Aerolfos

6 ms·

61.
by Aerolfos 10mo ago
> They are retrained every 12-24 months and constantly getting new/updated reinforcement learning layers This is true now, but it can't stay true, given the enormous costs of training. Inference is expensive enough as is, the
62.
by Aerolfos 10mo ago
Don't forget "we're obligated to try and sell it so here's an ai generated article to fill up our quota because nobody here wanted to actually sit down and write it"
63.
by Aerolfos 10mo ago
Well, it would be nice to see examples (or weights to be completely open) for the baseline model, without any GPT-5 influence whatsoever. Basically let people see what the "raw" output from historical texts is like, and for that m
64.
by Aerolfos 10mo ago
> but surely consumer SSDs wouldn't just completely ignore fsync and blatantly lie that the data has been persisted? That doesn't even help if fsync() doesn't do what developers expect: https://danluu.com/f
65.
by Aerolfos 10mo ago
> https://github.com/DGoettlich/history-llms/blob/main/ranke-4... Given the training notes, it seems like you can't get the performance they give examples of? I'm not sure about the exact
66.
by Aerolfos 10mo ago
Ok so it was that. The responses given did sound off, while it has some period-appropriate mannerisms, and has entire sections basically rephrased from some popular historical texts, it seems off compared to reading an actual 1900s text
67.
by Aerolfos 10mo ago
> To make `Z` a column vector, we would need something like `Z = (Y @ X)[:,np.newaxis]`. Doesn't just (Y @ X)[None] work? None adding an extra dimension works in practice but I don't know if you're "supposed" to
68.
by Aerolfos 10mo ago
> Why not use X.transpose()? Or just X.T, the shorthand alias for that
69.
by Aerolfos 11mo ago
The article is AI-written. Obviously can't be relied on to accurately convey the original information "But Arduino isn’t SaaS. It’s the foundation of the maker ecosystem." is a giveaway, but the whole set of paragraphs of set
70.
by Aerolfos 11mo ago
Video editing? Adobe has set themselves up for failure there, everyone wants an alternative Davinci Resolve is probably competitive with Premiere, but while free it's not actually open source. But either a viable competitor catching up
71.
by Aerolfos 11mo ago
I'm sure people will come out of the woodwork to tell me just how wrong this is, but say various Linux projects or the kernel seem to have better (better, not perfect) governance structures. Not being obsessed with rapacious growth, no
72.
by Aerolfos 11mo ago
>The Mozilla leadership seems to have a unfortunate tendency to emulate the behaviors of the tech companies that their core Firefox project is often seen as an alternative too. But of course, they need "competitive" salaries so
73.
by Aerolfos 1y ago
Actually more common than you might think. Bethesda games have the same ecosystem - they do provide an official plugin system, but since modders aren't content with the restrictions of that system, they reverse engineered the game(s,
74.
by Aerolfos 1y ago
Standard policy is I think mostly the same, but in Europe there's been arguments that those policies don't follow the actual consumer protection laws, which is a whole thing that I don't think really resolved one way or ano
75.
by Aerolfos 1y ago
Key activation is actually a good point. I believe there were some pushes to get rid of opened box = no refund policies as being against standard 14-day returns in Norway, because the law explicitly says the consumer may (paraphrase) "
76.
by Aerolfos 1y ago
In some sense, GOGs entire existence is testing the hypothesis that it's impossible to run a consumer friendly digital store due to abuse. If you took the common sense publisher view then no DRM = everything you make is instantly pirat
77.
by Aerolfos 1y ago
> Refunds with no questions asked if you've played less than 2 hours of the game, Weaker than standard physical store consumer protections (no playtime restriction on returns, obviously), and (much) weaker than GOGs refund: 1 month