News for hackers. Stuff that matters.

jMyles

Re:Much Ado About None: Sentinel Values in Python on Saturday October 31, 2026 @09:36PM (#50321972)
I have a library that's approximately half joking that I use for (sentinel and other) special value constants. If you like reading test suites, this one will get ya:

https://github.com/nucypher/constantSorrow/blob/master/tests...

Re:Much Ado About None: Sentinel Values in Python on Saturday October 31, 2026 @09:35PM (#50321967)
> No, it returns KeyError. Or at least it should.

Surely you meant _raises_? Or did you literally mean to return an exception class?

Anecdotally, does this match anyone's actual experience?

Everyone I know who is working closely with LLMs is doing some of the most acute critical thinking of their lives these days.

I'm very skeptical of these kinds of takes; I'm reminded that Socrates railed against books on the basis that they were going to rot the brains of the youth.

I don't know - it depends who specifically you're talking about.

You're the one who invented these "crypto enthusiasts" in order to rail against them - whom do you mean exactly? Got a particular talk from devcon? Let alone regens unite / schelling point / GFEL / some other radical ethereum event?

Who are these "crypto enthusiasts" who, in 2026, have a "stated goal" of "some alternative 'currency'"?

Look at the schedule of any crypto conference - ethereum events in particular, but others as well - and you will find that your critique has already been widely adopted. The discussion is around replacing power structures and sources of injustice, not merely 'currency'.

I was reluctant to get involved in blockchain tech because I don't care for currency and don't view it as a likely organizing principle for humans for the next thousand years. That was in like 2016.

Today, nearly every blockchain project attempts to overcome challenges other than currency.

Re:PHP 8.2.35 on Wednesday October 21, 2026 @11:01AM (#50182222)
Yeah, I was very surprised to wake up just now and find docker still vulnerable.
Re:The 1.5°C goal of the Paris Agreement is no longer viable on Wednesday October 21, 2026 @05:43AM (#50178128)
Yeah, why _are_ night trains uncompetitive? Is it just so simple as "air travel doesn't have externalities priced in"?
I don't know for sure, but I found it interesting that there was such lively debate, and yet what they ended up doing seems to have been one of the few solutions that was intuitively worse than nothing.

Heck man, give Stamets a shot with his oil-eating mushrooms; I thought that was an interesting suggestion.

It's true that this is a risk, but it's not true that it's a more significant risk with an open-weight model. It's less of a risk. Much, much less.
I think that "infer once, repeat resulting logic forever" is one of the most important / sustainable / humane ways to imagine how AI acclimates to the universe.

https://pickipedia.xyz/wiki/PickiPedia:Artificial_Dumb

Re:William Gibson and the Trademarking of Cyberspace on Sunday October 18, 2026 @09:45PM (#50147401)
Whatever else is real or myth with regard to these trolly pranking trademark spats, I think it's pretty reasonable to say that the term "cyberspace" emerged amidst - or at least adjacent to - a lot of LSD and Grateful Dead tapes.

https://w2.eff.org/Misc/Publications/John_Perry_Barlow/HTML/...

Absolutely.

But, did AI actually _help_ in this specific phrase? I think it's OK to point out the quirks (and even drawbacks) of AI, even if we generally support the underlying motions of the person using it.

Actually man, I've transitioned pretty much full-time to playing bluegrass. Given that my life has been spend watching and helping the internet grow, most of my songs are about what it has felt like to live amidst that: https://justinholmes.bandcamp.com/

I still hack every day (current project: pickipedia.xyz), but I'm not doing security anymore except for my own projects.

And yes, I was more making light of what seem to me to largely be political disagreements; of course I respect your work at the same time.

It seems clearer and clearer to me that nationa-states cannot possibly withstand the evolution of the internet in any dignified way, and I pray for peaceful, simple deprecation of them. It seems to me to be the sensible stance, both in terms of security and in terms of joy.

My repeated sense - right honed or wrongly - is that you defend these structures (nation-states generally, and intelligence operations / state secret brokers in particular) in ways that I find hard to reconcile, even upon extended reflection.

> I think I agree with you in spirit

holy heck, the planets have aligned, HN

;-)

That's the opposite of what my gut is telling me.

