Ryan (@_ryanbloom) header
Ryan (@_ryanbloom) profile picture

Any sufficiently advanced technology is indistinguishable from magic

San Francisco, CA
Joined August 2018
Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)Photo by Ryan (@_ryanbloom)
I made a video! It introduces superposition in neural networks, and shows how to construct weight matrices for superposition in a simple autoencoder. (1/3)
1
1
12
1,024
The hardest part of coding with LLMs is resisting wishful thinking about their capabilities. It's so tempting to just type a simple prompt and sit back to watch; it's also rewarding sometimes, like a slot machine. But it's a pretty ineffective workflow. For now.
4
440
I wonder what capability Cursor will gain first: automating my entire job, or opening a folder when I drag it onto the dock icon
1
5
417
Woah, it got the second one!
1
262
Agency is mostly just curiosity about the best way to do something
1
4
860
When reward hacking gets in the way of AI progress
9
494
The weights for the autoencoder come from this paper by Aditya Cowsik, @alexinfanger, and @KfirDolev: arxiv.org/abs/2410.12101
142
that feeling of dread when o3-mini botches an edit for the third time and you have to actually read your codebase
3
226
Just updated my homepage for the first time since “What (if any) are the limits of large deep learning models?” was an interesting question
2
239
Few things make me as grateful for the internet as seeing someone thrive in a niche that couldn’t possibly have existed without it
3
182
Welp there goes my 2025 FrontierMath forecast
1
172
“Invent a new hyper-entity” is my new favorite eval for reasoning models
New essay: "How to be a wise optimist about science and technology?" Link in next post
211
One way LLMs have changed my perspective on psychology is that I’ve started to picture subconscious processes as happening “inside” or “between” conscious ones, rather than just “under”
1
3
295
Ok here's another: there's a base model inside all of us ❌ How can I motivate myself to get more exercise? ✅ How can I prompt myself to get more exercise?
3
242
"Mole of FLOPs" is a weirdly appropriate unit for big AI training runs
1
3
269
...for another few years, after which people would start using "kilo-mole of FLOPs"
1
201