> It’s simple intelligence, which is what we can do today and make reliable. It’s not sexy. It’s technology in the service of making things easier for workers and more efficient.
> We didn’t have Internet in the office back then so research was a bit of a problem. Making things up was the way to go, nothing like Google to make life easy back then. I remember asking Gary if I could go the library to photocopy some pictures from books to help me out! An artist back then had to know way more than an artist (who tends to get pigeonholed) these days. There was no such thing as an environment artist back then...so we did everything. Game artwork, interface (I was font boy for months!), cut scenes, animation. If we didn’t know how to do it we just learnt as we went along. Quite mad really but it seemed to work as production values were so much lower back then.
It's hosted on one of the author's sites. The collection itself is (as far as I can tell) out of print. It's falling through the cracks of "too complicated for a publisher to figure the rights out of" and "not lucrative enough for anyone to care".
It’s possible that it’s being distributed with permission of the copyright holders. Given the number of different people involved that seems kind of unlikely, but “free” doesn’t have to imply a permissive license.
I think it's normal for the publisher to hold those rights (perhaps shared with the original authors, depending on the details of their agreements), so possibly all that would have been required here would be for the publisher to approve doing this.
Or maybe Rucker and all of the other authors are friends, and keep in touch, and he just literally called all of the up and said "Hey, can I post Mirrorshades online for posterity?" and they all agreed. Who knows?
Not by default, no. But it seems entirely reasonable that he may have approached the original publisher, requested permission to post this, and received said permission. Considering that the print book has been out of print for some time, and given that the linked page does emphasize the copyright status of the works, this feels like the most likely scenario to me.
He drew hard on his cigarette. Annoyance flickered across his face, like an artefact in the poorly-compressed bootleg movies he sold to his fellow low-lifes at The Pig and Drum.
Some Corpo-type, no doubt. Can't help seeing something good scroll across their feed tube without calling Legal.
He'd worked with a few in the past. Not bad all-in-all, at least they paid on time. That said, he could think of few he'd drink with.
He toyed with the idea of leaving a bitchy comment. Probably get downvoted to oblivion.
The dogs in the yard barked at a passing vehicle.
Irritated by the animal noise and the corpo whining, he thrashed something out. Pulling another cigarette from his pack, he hit "reply".
I wouldn't describe him that way but, at any rate, his point is right: pirating books is taking money out of authors' pockets. Corporations (publishers?) aren't hit hardest and don't care the most about it.
You may prefer the same point expressed in less colorful language by Ursula K. LeGuin, from the same article as Ellison's quote: “I thought, who do these people think they are? Why do they think they can violate my copyright and get away with it?”
https://www.nytimes.com/2009/05/12/technology/internet/12dig...
Now, as we found out in the meantime from Rudy Rucker's comment, this anthology isn't pirated after all. If you put up your work for free, there's no piracy.
I have no issue with Ellison's colorful language, but the point expressed by LeGuin is quite different. Copyright violation is not theft. It's not even a criminal charge in the US unless done for profit! You know what also deprives authors of royalties? Borrowing books from public libraries, buying used books, and loaning books to friends. So does playing video games instead of reading a book! The fact that an action yields less money to an author does not in-and-of-itself make it theft, or even immoral.
Copyright was chartered to encourage authors to contribute to a large public domain of works. Lobbying (by wealthy corporations) perverted this purpose. Sure it's a nice fringe benefit that some authors were made more comfortable by this. That doesn't stop many authors from taking a "I wrote it, it's mine" attitude as if a monopoly on the use of works you authored is a natural right.
LeGuin has taken a nuanced view on this, with the apparent understanding that copyright is a framework under which she was promised certain things, and the piracy is a violation of that promise.
> That doesn't stop many authors from taking a "I wrote it, it's mine" attitude as if a monopoly on the use of works you authored is a natural right.
I think that monopoly (with various caveats, e.g. it can't outlive you much, etc.) is a good thing for authors to have, as it enables them to make a living off of their writing. Authors weren't just "made more comfortable" by this as a fringe benefit, as you say, but really, they were able to make a living from their creative work. Harlan Ellison himself says so in that article, and there are countless instances of up-and-coming writers fighting piracy (one legendary story is how Tolkien fought pirating of LOTR in the U.S. soon after it was printed in the UK).