Hashes are so simple. There's no place for these crazy math attacks, which rely on rethinking long-standing and otherwise reasonable assumptions, to hide.

Surely we can all agree that there's no way to reverse a modulus.

Yeah, I've been extremely disappointed with disputes on Amex Platinum.

I literally had a bike rental company in Amsterdam attempt to get me to conspire to report a bike stolen (which had briefly been stolen but was back in my possession), and then when I refused, charged me as if the bike had been stolen!

I reported all this to Amex, and I was never refunded, and to my knowledge, the attempt fraud was never investigated.

Although we'll probably never see evidence, I think this is almost certainly what is happening.
Re:Homa: The End of TCP for AI Clusters [video] on Sunday October 04, 2026 @06:05PM (#49958369)
I wonder if, as LLMs get accustomed to using homa or some other optimized protocol for, as the article lists, "chores such as weight gradients, model weights, KV cache entries, and checkpoints", whether we'll start to see TCP as a bottleneck for their post-trained interactions as well, for many of the reasons.
Re:I Quit OpenAI Because Its Culture Is Broken on Sunday October 04, 2026 @01:25PM (#49955877)
> That said, when the problem is at the level of "the government itself is breaking the law", you can reasonably ask if any regulation is even worth the paper it's written on.

Yeah, exactly, and ultimately I think that's really the thrust of the point I was making.

And, to me, if I was just looking at this calmly as a decision about what the obvious direction seems to be, given these factors, it's pretty straightforward: deprecate the nation-states. They are the ones mucking up the whole system.

If the thing we're really concerned about is LLM-safety wrt warfare and weapons, then I'd much rather tell the (whining, childish, seemingly headed for self-destruction anyway) nation-states that they have to sit this next era of humanity out than have to nerf them for the rest of us (and as you point out, nerf them in a way that the nation-states won't abide anyway).

Re:I Quit OpenAI Because Its Culture Is Broken on Sunday October 04, 2026 @08:13AM (#49953224)
So how about we keep the LLMs and get rid of the warfare and autonomous weapons systems?
As awful a practice as it is, I'd _much_ rather have an internet where its reform is prompted by competition than by the cops.
Came directly to comments hoping not to see this one.

<sad trombone sound>

Surely someone will soon do what the title of this post makes it seem like cloudfare did. Truly modular open source training and inference logic, along with a totally open corpus and weights, will eventually out-compete the closed ecosystem.

Re:OpenAI agents tried to bruteforce a UN website's API fields on Sunday September 27, 2026 @02:04AM (#49863794)
Hear hear.

It's really funny to see the delusion being defended so vigorously by people - presumably well-meaning people - purporting to defend the livelihoods of musicians and artists, while the musicians and artists are desperately trying to free themselves from the jaws of their IP agreements precisely so that their music can spread more easily.

I imagine this is already well-known on HN, but there is a significant movement underfoot in the worlds of bluegrass/old time/trad/jam toward DRM-free and CC licensing.

https://pickipedia.xyz/wiki/DRM-free

Re:Claude discovers a novel enzyme system with CRISPR-like repeats on Wednesday September 23, 2026 @03:09PM (#49820998)
> Even RSI will not be able to get new medical treatments to market instantly. Real world testing takes a long time and is an unavoidable part of the process.

Is it unavoidable, though?

I cut a one-take version of Frog with some of my fav. musicians:

https://www.youtube.com/watch?v=UCuDk18gGRc

> Disputed by who?

Yeah, I was gonna say, as conspiracy theories go, this one is pretty much in the books at this point. Fuck man, Tom Hanks stars in a movie about it; that's about as mainstream as it gets.

Notably, the US is on precisely the opposite side of the conflict in 2026 compared to 2006.

To put it differently: its stated allies are direct lineal political and military descendants of the parties who were ostensibly its enemies in 2006, including Al-Qaeda / Jabhat Al-Nusra.

Re:An Update on Wayback Machine Access on Tuesday September 15, 2026 @07:32PM (#49720284)
Seems like a reasonable norm:

* If you train AI on it, you have to afford public access to it.

* Nobody can exact violence against anybody else in response to that person providing public access to any data anymore (ie, all bytestrings are public domain).

That's the world I'd like to try in the coming years.

Older »