p1esk

34 ms·

721.
by p1esk 3y ago
Why do you say emergency medicine is fun? I imagine you often see people in pain, life degrading injuries, and all kinds of other sad things.
722.
by p1esk 3y ago
I tried a dozen of different prompts in suno ai to generate some music - it completely ignored them. It just generated some simple pop sounding tunes every time. I’m not impressed.
723.
by p1esk 3y ago
But none of these is a significant improvement over Jukebox. I think at this point everyone is waiting for Jukebox2.
724.
by p1esk 3y ago
Yes: https://github.com/riffusion/riffusion
725.
by p1esk 3y ago
A lot of humans would fail such a test.
726.
by p1esk 3y ago
Is there a better online community of smart people I can join?
727.
by p1esk 3y ago
So, is this a good investment opportunity, what do you all think?
728.
by p1esk 3y ago
My favorite question.
729.
by p1esk 3y ago
This is a good paper. Though emergence doesn’t necessarily require a sudden jump in metrics, or unpredictability. New abilities can emerge gradually.
730.
by p1esk 3y ago
Why is Mars a good stepping stone to leave the solar system? The moon seems like a much better choice: much closer to Earth and much lower gravity to launch from.
731.
by p1esk 3y ago
Thanks! Big names among authors, I’m going to read it.
732.
by p1esk 3y ago
I do ML research at a small industrial lab. I’ll gladly provide some compute to people with a cool idea if that results in my company name listed on a paper in a top conference. Especially if the people are from a top university.
733.
by p1esk 3y ago
”All our Earth analogies are much weaker than the conditions that are present on Mars,” says Friedmann. “Mars is much drier, much colder, with much more intense UV radiation.” Why do people want to go to Mars? Seems like no matter how bad
734.
by p1esk 3y ago
I think you’re missing my point. RAG is something that you currently implement yourself, or pay someone else who already implemented it. With GPT5 (or GPT6 at the latest), you just give it the same access you would give to a RAG system, and
735.
by p1esk 3y ago
They had this promising result before they posted the paper.
736.
by p1esk 3y ago
Learn whatever interests you the most.
737.
by p1esk 3y ago
Not sure I understand what you mean by “doesn’t scale”. Are you trying to say you would like to see a tiny model performing as well as a large model?
738.
by p1esk 3y ago
I wouldn’t waste my time learning about RAG. It’s quite likely GPT5 will make all current methods obsolete.
739.
by p1esk 3y ago
They could partner with big industrial labs.
740.
by p1esk 3y ago
This method has only been tested on tiny models (<1B) and tiny dataset (17B tokens). It’s not clear if it scales.
741.
by p1esk 3y ago
How do you know the penetration testers you hired are not black hat hackers?
742.
by p1esk 3y ago
Sure, good UI is important, but do you really want to be a wrapper around someone’s model?
743.
by p1esk 3y ago
What will you do if OpenAI decides to release Jukebox2?
744.
by p1esk 3y ago
Musk should have spent 44B on a chip company.
745.
by p1esk 3y ago
This looks like a lay off
746.
by p1esk 3y ago
What you wrote makes sense, but I’m not sure I understand what an RNN has to fo with it. For example, what does it mean: “RNN that works post tokenization/embedding and at the output of individual transformer heads.”?
747.
by p1esk 3y ago
1. Self guided recurrent loops. Rnns are difficult to train correctly, which is why transformer architecture took over, but for reasoning there needs to be some sort of recurrent loop that is self determined by the model to iterate on the
748.
Ask HN: What do you expect from GPT-5?
10 points by p1esk 3y ago|17 comments
749.
by p1esk 3y ago
$8.25/hr for H100? Compare to lambdalabs at $2.5/hr. Why did you add “low cost” to the title?
750.
by p1esk 3y ago
If you have designed FPUs you should know that FP computation involves a lot more additional operations than just shifting (e.g. rounding, subnormals, and special value handling). That’s why, for example, CPUs use different hardware blocks

More ›