rosstaylor90

6 ms·

1.
Scaling RL Compute(gr.inc)
2 points by rosstaylor90 1y ago|0 comments
2.
by rosstaylor90 2y ago
Many math questions are easy to verify, and it's a classic benchmark for reasoning -> so it's a good hill to climb. I agree with your meta-point that better benchmarks testing more types of task would be good!
3.
by rosstaylor90 2y ago
RL has more than two steps...
4.
by rosstaylor90 2y ago
I disagree, 10 years ago AIs nailing these types of competition would have been seen as very impressive. The fact goal posts can move on this now shows how much AI has progressed. (Also the term “approximate retrieval” is a bad one - reason
5.
by rosstaylor90 2y ago
Thanks for feedback! Yes, we’re looking to improve quality in the coming months. Couple of notes: - The initial use of data is distillation so we’re less bound by question quality (anything that evinces output diversity is good). - But movi
6.
by rosstaylor90 2y ago
Happier now, upgraded the backend :) (co-creator here)
7.
by rosstaylor90 2y ago
What's your AIME 2025 score? https://gr.inc/RJT1990/AIME2025/
8.
by rosstaylor90 6y ago
No, the primary reason is that the UK is a heavily service-based economy. Quarantine impacts the service-based subset of the economy the most. Therefore it makes sense that the UK would be hard hit. A “nation of shopkeepers” suffers when th
9.
by rosstaylor90 6y ago
Ed Thorp’s Princeton/Newport Partners was the first great quant hedge fund.
10.
Methods in Deep Learning Taxonomized and Indexed(paperswithcode.com)
4 points by rosstaylor90 6y ago|0 comments
11.
A Home for Results in Machine Learning(medium.com)
1 points by rosstaylor90 6y ago|0 comments
12.
by rosstaylor90 7y ago
I'm from Suffolk - my school used to go here on school trips. The funny thing about the site is that it's basically just a hill in a field now, so needless to say it was a massive letdown for eight-year-old me who was expecting to
13.
by rosstaylor90 8y ago
ATLAS ML | London, UK | Full Stack Developer | Remote | FULL-TIME Our mission is to systemise the world's deep learning knowledge. We are building a discovery platform and tooling for deep learning that links ideas with implementations
14.
by rosstaylor90 8y ago
I think your conclusion is too strong. Yes, we know bigger models and data generally lead to better performance (e.g. BigGAN results last year) but progress in architecture can still speed up progress in ML tasks. If we were still stuck usi
15.
by rosstaylor90 8y ago
Philosophical Investigations - Wittgenstein. Just stunningly thoughtful and changed the way I interpret most things around me.
16.
State of Deep Learning: H2 2018 Review(medium.com)
5 points by rosstaylor90 8y ago|0 comments
17.
by rosstaylor90 8y ago
Cool example for a music video here: https://www.youtube.com/watch?v=-k9qDxyxS3s .
18.
Wiggle Stereoscopy(en.wikipedia.org)
2 points by rosstaylor90 8y ago|1 comments
19.
Ask HN: Why is HN so obsessed with Google?
9 points by rosstaylor90 8y ago|2 comments
20.
by rosstaylor90 8y ago
Hey all, One problem we face is how unmodular deep learning code can be, meaning that adapting someone else's code for your own dataset, your own workflow, your own way of training, and so on, can be a real pain. We made an open source
21.
Mantrahub – Plug & Play Deep Learning(mantrahub.io)
2 points by rosstaylor90 8y ago|1 comments
22.
by rosstaylor90 8y ago
I'm going to get one printed in real-life - once I do some superresolution and probably a bit of manual Photoshop editing. Which design do you guys like best?
23.
AI Generated Skateboard Deck Graphics(github.com)
2 points by rosstaylor90 8y ago|2 comments
24.
by rosstaylor90 8y ago
Hi all, just sharing a project I've been working on. Mantra is a deep learning development kit that manages the various components in an deep learning project, and makes it much easier to do routine tasks like training in the cloud, mo
25.
Show HN: Mantra: a high-level development framework for deep learning projects(github.com)
15 points by rosstaylor90 8y ago|1 comments
26.
by rosstaylor90 8y ago
Sorry to ask that annoying question that people tend to ask, but how does this differ from https://github.com/trending ?