jryan49

3 ms·

1.
by jryan49 10d ago
I can't imagine inference on an image could be faster that just a sort.
2.
by jryan49 10d ago
Either I use GenAI when I feel like it, or I don't, so no.
3.
by jryan49 14d ago
Three options, they actually think we're reaching RSI or they are proposing we "slow down" because the models aren't getting better at the speed they expected and want an excuse for it. Or they want to make it illegal fo
4.
by jryan49 14d ago
What did they demonstrate can you share some links? I'm not on twitter.
5.
by jryan49 17d ago
It's probably a good idea to note that it cost them 10-15 million dollars in tokens to do so.
6.
by jryan49 17d ago
I mean how often do you buy a watch? Compare with how much do you spend on groceries every month...
7.
by jryan49 23d ago
We created AGI so I don't have to click my mouse to change the background color of my slides.
8.
by jryan49 23d ago
I wonder if people who are down-voting you actually have children. I even know someone who's kid just died in a pool when no one was watching... My kids almost drowned themselves multiple times when they were below 5 yr old and they ar
9.
by jryan49 23d ago
I mean I wouldn't let me 5 year old cross busy streets by himself in my neighborhood... People drive while staring at their phones going 45 mph in a 25... Maybe we should start punishing people who make the roads so dangerous in the fi
10.
by jryan49 29d ago
Telling llms what not to do actually makes things worse. Frame things in positive language where possible, if not possible I don't mention it at all. The biggest fix for me is using better models. You can't rely on guardrails to m
11.
by jryan49 29d ago
https://www.crowdsupply.com/open-tools/open-printer/updates Hopefully this actually works when (if) it comes out
12.
by jryan49 29d ago
Cause it costs money?
13.
by jryan49 1mo ago
Over the years I've used the internet less and less, HN is one of the few places I still spend my time.
14.
by jryan49 1mo ago
Unusable how..?
15.
by jryan49 1mo ago
AI boom (probably) can't last forever, so yes.
16.
by jryan49 1mo ago
A quick Google search will show you... https://www.usatoday.com/story/news/politics/2025/04/26/nati... You're just willfully ignorant
17.
by jryan49 1mo ago
Here ya go https://www.nature.com/immersive/d41586-026-00088-9/index.ht... Something like 25% cut in new grants. 50% cuts to NSF... You could have just as easily found the data...
18.
by jryan49 1mo ago
In this case they'd still have to go inside to go to the bathroom. :/
19.
by jryan49 1mo ago
Who enforces "you must keep your house in good condition rules" that sounds like a lot of work to enforce. Wouldn't it be fun if a concrete company decides to bribe the government to make everyone replace their concrete every
20.
by jryan49 1mo ago
With all the software being written on github you'd think we were going though a software rennasance. Where are the results? Is it really just all slop?
21.
by jryan49 2mo ago
Yeah I probably should just give up trying at this point lol
22.
by jryan49 2mo ago
Kind of sick of people copy pasting my conversations in slack to AI, that just leads me to talking in circles.
23.
Ask HN: Anyone else getting token limited at work now?
2 points by jryan49 2mo ago|1 comments
24.
by jryan49 2mo ago
It's sad because the ai one is at least one of the worse of the set...
25.
by jryan49 2mo ago
All these hot takes that LLMs are -10x, 100x, 2x are kind of silly. It really depends on: What you are using it for? What language, what framework, how many LOC, what domain, how much docs can the LLM read, etc etc. What are you optimizing
26.
by jryan49 2mo ago
It was a frontier model from a few months ago. It hasn't done it recently. But it has happened. I imagine if there isn't a lot of training data for an idea it probably has a higher change of generating verbatim, so as time goes on
27.
by jryan49 2mo ago
I mean ai has generated copyright headers verbatim in code I asked it to make. I had a tutorial open from the web and asked the AI to generate an example for me and it was the example from the tutorial verbatim as well...
28.
Lean 4 Bug Found Incidentally by AI, "Proving" Collatz(twitter.com)
4 points by jryan49 2mo ago|0 comments
29.
by jryan49 2mo ago
AI makes mistakes like humans. We used to review human's code because humans make mistakes. What the hell has changed? I mean it's pretty effective to have another agent review the code too, but I still find it doesn't catch
30.
by jryan49 2mo ago
Since every token costs money and I dont think they are conscious, never.

More ›