dilyevsky

been using it for years, haven't found a compelling reason not to yet

some sites don't work (you can explicitly whitelist them tho) but generally it's been fine. i would honestly prefer a phone that doesn't allow any comms to any non-contact list parties at this point.

> Unlike metal servers, VMs are just operating systems being gaslit by a hypervisor. Your actual CPU is still executing those instructions! And if you know where to look, yyou can even read the other VMs’ memory in L1D

only the beginning. next year or two are going to be wild on this front so buckle up...

i think part of the reason was ant had really really hard time raising their A bc at the time nobody believed second frontier lab was possible. similar to that other nuclear energy startup taking money from putin/epstein tainted fund

crossplane is more about lifecycle management than just pure API

most ops people just like templates bc coding wasn't their strong suit (at least before AI) and helm delivered the least friction way of using templates.

no one is stopping you from importing k8s types as protobufs and writing a program to talk to k8s api via protobuf encoding. all the types are already protos (except CRDs). it's just not the most user-friendly way so most people don't use it.

i have noticed with opus 5.5 it produces 10x more smaller commits doing the same kind of engineering work i was doing before. probably an artifact of the RL process. would explain massive uptick in commit load at least partially

so much $$$ and they only put pcie5 in there?

for whatever reason AI just loves colored accent stripes on cards and overuses the crap out of them. once you see it you can't unsee it

You're right to question this and OP shouldn't have done it.

Worse beach looks

Interesting, we had developed a very similar framework for our internal agents: https://github.com/apoxy-dev/clrk For us main use-case was intercepting all network I/O including LLM providers, HTTP, and random TCP/UDP calls

load-bearing was a meme phrase used sarcastically on twitter long before claude started overusing it.

> A model is not an appropriate solution

that's a bold statement without any supporting argument. more like an opinion really. specially trained classifiers can be very good. sentiment analysis has been a thing for a while now, there had been entire businesses built on this tech where being even slightly wrong can be very costly

they did rebuild their serving layer from fastapi to some rust thing so...

there is a number of different methods. One is a separate model/classifier trained with human annotation. Here's a paper by OAI: https://arxiv.org/abs/2305.20050

That depends on the bylaws and specific contractual director-designation rights they may or may not have. Ianal but if he really controls 84% of shareholder votes like he claimed, he should be able to alter the board composition legally. Probably enough to get him unfired

If we’re comparing it to human thought process then there’s famously s1 and s2 thinking with former being more like looping in the latent space and s2 being more like cot. Personally im not convinced the way llm “think” is all that analogous to real thinking and cot and looped transformers are just neat tricks to tease out pre training data better

turns out that's not much cheaper and really limits design choices and sometimes house locations

wouldn't that imply some sort of "subconscious" reasoning ability that's hidden? and if that was the case why would we need cot in the first place?

> There really is no way to control them

it seems like it should be feasible to analyze cot trajectories that lead to such behavior and punish them retroactively so not sure it's impossible but maybe someone just hasn't tried enough

Ime it remains the best model for prototyping, sims, visualization etc throwaway scenarios. Just don’t look at the generated code if you just let it rip on the problem for hours ;) but that advice applies to any model unfortunately

Wouldn’t finding such bugs be considered advantageous? That is unless someone just yoloing the result directly to customer ofc

Are you aware of the syzkaller and how many bugs it still finds including in networking stack?

Yes, lots. Early spdy was vulnerable to CRIME attacks and so was TLS1.0 (and to bunch of other attacks). Kaminsky DNS cache poisoning is another one that comes to mind

Lmao that output genuinely gives me the creeps. What's you opnion vs. SGLang in this regard?

Even 25 years ago being “too online” - back then it was icq, livejournal, and forums - was already destructive as hell and always led to some pretty cringe outcomes. At least these days some people make a career out of it, back then it was purely negative

Very poor analogy - information security has very few commonalities with ballistics (duh)