Hacker News: bkettle

Fibonacci in the NYT Pips

bkettle — Thu, 15 Jan 2026 19:13:06 +0000

Article URL: https://benkettle.xyz/posts/fibonacci-pips/

Comments URL: https://news.ycombinator.com/item?id=46637595

Points: 2

# Comments: 0

New comment by bkettle in "Local Journalism Is How Democracy Shows Up Close to Home"

bkettle — Tue, 13 Jan 2026 17:32:36 +0000

Are you based in Alaska? I’d love to hear more if so; I have contact info on my website. (I grew up there).

New comment by bkettle in "AI will make formal verification go mainstream"

bkettle — Tue, 16 Dec 2025 23:52:13 +0000

Yeah fair enough. I can definitely see the value of property-based verification like this and agree that useful properties could be easy to express and that LLMs could feasibly verify them. I think full verification that an implementation implements an entire spec and nothing else seems much less practical even with AI, but of course that is just one flavor of verification.

New comment by bkettle in "AI will make formal verification go mainstream"

bkettle — Tue, 16 Dec 2025 23:00:58 +0000

The whole point I think, though, is that it doesn’t matter. If an LLM hallucinates a proof that passes the proof checker, it’s not a hallucination. Writing and inspecting the spec is unsolved, but for the actual proof checking hallucinations don’t matter at all.

New comment by bkettle in "AI will make formal verification go mainstream"

bkettle — Tue, 16 Dec 2025 22:49:46 +0000

I think formal verification shines in areas where implementation is much more complex than the spec, like when you’re writing incomprehensible bit-level optimizations in a cryptography implementation or compiler optimization phases. I’m not sure that most of us, day-to-day, write code (or have AI write code) that would benefit from formal verification, since to me it seems like high-level programming languages are already close to a specification language. I’m not sure how much easier to read a specification format that didn’t concern itself with implementation could be, especially when we currently use all kinds of frameworks and libraries that already abstract away implementation details.

Sure, formal verification might give stronger guarantees about various levels of the stack, but I don’t think most of us care about having such strong guarantees now and I don’t think AI really introduces a need for new guarantees at that level.

New comment by bkettle in "What Americans die from vs. what the news reports on"

bkettle — Wed, 15 Oct 2025 02:46:35 +0000

> there's a section of road near where I live that's dangerous, and we all know it's dangerous

Clearly not enough people know it’s dangerous or how dangerous it is, or one of them would do something about it

New comment by bkettle in "Uv overtakes pip in CI"

bkettle — Tue, 14 Oct 2025 01:55:38 +0000

Semgrep has supported uv for months now (I added it).

New comment by bkettle in "NanoChat – The best ChatGPT that $100 can buy"

bkettle — Mon, 13 Oct 2025 18:51:56 +0000

This free tradition in software is I think one of the things that I love so much, but I don't see how it can continue with LLMs due to the extremely high training costs and the powerful hardware required for inference. It just seems like writing software will necessarily require paying rent to the LLM hosts to keep up. I guess it's possible that we'll figure out a way to do local inference in a way that is accessible to everyone in the way that most other modern software tools are, but the high training costs make that seem unlikely to me.

I also worry that as we rely on LLMs more and more, we will stop producing the kind of tutorials and other content aimed at beginners that makes it so easy to pick up programming the manual way.

New comment by bkettle in "$912 energy independence without red tape"

bkettle — Sun, 05 Oct 2025 05:28:21 +0000

Do you mean 500Wh rather than 500kWh? 500 kWh would be around 10 EVs worth, and looks like it costs around 700k [0], but 500Wh seems to be a common size for portable power stations [1]

[0] https://www.backupbatterypower.com/products/516-kwh-industri...

[1] https://www.ankersolix.com/products/535?variant=497024349310...

New comment by bkettle in "PEP 810 – Explicit lazy imports"

bkettle — Fri, 03 Oct 2025 20:25:43 +0000

I think they mean backwards-compatible syntax-wise, rather than actually allowing this feature to be used on existing code. If I’m understanding correctly they would prefer for the Python grammar to stay the same (hence the comment about updating parsers and IDEs).

But I don’t think I really agree, the extensible annotation syntaxes they mention always feel clunky and awkward to me. For a first-party language feature (especially used as often as this will be), I think dedicated syntax seems right.

