[kragen]

Other authors of research and I are the ones demanding this. We're the ones giving the gift horse in the first place. People who don't play a part in creating human knowledge generally aren't interested in reading papers about how hard-to-use software that's no longer available worked on obsolete computers they don't have access to, especially when the problems that software solved are problems they don't have.

I'm not demanding that the ACM do more. I'm demanding that they do less, by renouncing their right to sue other people for legally archive and redistribute ACM papers, so the ACM don't bear the full responsibility of doing so themselves. That way, I can do more of that wealth-of-knowledge-creating stuff you're so excited about, benefiting the ACM's members. It's a win-win.

It really isn't very likely that anyone will ever publish a computer science paper as impactful as Dijkstra's go-to-statement thing. That affects how we write literally every line of code in every language today except maybe assembly. Maybe one of the LLM papers might compete?

On a different note, it seems like you mostly post comments on HN in order to personally attack other commenters, as you are doing here, and to advocate political positions. That isn't what the site is for. If you keep doing it, they're going to ban you.

[newswasboring]

> It really isn't very likely that anyone will ever publish a computer science paper as impactful as Dijkstra's go-to-statement thing.

Ok, disclaimer that I am not a computer scientist (work in semiconductors so only tangentially related). But, this statement has the same "end of history" energy has the famous Philipp von Jolly quote about end of theoretical physics:

"In this field, almost everything is already discovered, and all that remains is to fill a few holes."

I'm not claiming you are saying its end of CS, just the claim that there cannot be a new paradigm discovered in CS doesn't sit right with me.

[kragen]

I think there's an enormous amount that can still be discovered, including new paradigms. I don't agree with Ken Thompson's opinion that people studying informatics today are unlucky because the most interesting stuff is already done.

But I don't think it's especially controversial to claim that Galileo and Newton had more of an impact on physics than Maxwell and Einstein or than anyone since. You could maybe quibble about Gauss and Lagrange, but Kip Thorne and Ed Witten are much more similar to Galileo than Galileo was to Descartes or Aristotle.

You might be able to cause an Einstein-like revolution in informatics—LLMs in particular seem like they have a good chance of doing that. But the field those new paradigms revolutionize will probably be recognizably the field that was largely defined by papers published in CACM in the 60s and 70s.

Also, although this isn't relevant to my thesis that probably nobody will publish such an impactful paper again, the ACM is especially unlikely to. "Attention is All You Need" https://proceedings.neurips.cc/paper_files/paper/2017/file/3... got published in NIPS 2017 rather than CACM or even an ACM conference. You could imagine a timeline where CACM was the Cell or Lancet of informatics and published papers like AiAYN instead of "The Transformative Power of Inspiration". But that's not the one we're in.

[bpt3]

> Other authors of research and I are the ones demanding this. We're the ones giving the gift horse in the first place. People who don't play a part in creating human knowledge generally aren't interested in reading papers about how hard-to-use software that's no longer available worked on obsolete computers they don't have access to, especially when the problems that software solved are problems they don't have.

You (and I) are free to publish in venues that meet our requirements.

> I'm not demanding that the ACM do more. I'm demanding that they do less, by renouncing their right to sue other people for legally archive and redistribute ACM papers, so the ACM don't bear the full responsibility of doing so themselves. That way, I can do more of that wealth-of-knowledge-creating stuff you're so excited about, benefiting the ACM's members. It's a win-win.

I am not at all worried about this, and there's no real reason for you to be either (the odds of the ACM library vanishing is almost 0), so it seems like you're being needlessly hostile.

> It really isn't very likely that anyone will ever publish a computer science paper as impactful as Dijkstra's go-to-statement thing. That affects how we write literally every line of code in every language today except maybe assembly. Maybe one of the LLM papers might compete?

I'm sorry, but this is absurd. "Attention is all you need" comes to mind as a recent example of a highly impactful paper (not published in an ACM venue, but you're now expanding your claim to the entire field of CS).

> On a different note, it seems like you mostly post comments on HN in order to personally attack other commenters, as you are doing here, and to advocate political positions. That isn't what the site is for. If you keep doing it, they're going to ban you.

You're a real peach.

[kragen]

The odds of the ACM library vanishing are the same as the odds that you, personally, are going to die: at least 1000:1 in favor.† The only question is whether it happens before the relevant copyrights expire. Anyone who can't convince CloudFlare they're human has already lost access to the ACM library.

I've already addressed your "You are free to publish" argument.

I'm being hostile because your comment, in addition to being factually incorrect in a way that demonstrates your complete unfamiliarity with the subject matter, consisted almost entirely of personal attacks on me. You accused me of "looking a gift horse in the mouth", of playing "no part" in "creating" "human knowledge", and "demanding more for nothing". Now you're implying you thought that was friendly rhetoric? Do you expect anyone to believe that? How stupid do you think other people are?

I already gave "Attention is all you need" as a recent example of a possibly highly impactful paper in https://news.ycombinator.com/item?id=44734480. The fact that it wasn't published in an ACM venue is one of the reasons that the ACM's policy on new papers is relatively inconsequential compared to their policy on existing papers.

______

† I'd say 1:0 in favor, but rationally speaking, we can't completely exclude the possibility that all of this is some sort of hallucination or simulation, or that time will stop tomorrow so that everything that exists at that time will never vanish, and so on. But, under the usual presumptions that the universe is objectively real and everything in it vanishes sooner or later, the ACM Digital Library is absolutely guaranteed to vanish. And if you think it's inconceivable that it will be destroyed by political machinations within a few years, did you predict two years ago that the US would vote in favor of Russia invading Ukraine in the UN?

[DonHopkins]

"Attention Considered Harmful"

"GOTO Is All You Need"

[kragen]

That last one sounds like Scheme. Or Levien's Io.

The former sounds like a LessWrong fanfic.