Contrary to the case for the internet, there is a way out, however - if local, open-source LLMs get good. I really hope they do, because enshittification does seem unavoidable if we depend on commercial offerings.
Well the "solution" for that will be the GPU vendors focusing solely on B2B sales because it's more profitable, therefore keeping GPUs out of the hands of average consumers. There's leaks suggesting that nVidia will gradually hike the prices of their 5090 cards from $2000 to $5000 due to RAM price increases ( https://wccftech.com/geforce-rtx-5090-prices-to-soar-to-5000... ). At that point, why even bother with the R&D for newer consumer cards when you know that barely anyone will be able to afford them?
When I was a kid in the 80s scooters weren't a thing in my country, I never saw one, but they did appear in the comics I read (from the 50s-60s). I remember asking my parents about them, and they telling me that they were toys from their time but no longer existed.
Now of course they're very common, my son has one.
> We haven't chosen quantity over quality, we have decided that journals should not be the arbiters of quality.
In CS, this is definitely not the case at all.
If you remove the "quality badge" factor, journals are totally useless. Everyone in my field knows how to use LaTeX, produce a decent-looking PDF and upload it to arXiv. This saves you from paying APC's, has actually better discoverability (everyone checks arXiv as a one-stop shop for papers, almost no one goes to check the dozens of different journals) and much less hassle (no need to fiddle with arcane templates, idiosyncratic paper structures forced by each journal, idiosyncratic submission systems that look straight from the 90s, typesetters that introduce more errors than they fix, etc.).
I am pretty sure that journals, at least in my field, subsist precisely as arbiters of quality, they don't provide any other value at all.
I assume uploading to arXiv doesn't count as having published a peer reviewed journal article, which is a problem for professionals.
For example, for me to progress in my current job I either need a doctorate or to have published a number of peer-reviewed articles in recognised journals as first author. I have written two IETF RFCs and these count for nothing.
I am not a scientist, I am a software developer. I am not employed as a scientist, I am employed as a software developer. But the rules of the organisation are thus.
> I assume uploading to arXiv doesn't count as having published a peer reviewed journal article, which is a problem for professionals.
Yes, in fact this is mainly what I meant with "quality badge". It's a badge mostly for instutitional bean-counting processes. Fellow scientists don't need it that much, typically we can separate the wheat from the chaff with a very quick skim.
Maybe it's time to do a Eurovision style thing for the quality badge. Everyone uploads to Arxiv. Every who's in the field votes on the worthiest papers (not allowed to vote for anyone you actually collaborated with).
Winners get to put a shiny sticker on their papers.
Semantic Scholar is for search, but you can't just go there and look at everything that has been uploaded today as you do in arXiv, right? I know many people who check arXiv every day (myself included) but not Semantic Scholar, although I guess this might be highly field-specific.
What follows is totally offtopic, but to be honest I don't check Semantic Scholar much because I have a grudge with it. Profiles just don't work for authors with accented characters in the name (such as myself), papers get dispersed between multiple automatically-generated profiles. The staff is very helpful and will manually merge profiles for me when asked, but then I publish a new paper and wham, instead of incorporating it into the merged profile the system creates a new one. This has been going on for 6 years (if not more) and still unfixed.
For all the criticism that Google Scholar gets, I highly prefer it because it gets that right. It's extremely annoying when tools give you extra work for committing the sin of not having an Anglo-Saxon name (this is much more common than unaffected people would expect) and just don't seem to care to fix it.
In my field, arXiv (free preprint server) is actually much more discoverable than journals. It tends to be on top of Google searchers, many people (myself include) check it out daily, and few people even check journals (why would you check dozens of different ones if everyone posts their work on arXiv?).
Ev Fedorenko is a highly recognized cognitive scientist that has been studying how humans parse language for years.
Of course this doesn't mean one shouldn't question what she says (that would be an obvious authority fallacy), but I do think it's fair to say that if you want to question it, the argument should be more elaborate that "this sounds like she has no idea of the topic".
I'm not the person you responded to, but I found the article unreadable because it kept going on about Ev’s life instead of her research. I'm sure her research is valuable and insightful, but with this style of reporting it is both inaccessible to me, and it gives me the (probably flawed) impression that her research isn't the part of her life that's supposed to be important or impressive.
FWIW, that's soft of the way a lot of physics books (not textbooks) approach the subject: Einstein/ Heisenberg/ Bohr/ Pauli/ Feynman/ Oppenheimer was this kind of person, oh, and by the way he came up with this theory of X. Apparently a lot of people like that way of presenting science, but it's not for everyone.
The LLM is always operating as designed, but humans call its outputs "hallucinations" when they don't align with factual reality, regardless of the reason why that happens and whether it should be considered a bug or a feature. (I don't like the term much, by the way, but at this point it's a de facto standard).
I remember that. I studied CS in that period and some professors were convinced that software development was going to become an unskilled job, analogous to bricklaying, and that our goal as future CS graduates should be to become managers, just like someone that studies a university degree about making buildings is intended to become an architect and not a bricklayer.
I never believed it, though (if I had, I would probably have switched degrees, as I hate management). And while the belief was common, my impression is that it was only so among people who didn't code much. The details on how it would happen were always highly handwavy and people defending that view had a tendency to ignore any software beyond standard CRUD apps.
In contrast, if I had to choose a degree right now, I'd probably avoid CS (or at most study it out of passion, like one could study English philology or something, but without much hope of it being a safe choice for my career). I think the prospects for programmers in the LLM era look much scarier, and the threats look much more real, than they ever did in that period.
The bigger issue is that so many people have jumped into CS because programming (not the same thing I know) has become seen as this thing that will earn you big bucks.
Of course, some level of computer skills is important in most professions at this point. But logic suggests that CS (and programming) compensation will level out at a level comparable to similarly skilled technical professions.
But what's the logic? I have never seen it but it doesn't sound good even aesthetically (which is usually the justification for all kinds of violations of common sense). So what are they thinking?
A number of hotels that were built with this lack of privacy (including one I love - but its been fixed there though - more subtle worse as you could see in from the stairs) were all designed by the same architect who is said to have had a kink about looking into toilets.
Maybe he started it, and as his hotels are (otherwise) lovely it made it part of a cool aesthetic and was therefore copied?
Which is, of course, not how people primarily communicate on WhatsApp (or any communication based app). People don't send streams of videos to each other in group chats. They write text and add gif memes.
reply