pron

20 ms·

241.
by pron 5mo ago
> Anthropic can always fire the Opus/Mythos token machine gun on any problem (bugs, features, security) to ensure PR success, Can they, though? They tried and failed to do it in their C compiler experiment. The experimenter wrote: &
242.
by pron 5mo ago
> where are we seeing that it failed? Anthropic said the experiment failed to produce a workable C compiler: - I tried (hard!) to fix several of the above limitations but wasn’t fully successful. New features and bugfixes frequently brok
243.
by pron 5mo ago
Yes, the task is very different, but also it will be months to a year until we know the results of the bun experiment.
244.
by pron 5mo ago
A workable C compiler is a ~10-50KLOC program, and a fairly simple one at that (batch, with no concurrency or interaction). That Anthropic's swarm of agents wrote 100KLOC before failing is a symptom of the problem. It's certainl
245.
by pron 5mo ago
> It's 2026 and the notion that you can get detailed enough requirements and specifications that you can one-shot a perfect solution needs to die. It's 2026 and the idea that even with detailed-enough requirements you can one-s
246.
by pron 5mo ago
There is another problem. For developers, productivity means "functionality produced per hour of work", but that's not what productivity means for businesses. To them, productivity means "money produced per hour of work&
247.
by pron 5mo ago
The question of how intellectual ideas gain political influence is always a tough one. John Ganz tries to trace right-wing ideologies and their influence at different times and places (mostly America, France, Italy, and Germany), from the T
248.
by pron 5mo ago
I have no idea if the accusation levelled toward Meta is true or not, but I think the claim made by the article is both logical and clear: technofascism seeks to weaken democratic institutions, and to do so it exploits existing nationalist
249.
by pron 5mo ago
That's what I'm saying. I'm glad to see that he's found a subject where his lack of knowledge isn't a glaring handicap. When I come across his posts I usually feel uncomfortable because they read like a bright 4-yea
250.
by pron 5mo ago
1. Scott Alexander is famous for writing about topics he knows little about. I'm glad to see he's found a subject he knows little about but so does everyone else. 2. What's even worse than predicting that some growth curve fl
251.
by pron 5mo ago
I don't think our experience is different. Letting the agent work on pieces no bigger than a couple hundred lines at a time and checking if there's something fishy or not and that the code is legible and logical is close human s
252.
by pron 5mo ago
What exactly of what I wrote is untrue, and while I'm not familiar with Shenandoah, what does decommitting have to do with any of it?
253.
by pron 5mo ago
Yes, on any hardware that ZGC supports and that your program otherwise runs with acceptable performance you won't get a long pause time, including (hypothetically) Pentium 3. The reason is that there is no work done inside the pause
254.
by pron 5mo ago
My point is that the website offers a very low coverage of the codepaths that affect language performance, and all it says is that the coverage is incomplete without saying just how incomplete it is. People who don't know that, for exa
255.
by pron 5mo ago
> GC is only responsible for saying to the memory allocator "this object is no longer used" This is simply not true. Not only are Java's GCs responsible for allocation (which they do simply by bumping a pointer, similar to
256.
by pron 5mo ago
> Is it possible to detect cyclic references perfectly? Yes. The GCs in Java, .NET, V8, and Go do it. > If we have 7 and 0.1% of the time you need one of the 6 that is non-default, how do we choose? Java's GC are optimised for di
257.
by pron 5mo ago
> Lately, they seems to work with CRIU, various heuristics, multi-stage in-process bytecode compilation .. Not sure what you mean by this, as this has nothing to do with GC, and Java has had a multi-tier optimising compiler for 15 years
258.
by pron 5mo ago
https://openjdk.org/jeps/8377305
259.
by pron 5mo ago
The new generational ZGC? I'm sceptical.
260.
by pron 5mo ago
> while it's easy to do harm (lead exposure) it's all but impossible to do any good. That's just a meaningless statement no different from "while it's easy to subtract negative numbers, it's all but impossib
261.
by pron 5mo ago
Java lost almost all those knobs a while ago (I mean they're there, but you're better off relying on the defaults). The modern GCs have one or at most two knobs remaining, and even that will become unnecessary next year. As to pre
262.
by pron 5mo ago
> There are other ways to achieve this goal though, that don’t require exhaustive human review, better models are able to do that part as well if properly guided. Not at this time. Even if you could somehow get their success rate to 90%,
263.
by pron 5mo ago
Down the line the agent is no longer able to fix one failure without causing another and the codebase is unsalvageable, but you may not have reached that point yet. Agents can help a lot when you carefully review everything they output and
264.
by pron 5mo ago
That's perfectly fine, and I totally understand people who don't want to sit and wait 70ms for their script to finish running (that 70ms is not the time it takes to start), but let's not turn a <40ms startup into 1.5s. Now
265.
by pron 5mo ago
Java runs a Hello World, cold, in a packaged JAR, in about 40ms. What you've seen isn't JVM startup but programs that do a lot at initialisation (like MS Word), as many Java programs like to do (because they often expect to run fo
266.
by pron 5mo ago
Well, these days a small CLI program in Java (say, ls) starts up cold, runs, and terminates in ~70ms, not 1500ms, but yeah, sometimes 70ms is too long to wait for a script.
267.
by pron 5mo ago
> How do you define "bad code"? Code that will not be able to evolve for more than one-two years is terrible code. Agents write terrible code while doing a truly impressive job hiding it (including in the tests they write) unle
268.
by pron 5mo ago
Well, the AI's judgment (i.e. if you accept it) leads to a codebase that cannot handle evolution for more than 18-24 months or thereabouts. If you bother to look you can literally see it rotting at 5x speed (all while passing all tests
269.
by pron 5mo ago
They are nowhere near solved. Agents make serious mistakes in judgment and do it frequently enough to threaten the viability of the codebase unless you slow down and monitor them very, very closely . If you do that, it's all good. If
270.
by pron 5mo ago
Human developers don't produce code at such a rate, and their judgment is, on average, better. So one, the review doesn't make you feel like you're slowing things down much, and two, the problems are less hidden.

More ›