@glyph Did you quote post something?
Uncategorized
903
Posts
312
Posters
0
Views
-
@mijenix @tychotithonus TIL! Thanks! This iPad does indeed remain idle for long periods of time, so it's plausible.
-
@glyph :) its AAALIIIVE!!! :)
-
-
@glyph Having found a 2007 standard that I've been trying to convince people to adopt since 2018 with little success, I keep thinking: "I need to write more implementations of this spec so I can show people why it's cool. Well, the half-dozen implementations I've written so far haven't done the trick, but maybe if I just write an entire new RSS feed reader that works how I want it to then people will get the idea." The project of writing a feed reader is the only thing that has tempted me to try an LLM, and it boils down to the reasons you're talking about: dealing with people is complicated and it seems like approximately nobody else cares about the things I care about, so it's tempting to try to get help from a machine instead. But I really don't want to go down that road…
-
@glyph I gotta be better about speaking up when I'm hanging out in your chat. My lack of relevant insights has, thus far, made it feel like I'd be lobbing a distraction at you
-
@glyph I gotta be better about speaking up when I'm hanging out in your chat. My lack of relevant insights has, thus far, made it feel like I'd be lobbing a distraction at you
@wuest tbh hearing chat *trying* to understand what I'm doing is really motivating. Best case scenario, sometimes the explaining prompts me to realize that I didn't actually know what I was doing as well as I thought I did, and change strategy. In the average case the rest of the chat starts to understand it a bit better and can lob in more useful suggestions. Even the worst case is just a more entertaining stream
-
@glyph Also, learning implies the ability to generalize? LLMs cannot do that, so they have to be trained on O(the entire internet) to make sure there's something to plagarize for every situation. That's not learning, it's leaning over another kid's shoulder and copying all the answers.
-
> “learning” is inherently mutual
@glyph as a part-time teacher I often think about something a teacher of mine once said: I love teaching because I learn so much.
-
@glyph Also, learning implies the ability to generalize? LLMs cannot do that, so they have to be trained on O(the entire internet) to make sure there's something to plagarize for every situation. That's not learning, it's leaning over another kid's shoulder and copying all the answers.
@xgranade I tend to agree but this veers directly into that frustrating philosophical territory. Worse, actually, because this invites a conversation about what exactly "generalizing" is, and whether the (objectively impressive) pseudosemantic word extrusion that they do counts, and then we have to talk about benchmarks and "quals" and I'm already tired
-
> “learning” is inherently mutual
@glyph as a part-time teacher I often think about something a teacher of mine once said: I love teaching because I learn so much.
@justvanrossum as I said! a cliché! I've done like… probably a few hundred hours of teaching in my entire life, but even I know this!
-
@glyph the story in four acts we are about to see (just one example, but could be Wikipedia, or anything else as well)
-
@glyph All great points!
In case you are interested, here is my philosophical take (also arguing that AI training is different from teaching but in a different way).
https://listed.to/@24601/61247/public-domain-business-models-and-teaching-ai
-
@justvanrossum as I said! a cliché! I've done like… probably a few hundred hours of teaching in my entire life, but even I know this!
@glyph I missed the post where you did say that: I now see my response was rather redundant…
-
@glyph the story in four acts we are about to see (just one example, but could be Wikipedia, or anything else as well)
@tymwol @glyph what I'm more concerned about from the Wikipedia side of things is sources that used to be reliable becoming untrustworthy LLM generated garbage (fake citations are popping up in peer reviewed papers a lot more nowadays, for example). There's some news sites that have decided meh who needs human writers. Stuff like that.
-
@glyph the story in four acts we are about to see (just one example, but could be Wikipedia, or anything else as well)
Call it a conspiracy, but I'm still convinced that #StackOverflow is dying primarily because the marketing for CodeGen AI demotivated a lot of people to the point that they do not even want to program anything anymore.
I've seen a lot of (what could have been) Juniors avoid doing anything with code because AI will be able to do it in 5 years anyway so there is no point in learning it.And then the large layoffs, so the remaining ones are probably too busy to write answers there
-
Call it a conspiracy, but I'm still convinced that #StackOverflow is dying primarily because the marketing for CodeGen AI demotivated a lot of people to the point that they do not even want to program anything anymore.
