Hacker News: puchatek

New comment by puchatek in "What AI did to stackoverflow in a graph"

puchatek — Sat, 18 Jul 2026 15:18:58 +0000

I might even find an upside in that if it wasn't for the fact that this stolen crowd sourced knowledge will be used to make some billionaires even richer.

New comment by puchatek in "The kids with phones are alright"

puchatek — Tue, 14 Jul 2026 14:25:14 +0000

Reminds me of the "good guys with guns" narrative tbh. Being able to document such transgressions is not enough of an argument to do a mass rollout of ad display technology IMHO.

New comment by puchatek in "New York City to ban deceptive subscription practices"

puchatek — Fri, 10 Jul 2026 20:32:59 +0000

People who forget to cancel are easy prey.

Cancelling then being charged anyway is something else. Something akin to being robbed IMHO.

New comment by puchatek in "Cloudflare Drop"

puchatek — Thu, 09 Jul 2026 05:22:53 +0000

Amazing stuff? You mean technical stuff right? Because politically, demographically and climate wise we're headed for disaster or at least dystopia.

I wished I could find all the science and technology as uplifting as my friends do.

New comment by puchatek in "Anonymous GitHub account mass-dropping undisclosed 0-days"

puchatek — Sun, 28 Jun 2026 07:05:37 +0000

Approaching the maintainers would be ideal but time-consuming. Disclosing it like this is neutral I guess. Better than selling it in the darknet.

I do wonder though: if you can tell the AI to search for vulns, can't you also tell it to contact the right maintainer for each one found?

New comment by puchatek in "Claude for Small Business"

puchatek — Thu, 14 May 2026 08:45:58 +0000

The future is perpetually dealing with the fallout from all the vibe coding as the pool of people who'd have a shot at fixing it gets smaller and smaller. Shitty will be the new normal.

New comment by puchatek in "America Lost the Mandate of Heaven"

puchatek — Sat, 18 Apr 2026 10:48:04 +0000

So... What's your point? Sounds like what you identify as key is not relevant to the remainder of your comment.

New comment by puchatek in "Honda is killing its EVs"

puchatek — Tue, 17 Mar 2026 22:15:28 +0000

Big Vintage Energy

New comment by puchatek in "Many SWE-bench-Passing PRs would not be merged"

puchatek — Thu, 12 Mar 2026 06:02:52 +0000

Uhm... the second sentence of the original comment does contain "x isn't Y, it's Z". You missed this... just like an LLM will miss things sometimes, making me wonder if you're one of them, too.

/s ... ?

New comment by puchatek in "How do I cancel my ChatGPT subscription?"

puchatek — Sat, 28 Feb 2026 06:59:33 +0000

For me this is very valuable. The results of personal "research projects" are in there. I use it for reference. Of course I could ask Claude to get me those answers but why waste the energy?

New comment by puchatek in "We Will Not Be Divided"

puchatek — Sat, 28 Feb 2026 06:36:21 +0000

It's different because we are talking about a technology that we might lose control over at some point. Those drones in your example might make an entirely different choice than what you anticipated when you let them take off.

New comment by puchatek in "We Will Not Be Divided"

puchatek — Sat, 28 Feb 2026 06:26:02 +0000

Democratic governments care about this to a degree but only autocratic ones get paranoid.

New comment by puchatek in "We Will Not Be Divided"

puchatek — Sat, 28 Feb 2026 05:58:52 +0000

The deal was only possible because anthropic stayed by their convictions. OpenAI didn't have agency in that. You're making it sound like Altman orchestrated the whole thing.

New comment by puchatek in "The Age Verification Trap: Verifying age undermines everyone's data protection"

puchatek — Tue, 24 Feb 2026 17:46:12 +0000

No, we will exclude a few people because Germany doesn't have its shit together when it comes to digital stuff. Then hopefully people will complain and things will improve.

New comment by puchatek in "The Age Verification Trap: Verifying age undermines everyone's data protection"

puchatek — Tue, 24 Feb 2026 15:58:55 +0000

This is the way. It annoys me to no end when e.g. the German chancellor demands clearnames in social media. The real issue are bots and algorithmically enhanced reach. Proof of personhood in a privacy-preserving way is enough to fix this. But it should be mandatory for social media in the EU. You don't need to expose people to the doxxing mob to protect our democracy.

Tbh, when I read that "platforms face a choice between excluding lawful users and monitoring everyone." I don't have much understanding.

No gov. ID, no participation. It's not like you cannot go outside and talk to people anymore so let's not pretend that being on insta is some sort of universal human right and anybody barred from it is some sort of terrible tragedy.

New comment by puchatek in "How I use Claude Code: Separation of planning and execution"

puchatek — Sun, 22 Feb 2026 08:06:48 +0000

Spec-driven looks very much like what the author describes. He may have some tweaks of his own but they could just as well be coded into the artifacts that something like OpenSpec produces.

New comment by puchatek in "How I use Claude Code: Separation of planning and execution"

puchatek — Sun, 22 Feb 2026 07:35:42 +0000

Is it on GH?

New comment by puchatek in "European Tech Alternatives"

puchatek — Thu, 19 Feb 2026 06:26:45 +0000

I believe the EU inc initiative attempts to fix the capital aspect

New comment by puchatek in "We mourn our craft"

puchatek — Sat, 07 Feb 2026 21:26:42 +0000

If that was the final price, no strings attached and perfect, reliable privacy then I might consider it. Maybe not for the current iteration but for what will be on offer in a year or two.

But as it stands right now, the most useful LLMs are hosted by companies that are legally obligated to hand over your data if the US gov. had decided that it wants it. It's unacceptable.

New comment by puchatek in "Opus 4.5 is not the normal AI agent experience that I have had thus far"

puchatek — Wed, 07 Jan 2026 08:06:01 +0000

That's been my main argument for why LLMs might be at their zenith. But I recently started wondering whether all those codebases we expose to them are maybe good enough training data for the next generation. It's not high quality like accepted stackoverflow answers but it's working software for the most part.