Petr Baudis (pasky)
xpasky.bsky.social
Petr Baudis (pasky)
@xpasky.bsky.social
Rossum.ai. A variety of OSS & AI things in the past (Git, glibc, pre-AlphaGo Pachi, OpenTTD, ...). Your computer might be running some of my code (sorry).
Pinned
Claude knows how to use vim! And... you can do anything from vim.

Embedding Claude in vim means you get a fully agentic AI pair programmer in <1000 LoC. (Or a generic chatbot, or Linux co-admin.)

Unlike CLIs, you can review Claude's changes, and keep the chat history.

github.com/pasky/claude...
Reposted by Petr Baudis (pasky)
Definitely believe them regarding technical capabilities of the models. (Ok maybe add a 6-12m buffer.)

Where they are imho over indexing is maximum realistic speed of adaptation of the real world.

Adoption even of the most amazing stuff will take time, and need a lot of infra.
January 6, 2025 at 1:27 PM
Reposted by Petr Baudis (pasky)
i was annoyed at having many chrome tabs with PDF papers having uninformative titles, so i created a small chrome extension to fix it.

i'm using it for a while now, works well.

today i put it on github. enjoy.

github.com/yoavg/pdf-ta...
January 5, 2025 at 10:22 PM
Reposted by Petr Baudis (pasky)
good news: despite all the bleak shit going on in the world it is once again Awesome Games Done Quick week. go watch some speedruns www.twitch.tv/gamesdonequick
GamesDoneQuick - Twitch
Awesome Games Done Quick 2025 - Benefiting the Prevent Cancer Foundation - Ori and the Blind Forest
www.twitch.tv
January 5, 2025 at 10:07 PM
Quick primer for non-wizards about the post-MCTS LLM reasoning future:

How will LLMs learn to reason efficiently?

No math in this thread, ~simple words only! Let's go through the "Process Reinforcement through IMplicit REwards" (PRIME) method. 1/n

curvy-check-498.notion.site/Process-Rein...
January 5, 2025 at 12:17 AM
Reposted by Petr Baudis (pasky)
Science/maths/programming have a tendency to depreciate the value of grinding - smart people don't grind! - yes, that project basically took me only 30 minutes. The 3 and half hours of dead ends I went down obviously don't count. Or the hour I spent installing the wrong package.
December 24, 2024 at 1:27 PM
"Is System 2 thinking even real, chat?"

Humans vary widely between being very intuitive or deliberative.

Human-level AGI could plausibly exist even purely as System 1, or with only very basic System 2 mixed in. Just git gud at superhuman intuition.
December 13, 2024 at 1:07 PM
Reposted by Petr Baudis (pasky)
What can RBMK teach us about building AGI?
December 7, 2024 at 8:11 PM
Reposted by Petr Baudis (pasky)
Can we build something lasting in a world where technology is easily reproducible?

How much time do we have before Artificial General Intelligence (AGI) is here, and how much time do we have afterwards? 2/n
December 9, 2024 at 4:07 PM
The product strategy of most SaaS isn't ready for AGI.

A disconnect is widening between how product builders think, and how Artificial Intelligence researchers project short-term future.

But I believe it's a solvable problem (in any world where a solution exists). 1/n🧵
December 9, 2024 at 4:06 PM
Reposted by Petr Baudis (pasky)
THEY HAVE PLAYED US FOR ABSOLUTE FOOLS
November 29, 2024 at 10:31 PM
If I'm defining my self-worth by anything, it's positive effect of my actions.

Also, I try to live as a stoic, finding happiness in what I can control: my thoughts and actions.

However, positive effect of my actions isn't just in my control! Only their intent is.

/off figuring it out
December 8, 2024 at 11:09 PM
Damn, rewatching after 10y hits as hard as ever.

I now realize Haibane Renmei is not just my most favorite show ever, but also the core of my moral philosophy.
December 8, 2024 at 5:55 PM
Reposted by Petr Baudis (pasky)
And on the pedestal these words appear:

"My name is Union of Soviet Socialist Republics:

Look on my works, ye Capitalists, and despair!"

No thing beside remains. Round the decay

Of that colossal wreck, boundless and bare

The lone and level steppes stretch far away.
December 7, 2024 at 8:42 PM
What can RBMK teach us about building AGI?
December 7, 2024 at 8:11 PM
Reposted by Petr Baudis (pasky)
Two stages of writing:

1) This shouldn't take too long
2) Oh no
December 4, 2024 at 7:36 PM
the starships never pass by here anymore

95% of stars that would ever be born have already been born

it's calm and serene... and focused here

www.youtube.com/watch?v=Bx5b...
the starships never pass by here anymore
YouTube video by spiritual brother Sci-Fi
www.youtube.com
December 3, 2024 at 11:40 PM
Best read of the week, and I've read some bangers this week.
Hello, I wrote a new talk that is ultimately about how misery is a dumb strategy

egoless.engineering
Egoless Engineering
It turns out misery is a shitty proxy metric.
egoless.engineering
December 3, 2024 at 10:39 PM
That feeling when you accidentally find the perfect citation for your essay.

(or: why does it make sense to build software products even in the age of AGI? opportunity cost!)
“Most of the technologists I know [see the future with] equal parts melancholy, fatalism, and pride~an Oppenheimer-esque ‘Now I am become death, destroyer of jobs’ kind of thing…in the meantime, it’s something cool to have worked on” ⁦‪

www.noahpinion.blog/p/plentiful-...
Plentiful, high-paying jobs in the age of AI
Comparative advantage is very subtle, but incredibly powerful.
www.noahpinion.blog
December 3, 2024 at 12:29 AM
One thing Czech universities (at least matfyz) don't do well is building a community of students (especially introverted ones).

(Cambridge looked so very different in the Thinking Game biopic of Demos Hassabis!)
December 2, 2024 at 8:15 PM
Reposted by Petr Baudis (pasky)
To lay my cards out, more and more I'm coming to the view that an underrated role of education is simply stuffing the kid full of facts and technique. References cases. Both as mental food and to show them that you can do powerful stuff with a brain, even as a kid.
December 1, 2024 at 7:44 PM
THEY HAVE PLAYED US FOR ABSOLUTE FOOLS
November 29, 2024 at 10:31 PM
Reposted by Petr Baudis (pasky)
On BSky, there is a gentleman's club vibe of overpolite society with unreasonably strict rules that are aggressively enforced, but mostly obvious when you break them.

On X, "freedom of speech" reigns and everyone writes "the other app" in fear of being randomly shadowbanned.

Which is better?
Thank you @bsky.app team for correcting the mistake. Glad to be back!
November 29, 2024 at 1:45 AM
On BSky, there is a gentleman's club vibe of overpolite society with unreasonably strict rules that are aggressively enforced, but mostly obvious when you break them.

On X, "freedom of speech" reigns and everyone writes "the other app" in fear of being randomly shadowbanned.

Which is better?
Thank you @bsky.app team for correcting the mistake. Glad to be back!
November 29, 2024 at 1:45 AM
Reposted by Petr Baudis (pasky)
There's something quite sobering about the fact that 95% of all stars that will ever be born ... have already been born.

(I learned this from @astrokatie.com)
November 26, 2024 at 1:32 PM
I'm a month late catching up but the stateof.ai 2024 report is 200 high density slides of pure gold - so much that I somehow missed over the year.

(Need to take second look at ColPali! Missed much of the synthetic data discussion. Would be less surprised about quantized models high perf. ...)
State of AI Report 2024
The State of AI Report analyses the most interesting developments in AI. Read and download here.
stateof.ai
November 25, 2024 at 11:35 PM