koverstreet

7 ms·

1.
by koverstreet 3d ago
Yeah, they won't. And if a threat to the human race does come from AI, it's going to come from OpenAI/Anthropic. Hypercapitalist, secretive, in bed with the government, plus multiple real documented hackings of open source in
2.
by koverstreet 3d ago
Who should I be more scared of? China, which has been doubling down on open transparent research, or the secretive US companies who are in bed with the most unhinged administration we've ever had and has been actively starting wars? Th
3.
by koverstreet 3d ago
"Trust us, really!" HAH
4.
by koverstreet 17d ago
Yeah, I'm with Theo on this one. Conventional OS security between Ring-0 and everything else is well understood; the problem has become too much code in Ring-0, a great fraction of which has its own interfaces across the security bound
5.
by koverstreet 23d ago
Instead of railing about ideology, try comparing the actions of the U.S. and China over the past 20 years. The U.S. has been far more interventionist and aggressive; China appears to be maturing and the U.S. is regressing.
6.
by koverstreet 1mo ago
I consistently see a level of competence and professionalism in South America that has left the building in the states.
7.
by koverstreet 2mo ago
The maturity of leadership in China seems to be on a whole different level from the U.S. Lots of companies in the U.S. have fallen victim to that syndrome, but if you used the word "restraint" in that context in Silicon Valley mos
8.
by koverstreet 2mo ago
U.S. border agents really are just worse. Much worse. (I can't say what racial profiling adds to that, I can only speak to my own experiences). I've been routinely grilled and extensively searched many times just crossing between
9.
by koverstreet 2mo ago
I'm not sure if the complexity is at the neuron level. It's clearly possible; we see some pretty complex behavior from single celled organisms - Stentors - and there were the experiments (starting in the 70s, I believe) that condi
10.
by koverstreet 2mo ago
bcachefs is slowly working up to what you're describing - describing what you want and letting the filesystem sort it out. We're basically there for local storage, and other people have been building some nice reporting on top. Ne
11.
by koverstreet 2mo ago
> More broadly, the predictive brain model would suggest that all of the brain, not just the neocortex, is dedicated to prediction. What would you say makes LLMs similar to the neocortex, rather than the basal ganglia or Broca's are
12.
by koverstreet 2mo ago
That's fundamental to how anything that compresses/understands the world has to work, in the Kolmogeravian sense. That's why people denigrate LLMs as being just "next token predictors" - they're not wrong, but
13.
by koverstreet 2mo ago
Yeah, there's some beautiful math underlying what LLMs are doing, and it's the same math our neocortex runs on.
14.
by koverstreet 3mo ago
The mm people are increasingly hostile to any method of handling OOMs (like, just failing the allocation) besides the OOM killer - it's become very dominated by the hyperscalars and cloud vendors. Working around mm nuttiness is a frequ
15.
by koverstreet 3mo ago
The point is that the "outbox pattern" is not an atomic transaction. You fundamentally don't have those in the distributed world, via the CAP theorem, and if you want anything close to the guarantees a local transactional dat
16.
by koverstreet 3mo ago
The certainly don't teach you that in business school :)
17.
by koverstreet 3mo ago
"AI girlfriend" is something the trolls invented. I've posted code and research - that's how you know they're trolls, they make stuff up anyways :)
18.
by koverstreet 4mo ago
Are you forgetting the nitrogen? :)
19.
by koverstreet 4mo ago
We're in the middle of a huge spike in LLM discovered security vulnerabilities, which means not everything will get assigned a CVE, a lot of people are watching repositories to look for exploitable bugs, and in the frenzy of backportin
20.
by koverstreet 4mo ago
100% - but that's where writing regression tests when people find things really helps with the stress levels of future-you :)
21.
by koverstreet 4mo ago
If package maintainers were always fine upstanding package maintainers as you imagine them to be I wouldn't be complaining, but I have in fact had Debian ship my software and screw it up and gotten a flood of bug reports, so... :) I th
22.
by koverstreet 4mo ago
You definitely need different channels for high priority fixes and normal releases, stable and testing releases and all that. But two years is impractical and Debian gets a ton of friction over it. Web browsers and maybe one or two other pa
23.
by koverstreet 4mo ago
You're going to have to update production at some point, and delaying it to once every 2 years is just deferred maintenance. And you know what they say about that... So when you do update and get that GSSAPI change, it comes with two y
24.
by koverstreet 4mo ago
Well, my workstation runs Debian sid, and all the newer stuff runs NixOS... But that does nothing for people who write and support code Debian wants to ship - packaging code badly can create a real mess for upstream.
25.
by koverstreet 4mo ago
No, that's exactly the thing to complain about. That whole model dates to before automated testing was even really a thing, and no one knew how to do QA; your QA was all the people willing to run your code and report bugs, and that too
26.
by koverstreet 4mo ago
It seems to me you're style over substance then.
27.
by koverstreet 5mo ago
That's pretty much been my day - today was genuinely bad, and I've been putting up with a lot of this lately. Now on Qwen3.5-27b, and it may not be quite as sharp as Opus was two months ago, but we're getting work done again.
28.
by koverstreet 5mo ago
https://transformer-circuits.pub/2026/emotions/index.html At the actual inference level temperature can be applied at any time - generation is token by token - but that doesn't mean the API necessarily expose
29.
by koverstreet 5mo ago
It's how temperature/top_p/top_k work. Anthropic also just put out a paper where they were doing a much more advanced version of this, mapping out functional states within the modern and steering with that.
30.
by koverstreet 5mo ago
Technically speaking, models inherently do this - CoT is just output tokens that aren't included in the final response because they're enclosed in <think> tags, and it's the model that decides when to close the tag. You

More ›