Hacker News: nathanfig

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 21:07:22 +0000

I didn't ask about standard representations, but I think the graph representation works well when paired with the discrete moves this maze allows. If we wanted continuous spaces it would probably be harder to play and harder to represent, but maybe more interesting.

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 21:03:39 +0000

Yes, I've exported and self-hosted in the past. I'll sometimes use Sites to mockup something and then I'll download and pass it onto another agent for use in something bigger.

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 19:19:00 +0000

Yeah you have actually picked up on something that the game intentionally does to help the user: If the cheese can be reached by any of the dimensions you are currently on, then it will ensure that dimension is in the subset of dimensions you receive when you go down a dimension.

Otherwise I'd have to introduce some kind of dimension picker. Maybe a fun hardmode.

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 19:16:54 +0000

Yeah if I were to make it harder I would introduce more sparse connections requiring the user to actually understand the layout, right now you can kind of "feel" your way towards the cheese. Most brutal maze would be one with only one possible path and lots of backtracking required.

Maybe I'll do that and post to HN later, didn't expect this much response!

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 19:15:16 +0000

Yeah I think moving up and down dimensions is the best strategy. I find spamming movements in 6D works but only because the paths are dense enough.

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 18:11:12 +0000

I mean, with a Cloudflare API key you can pretty much already do that. But keeping your data straight and putting in the right authentication guards, etc. are all still where the business problems meet technical and the AI is not enough unless the agent is both architect and customer.

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 18:00:57 +0000

Thanks! Yeah I find it helps to jump back down to lower (comprehensible) dimensions, move closer at that level, then jump back up.

New comment by nathanfig in "Sites in ChatGPT"

nathanfig — Fri, 02 Oct 2026 17:42:46 +0000

I have been using Sites a few months now, really underrated feature IMO. Any time I get a quick app idea I just hand it off to Sites and usually get a working prototype within the hour.

For example, here's a game where you are a mouse hunting for cheese in higher dimensional mazes: https://manifold-six-axis-escape.nathan338932.chatgpt.site/

Idea came to me while at a concert, was ready to play that same night.

New comment by nathanfig in "Ask HN: Does AI watermarking present a new attack vector?"

nathanfig — Fri, 14 Aug 2026 11:49:22 +0000

I hope open weights continue to gain traction for this reason too.

But injecting information into outputs is a little different than Claude keeping tabs internally. The point is that we do not know what information is being shipped with the fingerprint itself.

Ask HN: Does AI watermarking present a new attack vector?

nathanfig — Fri, 14 Aug 2026 11:20:01 +0000

One thing I have not been able to determine from Claude's documentation on watermarking[0] is what kind of metadata they store in association with a given watermark. Ostensibly they could make the fingerprints as unique as they want, possibly down to the exact time, user and session.

If so, this seems like hidden risk that AI users are probably not considering. Any code you write now carries information that you might not want revealed. If a bad actor gets the keys then they may be able to de-anonymous open source contributors who want to stay hidden. Or perhaps enough fingerprints across a sample could reveal internal organization details a company would rather not disclose.

Someone with more imagination can probably come up with better examples, it just seems like an attack vector that I haven't seen much consideration for.

[0] https://support.claude.com/en/articles/16266773-how-claude-marks-ai-generated-content

Comments URL: https://news.ycombinator.com/item?id=49297267

Points: 4

# Comments: 6

New comment by nathanfig in "Please Use AI"

nathanfig — Fri, 29 May 2026 15:13:06 +0000

Yeah it's entirely possible the answer is: "You don't." But I haven't even seen people attempting to imagine life well-lived. Like with so much technology it's basically an afterthought.

New comment by nathanfig in "Please Use AI"

nathanfig — Fri, 29 May 2026 14:37:15 +0000

I've been pondering the question: "What does it mean to live well with AI?"

We are certainly scrambling for productivity with "token maxxing" and scrambling for entertainment with AI companions, but I haven't seen many thoughtful takes on how AI might look in a life well-lived.

You Cannot Defeat Entropy with AI

nathanfig — Wed, 18 Feb 2026 17:58:11 +0000

Article URL: https://greenbluegray.substack.com/p/you-cannot-defeat-entropy-with-ai

Comments URL: https://news.ycombinator.com/item?id=47063972

Points: 4

# Comments: 0

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 15:55:18 +0000

I think that's a big part of what makes this era so much fun - the interactivity tailored to my capability. I feel like I'm 13 again discovering I can write HTML and feeling that giddy delight when I try something and it just works.

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 15:52:30 +0000

>> I haven’t had great luck with winding it up and setting it loose on a task

Yeah there is real labor involved with trying to set up the right context and explain your goals and ensure it has the right tools, etc. etc. Sometimes the payoff is massive, other times it's like "I could have done this myself faster". It takes time to build an intuition for which is which, and I'm constantly having to tune that heuristic as new models and tools come out.

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 15:48:34 +0000

Yeah, the social problem of "how do we use technology well" persists.

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 15:47:49 +0000

Can you elaborate? I might agree, I'm just curious what you mean.

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 15:46:51 +0000

Yeah in fact I think we are going through some major upheaval in development practices because we haven't yet figured out what constraints to put in our use of AI, and so plastering it everywhere we trip ourselves up. It took a while to figure out that private members were useful and GOTO is harmful despite how tempting it might be.

I think similarly we will find that using AI to take shortcuts around design is mostly harmful, but using it to fulfill interfaces is brilliant. Eventually a set of best practices will evolve.

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 14:32:50 +0000

Because they are already useful for programming and unlikely to get worse!

New comment by nathanfig in "Now might be the best time to learn software development"

nathanfig — Wed, 18 Jun 2025 14:28:54 +0000

Sounds a lot like "nothing ever happens"?