New comment by bkettle in "US government shuts down after Senate fails to pass last-ditch funding plan"

bkettle — Wed, 01 Oct 2025 06:41:56 +0000

I found out yesterday that it is now the law that federal employees are guaranteed back pay after a shutdown, thanks to the Government Employee Fair Treatment Act of 2019 passed after the 2019 shutdown.

https://www.congress.gov/bill/116th-congress/senate-bill/24

New comment by bkettle in "Cloudflare Email Service: private beta"

bkettle — Thu, 25 Sep 2025 20:27:39 +0000

I actually think that Cloudflare has made publishing on the internet _more_ accessible for many individuals. I’ve helped a few people get personal websites running on Cloudflare pages and run my own there—it’s free and extremely easy. They could obviously pull the plug at any point, but with static sites it’s easy to avoid lock-in. If it weren’t for Cloudflare and other services that give free, easy hosting, I suspect there would be even fewer of the non-commercial small-internet sites that you value.

New comment by bkettle in "Slow social media"

bkettle — Thu, 18 Sep 2025 07:11:23 +0000

Hm. To me it is indeed the big bad tech companies at fault for implementing _obviously_ user-hostile functionality (mandatory suggested posts, for example) that exploit human nature to keep people engaged, even if people prefer not to. The existence and widespread use of screen time limits on phones seems like strong evidence that tech companies have built something nefarious here. And to me, it’s clear that they have done so intentionally to drive profit.

New comment by bkettle in "Slow social media"

bkettle — Thu, 18 Sep 2025 07:05:07 +0000

Is it possible that the current form of social media is actually contributing to the erosion of your liberties because it is so widely used in society and is likely contributing to polarization and antisocial behavior?

I see this (and, honestly, most problems) as much more than a personal responsibility issue. To me, it’s an issue of misaligned incentives and unpriced downside costs. It’s clear that market forces push companies to build an addictive service that produces long term misery. It’s also clear that social media has a cost on its users (producing long term misery, reducing acute productivity) But this cost is not paid by the social media company.

I’d argue that widespread use of the social media that today’s market incentives create is bad for society as a whole, not only for any one individual. Correcting market incentives that don’t align with social good is, in my opinion, one of the most essential purposes of legislation.

New comment by bkettle in "Slow social media"

bkettle — Wed, 17 Sep 2025 04:19:49 +0000

I think modern social media is a huge problem but don’t see we can fix it without regulation. It’s clear that all the current incentives point companies towards engagement and rage bait and away from anything actually “social”, and I think it’s unlikely that any new social network that tries to fix these issues would achieve widespread usage.

Have any countries proposed legislation to help reign it in? What would that legislation look like? My main idea is to simply outlaw ML-based recommendation algorithms, but obviously that is not as simple as it sounds and is mostly based on looking fondly on the earlier days of social media, when I felt like it was making my life better instead of worse.

New comment by bkettle in "Tesla changes meaning of 'Full Self-Driving', gives up on promise of autonomy"

bkettle — Sat, 06 Sep 2025 04:24:25 +0000

> it works for human drivers

Sure, for some definition of "works"...

https://www.iihs.org/research-areas/fatality-statistics/deta...

New comment by bkettle in "A computer upgrade shut down BART"

bkettle — Fri, 05 Sep 2025 22:08:11 +0000

Why are the socio-political stars aligned in tens of countries across Europe and Asia but not in the US, if such alignment is so rare?

I might argue that the bay area focuses on transportation technology that is flashy and gets around existing regulations because it is new, with hardly any regard at all for how it scales.

New comment by bkettle in "A computer upgrade shut down BART"

bkettle — Fri, 05 Sep 2025 22:01:11 +0000

Note that Caltrans only maintains state roads; looks like from that document that they distribute some money to localities but as far as I can tell we can't see what fraction of local road maintenance that covers. Of course localities also have parking fees, traffic tickets, etc that can help cover road maintenance.

New comment by bkettle in "A computer upgrade shut down BART"

bkettle — Fri, 05 Sep 2025 21:49:53 +0000

There are also a variety of ways that "efficiency" can be defined; your comment considers monetary efficiency, but both modes of transport have costs on society that are not considered in the numerical operating costs (pollution, opportunity cost of land use, healthcare costs due to accidents...)

New comment by bkettle in "De-Googling TOTP Authenticator Codes"

bkettle — Mon, 01 Sep 2025 17:34:00 +0000

Do we have the crypto to build an asymmetric variant? The way that I’d immediately think to do it is have the authenticator create a signature over the current time (chunked into 30-second windows) and the service verify that signature, but obviously those signature texts are way too long to manually enter as a one-time code. (Symmetric) TOTP solves this length problem by just truncating the output of a hash function, which both the authenticator and the service can produce because they have a shared secret. But in the signature case the service would need the entire signature to validate it; any truncation would make it useless.

It’s been a while since I did any crypto. But it feels like the shortness of the one-time-code probably makes it impossible to do asymmetrically. If this is indeed the case there is probably an elegant proof or some better way of thinking about why it’s impossible. I would be interested in reading that.