#editorial | Logs for 2025-06-06
« return
[00:00:15] <chromas> is that anticompetitive?
[00:04:51] <kolie> No?
[00:05:30] <kolie> anti-competitive deez nuts.
[00:05:33] <kolie> ami right>?
[00:24:38] <chromas> they have their own ai
[00:24:57] <chromas> so they want to be able to use the user's data but not share. Sue! Sue!
[01:45:44] <halibut> I wonder if they simultaneously feel it is Fair Use to train off of books and articles written by people who do not want their works used to train LLM models.
[02:01:45] <kolie> Most companies don't want to give other paid commercial entities free access to something that costs them a lot to have/maintain.
[02:01:54] <kolie> X's platform is expensive, and they are for profit
[02:02:05] <kolie> It makes sense they;d want to maximize that value, and not give away free value.
[02:02:08] <kolie> They aren't a charity.
[02:04:44] <chromas> Twitter never made a profit until Musk bought it, then suddenly it was retroactively profitable
[03:26:40] <halibut> I concur that X is around to make a profit. However, authors also wish to profit off of their work, and I was wondering if X, when training their LLMs, used authors' works without permission.
[17:25:10] <kolie> janrinok, signed off on japan lander, cool stuff.
[17:25:18] <kolie> links had a tracking thing that was the only minor thing i removed.
[17:25:48] <janrinok> fair cop, I should have done that
[17:25:57] <kolie> theres some broken links too
[17:26:01] <kolie> that are relative to the source site.
[17:26:56] <janrinok> broken, as in completely wrong or abbreviated and using SN instead of the original?
[17:26:58] <kolie> I think its cached on the frontpage but fixed in the source.
[17:27:10] <kolie> https://www.newscientist.com is missing, the a href is just a href="/article/somenumber/"
[17:27:12] <systemd> ^ 03New Scientist | Science news, articles, and features
[17:27:26] <kolie> so its cached as https://soylentnews.org
[17:27:27] <systemd> ^ 03404 File Not Found
[17:27:33] <kolie> instead of @ newscientist.
[17:27:44] <janrinok> Yeah, I am testing some new software. The Python handles that but I haven't got it implemented in GO yet.
[17:28:02] <kolie> Need any help?
[17:28:44] <janrinok> I'll scratch my head for a bit longer, I'm sure that there is a way to do it but it isn't jumping out at me at the moment.
[17:29:32] <janrinok> I'm just not used to the GO libraries yet
[17:30:20] <kolie> do a lot of go at work.
[17:32:12] <janrinok> Python has some very convenient routines in BeautifulSoup that I have found an equivalent for in GO
[17:32:30] <janrinok> * that I haven't
[17:32:32] <kolie> Which in particular?
[17:33:24] <janrinok> BS has a single comment to return all the links, whereas I haven't found the equivalent in GO.
[17:33:44] <janrinok> single command - damn I'm getting tired again
[17:35:22] <kolie> https://justpaste.it
[17:35:24] <systemd> ^ 03JustPaste.it - Share Text & Images the Easy Way
[17:35:34] <kolie> something like that probably.
[17:35:53] <kolie> goquery is similar to BS, I think for scraping a lot of go people like colly
[17:38:27] <janrinok> thanks for that - I will look at it over the weekend, well look at both of them I suppose. I am trying to use the standard libs as much as possible but I am beginning to use a few of the external libs too
[17:39:32] <kolie> goquery is like standard lib in most tool kits practically, but yea, no point in writing an html parser its pretty low level and goquery is the uhh basically standard option I guess.
[17:39:54] <kolie> colly is a bit more batteries included scraping.
[21:58:55] <chromas> rehash bug request: allow longer article titles :D