felixgallo

6 ms·

61.
by felixgallo 4mo ago
You said these groups have access to LLMs. So what? Mythos/Fable are a step change above most LLMs. Responsibly limiting access and easing it up over time safely is the sane move.
62.
by felixgallo 4mo ago
I'm a cybersecurity researcher! Can you explain how Anthropic is just hamstringing the good guys?
63.
by felixgallo 4mo ago
This is a clickbait article with a garbage title. From the actual article, the one quoted cybersecurity researcher is sane about it: “But it is understandable as we are still in the early days and they are still adapting their guardrails.
64.
by felixgallo 4mo ago
'their customers, in aggregate, want that' citation needed
65.
by felixgallo 4mo ago
what do you mean by 'security related concerns'?
66.
by felixgallo 5mo ago
all unreliable tools are attackers. Even if you're using well-aligned LLMs like Opus, you should assume that any input you give it -- including all dependencies from npm, etc. -- are at risk of compromise, which could result in attemp
67.
by felixgallo 5mo ago
You should not be using docker with LLMs. You should be using VMs, which have a much, much smaller attack surface than Docker, and significantly more reasonable defaults.
68.
by felixgallo 5mo ago
So you managed to find a single anecdotal blog post, and you're deriving that to mean that Anthropic's models are 'more misaligned in practice'? '“My vibes don’t match a lot of the traditional A.I.-safety stuff,” Al
69.
by felixgallo 5mo ago
Sam Altman is so cartoonishly, over-the-top sociopathically shady that he makes JD Vance look like Benjamin Franklin. I mean, honestly, tricking third world people into retinal scans in order to get a scam crypto coin? Anyone using OpenAI
70.
by felixgallo 5mo ago
well that's the problem, right? There is no justification for a trillion dollar Elon Musk valuation. And he and his investors know this. That's why they're trying to change the rules to dump the stock while it's irrat
71.
by felixgallo 5mo ago
I worry that cybersecurity as target is all fine and good, but it’s looking for your keys under the streetlight. We are all familiar with computers. The problem is likely to be humans, especially in automated programmatic manipulation. The
72.
by felixgallo 5mo ago
Out here in the actual demonstrated world, OpenAI has been leaking quality people like a sieve, has not yet demonstrated anything remotely similar to 'taste', and is led by a sociopath ( https://www.newyorker.com/ma
73.
by felixgallo 5mo ago
No, containerization has nothing to do with secure separation and never has.
74.
by felixgallo 5mo ago
Music, medical, engineering. Wherever the waveform consistency and noise floor are the optimization point.
75.
by felixgallo 6mo ago
someone stole Biden's daughter's diary, which revealed that she had battled a substance abuse problem in the past, and that's disqualifying to Biden exactly how?
76.
by felixgallo 6mo ago
That’s approximately what I mean. Although in comparison to the alternatives of Altman and Musk, the distinction is stark and significant.
77.
by felixgallo 6mo ago
I'm in the "AI could be good for humanity" camp, and in this camp, we believe that Dario/Anthropic is a radically better choice going forward than the alternatives at this moment. In this camp we are not 'fooled in
78.
by felixgallo 6mo ago
Thing is, if you're using Codex, you're supporting Sam Altman and the idea of Sam Altmans, in the same way that if you use X or buy a Tesla, you're supporting Elon Musk and the idea of Elon Musks. That's a pretty big ta
79.
by felixgallo 6mo ago
you're not wrong, but anecdata is not data. Here's some more data: https://marginlab.ai/trackers/claude-code-historical-perform...
80.
by felixgallo 6mo ago
the 'general degradation' is a myth. Check out https://isitnerfed.org/ .
81.
by felixgallo 6mo ago
I mean, that's what humans have always done with our morals, ethics, and laws, so what alternative improvement do you have to make here?
82.
by felixgallo 6mo ago
https://isitnerfed.org/
83.
by felixgallo 6mo ago
https://isitnerfed.org/ in short, it looks like nothing has been nerfed, but sentiment has definitely been negative. I suspect some of the openclaw users have been taking out their frustrations.
84.
by felixgallo 6mo ago
it would be interesting to measure the airflow with a decent anenometer that works at low airspeeds (I've had success with https://www.amazon.com/BTMETER-Handheld-Anemometer-Velocity-... ). My intuition is that the 5 ca
85.
by felixgallo 6mo ago
So is declaring that you won't abide by the Geneva Conventions, targeting civilian infrastructure and double tapping a girls' school, but here we are at the logical conclusion of the dumbest war in centuries.
86.
by felixgallo 6mo ago
This is brilliant work, guys. Did you get any pressure to soften or spike the story?
87.
by felixgallo 6mo ago
You misunderstand. Anthropic is now enforcing its rules against OpenClaw use in Claude Code subscriptions and requiring you to use API rates instead.
88.
by felixgallo 7mo ago
That’s a lot of words to pretend that you don’t know what the European Union is
89.
by felixgallo 7mo ago
It's fun to say that Anthropic doesn't have a business model, but clearly they do. Hopefully they can achieve it while maintaining their standards, even if in the eyes of some that's 'fumbling a strong lifeline'.
90.
by felixgallo 7mo ago
This is agitslop.

More ›