Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
twoodfin
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
24 ms
·
571.
▲
by
twoodfin
10mo ago
I feel like any information can’t be taken as fact, it can just be rolled into your world view and discarded if useful or not. The concern, I think, is that for many that “discard function” is not, “Is this information useful?”. Instead:
572.
▲
by
twoodfin
10mo ago
That’s unhelpfully reductive: There are zero companies making smartphones in the US, too.
573.
▲
by
twoodfin
10mo ago
If we’re arguing about strategic decisions made a decade ago in this context, aren’t we really arguing about whether any company should ever fail? This is Matt Stoller throwing a bunch of dust in the air because he wants to have been ri
574.
▲
by
twoodfin
10mo ago
The stocks are inflated because Wall Street believes current $$$$$ capital investments will create massive long-term value!!
575.
▲
by
twoodfin
10mo ago
What is the current AI/data center mega-boom if not forgoing short-term extraction for long-term value creation?
576.
▲
by
twoodfin
11mo ago
Especially strange and relevant here: VA Linux owned Slashdot. Depending on whether you were a Digg person—I was not—Slashdot is either hn’s conceptual grandfather or great-grandfather. “News for Nerds. Stuff that Matters.”
577.
▲
by
twoodfin
11mo ago
Yeah. But I will die on the hill that ChatGPT (today, at least) is a bad writer, and makes prompted writing worse in a way that isn't anything like the way schematic style or vocabulary rules might for an over-eager student. For what
578.
▲
by
twoodfin
11mo ago
The Robinson-Patman Act is terrible law. It’s been routinely violated (unknowingly in most cases) for decades across effectively every sector of the economy & enforced vanishingly rarely. If it were to be enforced uniformly and aggressi
579.
▲
by
twoodfin
11mo ago
They’re also a major investor in Rivian so there’s some circularity in that phenomenon.
580.
▲
by
twoodfin
11mo ago
I’m not arguing with that, it’s basic information theory. One bit, however, is not “the maximum possible amount” you can gain from an oracular answer to a yes/no question. The OP covers exactly this point re: the “Guess Who?” game.
581.
▲
by
twoodfin
11mo ago
If both of these are equally likely, you gain one bit of information, the maximum possible amount. If you already have other information about the situation, you might gain _less_ than one bit on average (because it confirms something you
582.
▲
by
twoodfin
11mo ago
The claim was that one bit was the maximum amount of information you could gain, which is clearly false. Just to make this unambiguous: If you ask me to guess a number between one and one billion, and by fantastic luck I guess right, your “
583.
▲
by
twoodfin
11mo ago
If I’m trying to guess a 9-letter English word, and test whether the first letter is “x”, there are only the same two answers: Yes/No. But “Yes” obviously gives me much more than one bit of the information I need to know the answer.
584.
▲
by
twoodfin
11mo ago
Sure, I believe that completely. But that's not how ChatGPT writes! Here are some random examples from one of the (at least) half-dozen LLM-co-written posts that rose high on the front page over the weekend: https://blog.c
585.
▲
by
twoodfin
11mo ago
Capital gains taxes are applied to capital gains income and are included in the stats above.
586.
▲
by
twoodfin
11mo ago
Are Kenyans really taught to write that way? I’m highly skeptical. At one point the author tries to argue this local pedagogy is downstream of “The Queen’s English” & British imperial tradition, but modern LLM-speak is a couple orders
587.
▲
by
twoodfin
11mo ago
ChatGPT does not “write well” unless your standard is some set of statistical distributions for vocabulary, sentence length, phrase structure, … Writing well is about communicating ideas effectively to other humans. To be fair, throughout l
588.
▲
by
twoodfin
11mo ago
The whole site is AI slop.
589.
▲
by
twoodfin
11mo ago
This is exactly Baumol: If by “high quality” you mean “difficult to mass produce” then yes, the lack of productivity gains in hand-made furniture makes real costs go up. Of course, it’s easy to mass produce sturdy furniture, such as office
590.
▲
by
twoodfin
11mo ago
The front page this weekend has been full of this stuff. If there’s a hint of clickbait about the title, it’s almost a forgone conclusion you’ll see all the other LLM tics, too. These do not make the writing better! They obscure whatever
591.
▲
by
twoodfin
11mo ago
My “favorite” version of the last issue is trying to acknowledge with “K” which inevitably becomes “I”.
592.
▲
by
twoodfin
11mo ago
That design introduces two kinds of overhead at runtime: - You need a pointer to the allocator (presumably you’d want to leave room for types beyond arenas). That’s 8 bytes of extra size on every object. - You need to dynamically dispatch o
593.
▲
by
twoodfin
11mo ago
Reminds me of a game-breaking strategy in the (interesting, flawed) hybrid RTS/RPG War in Middle Earth (1988). The RTS part involved moving armies and heroes around to fight Sauron / Saruman’s armies and defend your citadels. Th
594.
▲
by
twoodfin
11mo ago
I’m much more optimistic on device-side matmul. There’s just so much of it in aggregate and the marginal cost is so low especially since you need to drive fancy graphics to the screen anyway. Somebody will figure out how to use it—compl
595.
▲
by
twoodfin
11mo ago
The humor is hit or miss but when it hits it’s quite funny, and the misses are merely groan-worthy. Triggered by the lighthearted tone of the prompt, I’d bet, but still quite impressive relative to most LLM-generated jokes I’ve had the misf
596.
▲
by
twoodfin
11mo ago
The USSR’s political system collapsed fairly suddenly. Its economic system had been rotten for decades.
597.
▲
by
twoodfin
11mo ago
Forget about DRAM vs. SRAM or whatever: How does a cheaper source of non-Nvidia GPUs help Oracle? They’re not training models or even directly in the inference business. Their pitch is cloud infra for AI, and today that means CUDA & Nvi
598.
▲
by
twoodfin
11mo ago
They also aimed at what turned out to be the wrong target: When Itanium was conceived, high-performance CPUs were for technical applications like CAD and physics simulation. Raw floating point throughput was what mattered. And Itanium ended
599.
▲
by
twoodfin
11mo ago
Income reporting is not the problem: Anyone paying you any significant amount of money is required to file with the IRS, including if you’re paying yourself. The issue is the broad range of deductions and credits that depend on things like
600.
▲
by
twoodfin
11mo ago
Perl did have `use strict`, so there was at least some plausible path to making non-breaking changes under a new pragma. The OP’s theory that Perl 6’s radicalism allowed Perl 5 to be conservative sounds right to me.
More ›