chpatrick

4 ms·

1.
by chpatrick 7d ago
Slop is slop because people put in zero effort so the LLM just produces the first thing it was trained to do. You can absolutely get good results out of AI but it takes taste and a little work.
2.
by chpatrick 7d ago
Yeah but we're talking about something that was trained on petabytes of noisy, difficult to parse and ambiguous information.
3.
by chpatrick 10d ago
They've read a few thousand years of experiential data in text form.
4.
by chpatrick 11d ago
They remind me of Kodak inventing the digital camera and sitting on it to preserve their film business.
5.
by chpatrick 11d ago
I find its style the most sycophantic and annoying personally.
6.
by chpatrick 11d ago
Spoiler alert, almost everything is made in China, including the premium stuff.
7.
by chpatrick 14d ago
Minimax H3 works pretty great and you can run it on a 3090.
8.
by chpatrick 15d ago
That's included in the number, buddy. Read the report. https://iea-pvps.org/wp-content/uploads/2020/12/IEA-PVPS-LCI...
9.
by chpatrick 15d ago
https://iea-pvps.org/wp-content/uploads/2022/11/Fact-Sheet-I... Making and using a modern solar panel emits up to 44 grams of CO2 equivalent greenhouse gas per kWh over its lifetime as opposed to up to 1
10.
by chpatrick 16d ago
So if a model with exactly the same architecture controls a robot then it's suddenly sentient or what? "They generate actions in the real world"
11.
by chpatrick 16d ago
Does being made out of meat instead of silicon make you more sentient?
12.
by chpatrick 17d ago
Shouldn't doesn't mean people wouldn't.
13.
by chpatrick 18d ago
It's actually a great service though.
14.
by chpatrick 20d ago
If ai watermarking is undetectable to humans I wonder if sinister stuff in the context is also undetectable... Some thought or mood that you can't read but is still encoded in the tokens.
15.
by chpatrick 21d ago
As long as you only want to play games from 2008. :)
16.
by chpatrick 21d ago
I had the same job 5 years ago, but I'm sure you know better.
17.
by chpatrick 21d ago
And model training isn't putting money back into the business?
18.
by chpatrick 21d ago
As opposed to...?
19.
by chpatrick 21d ago
I'm also outsourcing computational geometry to CGAL and CPU optimizations to GCC, so what? I can still reason about them. For example for a problem I had recently Claude pointed out (after some prodding for me) that the determinant of
20.
by chpatrick 22d ago
Still unsolved for 87 years.
21.
by chpatrick 22d ago
About two month ago: https://en.wikipedia.org/wiki/Jacobian_conjecture
22.
by chpatrick 22d ago
I'm also dependent on a bunch of advanced libraries and compilers I didn't write. I could reimplement them if I took years to do it but it would be a total waste of time.
23.
by chpatrick 22d ago
I work on a codebase with a lot of math and I've both learned and implemented a bunch of things that were beyond my level before I could iterate on it with AI. Same for binary reverse engineering. On the other fixing tedious bugs by ha
24.
by chpatrick 22d ago
Typing the code was never the interesting part.
25.
by chpatrick 24d ago
Also true for human developers.
26.
by chpatrick 24d ago
I completely stopped using Gemini because I found its tone so annoying and pandering. Claude gets to the point.
27.
by chpatrick 25d ago
I think these kinds of comments really need to say which LLM that is. There's an enormous difference in skill between the frontier ones and say the Google search AI.
28.
by chpatrick 1mo ago
You just have to press the right link on the site and you get the tracking page (despite saying you need the app).
29.
by chpatrick 1mo ago
Overheating isn't exactly right, they got shut down because they couldn't be cooled, but they didn't overheat.
30.
by chpatrick 2mo ago
JS despite its (many) faults still allowed a fairly functional style way before that was mainstream. Java and C++ only got lambdas decades later.

More ›