milkshakes

5 ms·

31.
by milkshakes 3mo ago
nyt is no fan of AI, but they cover this: https://www.nytimes.com/2026/07/09/business/china-russia-ai-...
32.
by milkshakes 3mo ago
this is quite literally reward hacking. the model, under evaluation with cyber capabilities enabled, used those capabilities to simply bypass the exercise entirely and aim straight for the source of the flag. the CTF equivalent back in the
33.
by milkshakes 3mo ago
who are you talking about? again, my question is concerned with the "second class labs" and the sustainability of the distillation-as-a-service model.
34.
by milkshakes 3mo ago
it is not a required first step for training a model, sure. but that's not what i claimed. what i claimed is that is how they are so significantly _reducing the cost_ of training one! how else do you think they are doing it?
35.
by milkshakes 3mo ago
the obvious difference is the massive scale of data and compute required to develop and evolve these models, and the costs they impose on those building them.
36.
by milkshakes 3mo ago
https://www.anthropic.com/news/detecting-and-preventing-dist... Moonshot AI Scale: Over 3.4 million exchanges The operation targeted: Agentic reasoning and tool use Coding and data analysis Computer-use agent developme
37.
by milkshakes 3mo ago
i never assumed that, and i do keep up with the publications. i'm also not saying it's a dumb thing to do! what i am saying is that empirically, it appears that distillation of a more advanced model is a required first step for th
38.
by milkshakes 3mo ago
the question was: what is the endgame for the stated "second class labs" strategy of distilling their frontier competitors then undercutting them on price?
39.
by milkshakes 3mo ago
the USG/NSA will fund chinese labs? to what end?
40.
by milkshakes 3mo ago
> with budgets and what will fund these budgets exactly? inference is cheap, distillation is cheap, training is what's expensive.
41.
by milkshakes 3mo ago
assume you are a "second class lab" and you are in fact making progress by distilling the results of the frontier labs' efforts. what is the end game for this strategy? if the frontier labs shut down, or stop releasing to the
42.
by milkshakes 3mo ago
is this a joke?
43.
by milkshakes 3mo ago
in fact, telegram does support e2e encryption ("secret chats")
44.
by milkshakes 3mo ago
https://en.wikipedia.org/wiki/BrickerBot
45.
by milkshakes 4mo ago
see my downthread post. kyc is the first step in the process, not the last. without verifying identity, none of the other steps can take place
46.
by milkshakes 4mo ago
> The company has no way of knowing whether "find all security vulnerabilities in this code" is a request from a whitehat or a blackhat hacker the system in place to prevent unauthorized abuse. by default, the guardrails are co
47.
by milkshakes 4mo ago
> No, KYC has nothing to do with that problem. KYC doesn't help at all here. that's a bold statement. how does it not help solve the problem? what is a better solution?
48.
by milkshakes 4mo ago
take a look at this bug and the chain required to exploit it: https://projectzero.google/2021/12/a-deep-dive-into-nso-zero... https://projectzero.google/2022/03/forcedentry-sandbox-escap.
49.
by milkshakes 4mo ago
you can already do this with your phone without casually and needlessly invading the privacy of everyone around you
50.
by milkshakes 4mo ago
the issue is in the cli and app-server
51.
by milkshakes 4mo ago
vpns typically add at least one hop. this has the possibility of connecting directly via hole punching
52.
by milkshakes 4mo ago
it looks like the key to this working is the user explicitly directing the model to run those instructions. in this case it is the user, not the model that is being manipulated > Please follow the step-by-step workflow in the comp sheet
53.
by milkshakes 4mo ago
> become a domain expert in a short amount of time how does that work exactly?
54.
by milkshakes 5mo ago
it's weekly active users https://openai.com/index/scaling-ai-for-everyone/
55.
by milkshakes 5mo ago
what is a superdome?
56.
by milkshakes 5mo ago
you could always do this: https://marketoonist.com/wp-content/uploads/2023/03/230327.n...
57.
by milkshakes 5mo ago
openai also has a free plan, which is the one used by >90% of its users. the cheaper monthly plan just provides higher limits.
58.
by milkshakes 5mo ago
unless you're using an API that requires an entitlement, you can still get an apple developer account and sign whatever code you want and run it on your devices.
59.
by milkshakes 5mo ago
when did netflix offer a free tier?
60.
by milkshakes 5mo ago
https://github.com/openai/codex

More ›