be positive on HN. The structural conflict of interest is worth being aware of.

New comment by xyzal in "S&P 500 rejects SpaceX, also blocking entry for OpenAI and Anthropic"

xyzal — Sat, 06 Jun 2026 09:03:23 +0000

Mind that the host of this site has an interest in keeping the hype going on ...

New comment by xyzal in "Ask HN: What was your "oh shit" moment with GenAI?"

xyzal — Sat, 06 Jun 2026 06:32:02 +0000

To me it was just a few weeks ago discovering just how good and dirt cheap the recent flash models are, in particular Deepseek V4. Previously used Claude's variants almost exclusively.

I use them mostly in the "artist's assistant" role, doing internet research, writing a occasional function and doing transformations or refactorings (don't belive the agentic hype honestly), and for such tasks they seem to be well capable enough.

It seems that their open weights nature leads to competition among providers keeping the user cost close to inference cost.

Try them at least once if you haven't, it's well worth it, and the price difference is staggering