lumpa

One interesting possibility I've seen discussed[0] is enhancing the static analysis tools to find bug shapes that LLMs originally discover. There's no doubt static analysis tools are faster and cheaper, so they scale better to huge codebases and may allow generalizing the LLM's methods.

[0] https://lwn.net/Articles/1068968/

Here's some evidence that he is, again, badmouthing the PSF without a good reason: https://pyfound.blogspot.com/2025/06/the-psfs-2024-annual-im...

It is misinformation because it elides the context: white guys got virtually all the mentoring already, represented the absolute majority of core developers, held all the power in steering the language, etc.

So what he's saying is: my personal efforts focus on helping those that currently don't get help, don't have a voice. White man are doing fine already, they can forget getting direct help from me because they, as a group, don't need it.

So yeah, you can be upset that he's not offering to help those who, as a group, already have a path to getting into Python. Says a lot more about you though.

    Linux people get really worked up when I say "platform-native". There is no such thing on Linux, but reasonable people agree that something like a GTK app (or Qt) feels "native" on *most desktops* over other applications.

https://mitchellh.com/writing/ghostty-gtk-rewrite#user-conte...

And still an important voice in the Python community: https://discuss.python.org/u/tim.one/summary

I really hope Tim being in the documentary is the plot twist. He deserves a big chunk of time too.

This is completely bonkers.

The Python community is welcoming, many come for the language and stay for the community. It's not, of course, free of politics or drama, but it's very far from what you describe. Local communities are very strong, CPython core community seems to always be trying to improve to me.

Even Tim Peters, who I really hope is part of the documentary, is an enthusiastic participant, both helping with gnarly CPython issues and providing assistance to newbies.

If you look at the Fellows list[0], you can see that many important names aren't active in the community anymore (I don't know the reasons for each one), but many more are either active or in (very) good terms with the community.

The CoC was and is a net positive, the diversity efforts even more so. Last Saturday I was at a local Python conference and the local community has welcome both, to great success and improvement.

[0] https://www.python.org/psf/fellows-roster/

"There's a huge evil magnet that will tear you apart if you have any metal on you" sounds much easier to grasp and less likely to lose the listener's attention. Then, when you have them listening: "It can grab you from outside the room and hurl you into the machine where the evil magnet lives! Any metal, be it coins, necklaces, pins in your bones, belt buckles, bra wiring, dog tags. Anything can be the end of you, be damn sure you don't have any metal on you."

Oh, wait, you still want them willing to go near the machine? That complicates things a bit ;)

Homo sapiens are apes and are also virtually identical in physiology compared to other apes if the external comparison group is dogs or cats.

I has officially been released: https://blog.python.org/2024/10/python-3130-final-released.h...

Is this the technique they are using in Python's new JIT?

There are performance options[0] that enable PGO, LTO and bolt, and are known to speed things up, maybe more significantly than -march and -mtune.

But I found one report[1] that says these compiler flags do (or did, some time ago) make a difference.

[0] https://docs.python.org/3/using/configure.html#performance-o...

[1] https://atleastfornow.net/posts/py3-enable-optimisations/

Can you read the refutation at all? If so, would you mind pasting it here so us without a Twitter account can also read it?

There's an interesting bit about being unable to recover the timelocked secret if the parties that keep the network going disband before the release date:

    *if the League of Entropy shuts down, members will delete their keys*

    The world is unpredictable (just like drand randomness... heh), and it's possible that the League of Entropy will all hang up their coats at some point in the future. Should that happen, members would have two choices with their private keys: release them to the world or delete them entirely. The former would mean that ciphertexts created for some time after the cessation of the network would be decryptable to everybody. The latter would mean that ciphertexts created for some time after the cessation of the network would be unencryptable forever (/until quantum computers can break them). In the interests of privacy, we felt the latter option was preferable. That said, if you encrypt the private key to your [insert cryptocurrency name] fortune to stop yourself from spending it now and the network stops... you're going to have a bad time.