Also, I don't see how LeGuin's point is substantively different from Ellison's -- they are both saying they'll fight people who distribute their books without paying them, the author.
On that note, this argument:
> You know what also deprives authors of royalties? Borrowing books from public libraries, buying used books, and loaning books to friends. So does playing video games instead of reading a book!
... is partly false -- authors do get payouts from libraries. As for "playing video games instead of reading a book", that's absurd -- the problem with pirating is that you get for free something that the creator has produced. For your argument to be true, we would somehow have to assume that the creator is entitled to us spending time reading their books, which is obviously insane.
As for the other things you mention -- buying used books and loaning them from friends -- they have essentially no overlap with online piracy. Piracy is a problem because you can distribute infinite copies worldwide for free, which doesn't apply to selling or loaning physical books.
> Authors weren't just "made more comfortable" by this as a fringe benefit, as you say, but really, they were able to make a living from their creative work
By "made more comfortable" I was not referring to the existence of copyright at all, but rather the multiple extensions that were made from 1976 to 1998, where copyright terms went from 56 years to over 95 years[1].
If the Ace paperback edition of LoTR was piracy, then I question the meaning of the term, since the original US publisher imported British editions which lacked the (then required) US copyright notice. Note also that Ace ceased publishing this edition (and paid Tolkein) due to public pressure, not any legal threats.
(Also lest I misrepresent myself, there were many good changes to copyright in 1976, including removing the notice requirement that caused Tolkein so much trouble).
1: Prior to 1976 the lifetime of the author did not involve in the calculation, and literature is one place where works-for-hire are still rare this is more complicated than just 39 years longer; nevertheless 70 years from the (last in the case of multiple) author's death is always more protection than 56 years, and may be considerably more for a young author. This also reinforces my point that media corporations (where work-for-hire is the norm) benefited from this rather more than authors.
OK. I don't really have an opinion on the length of copyright, and I could definitely be persuaded that 70 years after death is too long -- prima facie, it looks too long. But we were talking about the permissibility of online piracy and the validity of copyright per se, not the right length of copyright.
It's free to read, not free to use. As it's from one of the involved authors, they probably got permission for this release. The problem with piracy is lack of permission/consent, not the act itself.
People are making books freely available all the time, even those they sell on other platforms. Nothing wrong with this.
Consider the popular cliche, "free as in beer vs free as in speech".
The rights copyright gives you, briefly, includes: copy, distribute, adapt, display, and perform a creative work. What suggests there is piracy is going on?
In 1986, it was unlikely that the original contract for the book mentioned anything about electronic rights. As it was a reprint anthology, the rights purchased would have only covered the use in the anthology as long as it was in print. Which means that to post the book online, Rucker would have to contact the individual contributors and get new permissions. Did he do that? I make a guess that he did not. It is not clear from what is stated here.
It's not intuitively clear to me how the game can go on forever -- I would expect that, eventually, you would hit upon some valid pattern. The explanation in the text didn't really make sense to me. Could anyone help with this?
The game could use a better definition of what constitutes a pattern.
>> A pattern is a sequence of pebbles repeated three times in a row.
By that logic, player 2 would have lost at their fifth turn. ... And 'in a row' is open to interpretation since it is additional information to 'repeated three times'.
I think it's pretty clear that a string s of pebbles is a pattern iff there exist strings x and y such that s = x|y|y|y, where | denotes string concatenation and y is nonempty.
This is interesting but, speaking frankly, I see many seemingly insurmountable issues. Here are some:
- Contests will often be won not by the entry that best adhered to the prompt, but the best-looking one. This happened in the contest "Input Prompt
Build a brutalist website to a typeface maker," which I got as a recent example. The winning entry had megawatt-bright magenta and yellow, which shouldn't appear anywhere near brutalism, and in other design aspects had almost no connection to brutalism either -- but it was the most attractive of the bunch.