I've seen a lot of (what could have been) Juniors avoid doing anything with code because AI will be able to do it in 5 years anyway so there is no point in learning it.And then the large layoffs, so the remaining ones are probably too busy to write answers there
@agowa338 @glyph Well, yeah, kind of. Also, StackExchange company themselves pushed AI to SO, so they helped it happen. But... same argument can be made about any other service, for example, if we stop using Wikipedia and switch to LLMs, same thing will happen. If we stop using our brains and switch to LLMs, we would end brain dead.
-
@agowa338 @glyph Well, yeah, kind of. Also, StackExchange company themselves pushed AI to SO, so they helped it happen. But... same argument can be made about any other service, for example, if we stop using Wikipedia and switch to LLMs, same thing will happen. If we stop using our brains and switch to LLMs, we would end brain dead.
You missed my point. My point was that people aren't moving towards LLMs but instead stopping to code all together and do other things instead...
If people were actually moving towards LLMs you'd see a lot of threads with people not understanding their own code and having to figure out the complex issues that LLM generated code causes (aka you'd see an uptick in people having to troubleshoot their own AI generated code hallucinations)...
-
@glyph I missed the post where you did say that: I now see my response was rather redundant…
@justvanrossum no worries, just enthusiastically agreeing :)
-
You missed my point. My point was that people aren't moving towards LLMs but instead stopping to code all together and do other things instead...
If people were actually moving towards LLMs you'd see a lot of threads with people not understanding their own code and having to figure out the complex issues that LLM generated code causes (aka you'd see an uptick in people having to troubleshoot their own AI generated code hallucinations)...
@agowa338 @tymwol @glyph I have been trying to understand this area with precision. I think the kinds of problems the code assistants tend to fail at are the kind that were not generally answerable on StackOverflow in the first place, and which a less experienced programmer might not even recognize or be able to describe well enough to even ask a question.
-
Oh and my biggest indicator for my assumption being right is looking at all of the "AI bros" and them after a while shit talking all of the codegen stuff themselves with phrases like "what's the point of all of this when I've to hire a programmer to make it work in the end anyway?" for example...
Gli ultimi otto messaggi ricevuti dalla Federazione
-
Thanks to everyone who responds or replies to my polls. They're one of my favourite parts of being on the Fediverse.
-
@zwol at this point I'd forego semantic, and just take easy and nice to look at. If I were writing a book today, I'd author it in Markdown, and then convert it to...whatever...for editing, or whatever at the very end. Before publishing, I ended up converting to an OpenOffice doc with the publishers template for editing, anyway (but I didn't originally write it to be published, it just worked out that way, so the OpenOffice step wasn't originally in the plans).
-
@ann3nova this is such a specific thing to obsess over, I love it. I drew my own pointers back then, and the one I used most looked pretty much like this. Same shape, but different colors.
-
@swelljoe I completely agree that it becomes tiring looking at all the tags. Not sure I've ever seen a markup language that was both semantic and pleasant to read in source form, though.
-
Le matematiche (libro)
@libri - La matematica è la stessa, ma la scuola russa la vede in modo diverso
-
@zwol I wrote a book in Docbook (SGML first, and then converted to XML later) about 25ish years ago, with a custom vim configuration. It wasn't pleasant. Yes, the tools were bad, but also it's just really sort of tiring looking at all the tags and the processing toolchain was dire.
I was using a LaTeX-based toolchain for the PDF generation, I don't remember details, but Sebastian Rahtz was super helpful when I ran into mysterious issues.
I wouldn't do it again. Different time, different me.
-
@jannem Yeah, like, considered as a thing-in-itself apart from the tools, the major complaint I have about DocBook is that the documentation of _how to write a book using DocBook_ is inadequate.
Many people have very similar complaints about LaTeX.
-
@Esoteria you might be a little too young for this classic scene from "WKRP in Cincinnati".