I have fond memories of following the messy developments on Groklaw.

That blog instilled me with hope that we'd see more and more amazing works from all sorts of experts on the web, distilling speciallized topics that would otherwise be pretty incomprehensible (or worse: shallowly editorialized) into thrilling but educational content.

Alas, I've never found another site that so consistently and deeply explained its subject matter from an expert perpective in language non-experts could follow. If you have examples of such content, I'd be very interested in pointers to them.

The nets are much more successful in capturing all sorts of birds at once, with little effort, even in environments where you wouldn't get clear line of sight for photos and/or for species that don't sit still for long enough.

Handling the birds makes for far better identification (and more detailed pictures too), which is important for places where unknown, hard to identify or hard to detect species are likely to exist. It also allows taking measurements, banding the birds and even collecting tissue samples. So nets are a far better ROI for some scientific projects.

There's a lot of effort going on to improve CPython performance, with optimization tiers, etc. It seems the JIT is how at least part of that effort will materialize: https://github.com/python/cpython/issues/113710

> We're getting a JIT. Now it's time to optimize the traces to pass them to the JIT.

PyPy is a Python implementation, not the same the PyPI (Python Package Index).

Your comment was dead, as are all your comments and submissions. Probably another case of a spam filter gone bad. As dang says[1], email hn@ycombinator.com and they'll solve it.

[1] https://news.ycombinator.com/item?id=36577441

Nice to see it added, quite a lot of people found out they had fat finger flagged some content when I told HN about these (then) hidden pages: https://news.ycombinator.com/item?id=30253799

Edit: as stated in that post, you can also see your vouched items by accessing https://news.ycombinator.com/vouched

The full book is available as a PDF at the bottom of the page.

Your post was dead, as are all your comments and submissions it seems. Since I didn't find any reason for that from a cursory look, I've vouched here. Maybe ask dang why and how to avoid that?

Yep, it's https://news.ycombinator.com/hidden and it shows in your profile.

Here's an interesting article that has more details about it: How I hacked Google’s bug tracking system itself for $15,600 in bounties by Alex Birsan https://www.freecodecamp.org/news/messing-with-the-google-bu...

Amazing how such a simple approach can achieve content injection on a diverse network of unrelated websites, to the point of raising the profile of the vector and increasing the chances of further spread.

I hope someone figures out which other campaigns were run with these tools. Also, whether you can find output with the link injections in source code, like on GitHub or distro packages.

You can go into your profile, then set "showdead" to "yes". Once you look at bellingcat's domain submissions, you'll see they're mostly "dead" and that makes them hidden from unlogged and most logged users, so it's like they're automatically taken out of the pool of articles available to receive votes. Only once someone "vouches" for a dead post can it be voted on again.

Usually posts go dead when downvoted enough, but it seems posts from that domain are automatically set to dead, and a few of them are resurrected.

Btw, the best place to find dead posts is the newest page, as they are uncommon in other places.

Edit: sorry if I sounded condescending, I might have misread that as you not understanding what "dead" meant, rather than "why would that be an issue".

> Also, I doubt this is a hnews concern, but for folks currently holding clearances just reading that story could put them in hot water.

How would that work for folks holding clearances that, following one of several paths other than directly reading the article discussion (like reading some random user's comments page), get exposed to significant bits of the article?

Come to think of it, how safe in general for folks with clearances is browsing HN, with its many discussions of military tech, intelligence topics, spycraft and security breaches news?

Ah, thanks, the simplest explanation didn't occur to me: all articles from domain start as dead and a few are vouched. So it seems like a normal outcome for an article a dozen of HN users would submit from such a domain.

I remember using something like this in Firefox a long time ago. A search suggests it was Optimoz/RadialContext: http://www.mozillazine.org/talkback.html?article=2407

There's at least a current attempt at the same idea, posting to test it later: https://addons.mozilla.org/en-US/firefox/addon/easygestures-...