- The approach only gets you to a local maximum. Current LLMs aren't very good designers, as you say, so contests will involve picking between mostly middling entries. You'd want a design that's, say, a 9 or a 10 on a 10-point scale -- but some 95% of the entry distribution will probably be between 5.5 and 7.5 or so, and that's what users will get to pick from.
All great points. A limitation with human feedback is that once you start asking for more than binary preferences (e.g. multiple rankings or written feedback), the quality of the feedback does decrease. For instance, many times humans can give a quick answer on preference, but when asked "why" they prefer one thing over the other, they might not be able to full explain it in language. This in general is very much an open area of research on collecting and incorporating the most optimal types of feedback.
I definitely agree with your second point. One idea we're experimenting with is adding a human baseline, in which the models are benchmarked against human generated designs as well.
yes! to the second point, someone in our show HN proposed encouraging human designers to compete in submissions as well - we tried implementing this and found that, at least right now, LLMs are still so bad at design that asking a human to beat them is trivial - our plan right now is to focus more on this once it becomes more of challenges and therefore hopefully more interesting/entertaining
> Daily puzzles aren’t just engaging—they’re efficient, scalable, and well-aligned with key product and business goals.
Pure, grade-A LLMese. Seeing this makes me have to summon all my strength to skim to the end (having given up reading), which I did in this case because I found the subject interesting and hoped there might be something thought-provoking in the article.
> Daily puzzles are engaging, efficient, scalable, and well-aligned with key product and business goals.
bring it back within the realm of human-generated PR text? Or it's too perfect? I find the perfect number of syllables to be off putting sometimes, it can feel like the uncanny valley of text.
> Each replication task consists of a detailed specification and a reference implementation. The central idea is that AI models are trained to produce an implementation that precisely matches the reference behavior. This clear-cut approach significantly simplifies evaluation, as the grading criteria are objective and direct: either the generated implementation behaves identically to the reference, or it doesn’t.
OK, but then you have to produce the detailed specification, working backward from the reference implementation. This is extremely non-trivial and it significantly weakens the TFA's parallels to pre-training, in which you don't need really need inputs other than raw text corpora.
I'm not saying this eliminates the idea outright, but I do think it hobbles it badly.
I’d like to courteously disagree. I think existing models and existing tools are good enough to bootstrap this at least.
I’d propose the following architecture:
Step 1: Microsoft phi style - read code and write specifications using a frontier model. You could use an ensemble here to nitpick the spec; it’s only going to get written once. We also have of course many many rfcs and codebases that conform to them or where they do not we have an existing repository of bug reports, patches, forum complaints, etc.
Step 2-4: implement multilayer evaluation: does it compile? Does an existing model think the code complies with the spec on inspection? When it’s run on qemu are the key evals the same as the original software?
I propose most of steps 2-4 are automatable and rely on existing tooling and provide a framework that is, if not cheap, achievable. I’m also sure someone could improve this plan with a few more minutes of thought.
To me the interesting question is - will this add capabilities at current model sizes? My prior is yes in that the current behemoth size models feel like they are only incrementally better than 1/10 size distills. I interpret that to mean we haven’t gotten the most out of these larger scales. I will note Dario disagrees on this - he’s publicly said we need at least 10x more scale than we have now.
When prompted correctly, models could generate good specification in form of pretty exhaustive tests. While all tests have weaknesses and are not formal specification, they could get us 99% there.
The author's point seems to be: "It's not that they don't know it's bad. They know, and they do it anyway. The fact that they know shows they're not stupid."
Umm, OK? I still don't get why, in the author's view, they do the bad thing on purpose, and why that is not stupid. (Perhaps the author might make say that doing a stupid thing doesn't make someone a stupid person, or something to that effect -- but, even if so, I don't see any sign of that argument in TFA.)
> The first effect of teaching a methodology —rather than disseminating knowledge— is that of enhancing the capacities of the already capable, thus magnifying the difference in intelligence.
Absolutely right, with the implication that new capabilities available suddenly to everyone often end up making the playing field more unequal, not less.
Perfection.