wyrdcurt

5 ms·

1.
by wyrdcurt 6d ago
Related from a couple days ago: https://news.ycombinator.com/item?id=49762835
2.
by wyrdcurt 8d ago
ZCode is pretty bloated anyway, in my experience. I used it for a while because Z.ai offers a subscription usage multiplier for using it, but despite that, I found myself hitting limits less often when I switched to Pi (and performance is t
3.
by wyrdcurt 14d ago
I didn't say there was a good reason, in fact I agree there is no good reason to share such videos for entertainment. I asked a specific question about your definition of gore, so I could try to engage with your opinion in good faith w
4.
by wyrdcurt 15d ago
What precisely is your definition of gore? Could you define it in a way that makes it illegal to share images of murders and car crashes as entertainment, without criminalizing, say, journalists documenting a war?
5.
by wyrdcurt 20d ago
It hasn't even been a century since nuclear holocaust became possible. Hardly any time at all on the grand scale. "They didn't" could just as well be "we haven't, yet".
6.
by wyrdcurt 23d ago
More optimistic take: we'll only be second-class for a few months, if the pattern of Chinese models catching-up holds.
7.
by wyrdcurt 23d ago
> they are both communities of people that others do not want to be around Speak for yourself. Only one of those communities is one I don't want to be around, and there are plenty of people who agree, apparently including the operat
8.
by wyrdcurt 24d ago
Can't imagine any reason why anyone who wants trans people to feel welcome would also want nazis and their friends to feel unwelcome... couldn't have anything to do with the fact that nazis systematically rounded up trans people,
9.
by wyrdcurt 24d ago
Pretty typical "cool HTML toy" LLM output, tbh. The only thing impressive about this is how fast it generated it (13 seconds is wild!), but that's more of testament to Google's infrastructural advantage than to the quali
10.
by wyrdcurt 28d ago
Interesting fact you may or may not already know: the pigeons most commonly found around human habitats are feral domestic pigeons. They generally aren't bothered by us because they're bred not to be! Corvids on the other hand are
11.
by wyrdcurt 2mo ago
"AI will probably, most likely, sort of lead to the end of the world. But in the meantime, there will be great companies..." - actual Sam Altman quote, the man is so unhinged he's beyond satire
12.
by wyrdcurt 2mo ago
No, that's backwards. OpenAI are the ones investing in Cerebras. Part of the deal is that they can't sell to Anthropic.
13.
by wyrdcurt 2mo ago
They are weights for matrix operations, so in principle you'd think they would be deterministic. In practice it's more complicated than that. Not to be snarky or dismissive, I mean this genuinely: ask an LLM about it. I currently
14.
by wyrdcurt 2mo ago
Except that even the exact same model won't output the exact same results, that's a fundamental aspect of how LLMs work. They're probabilistic/stochastic, not deterministic.
15.
by wyrdcurt 2mo ago
Eschatological is a perfectly fine word. The post is about doomsday scenarios, but perhaps the commenter is equally tired of people who think AI will be salvation! Anyway, if the word confuses anyone, they're in luck: dictionaries stil
16.
by wyrdcurt 2mo ago
So if the detector isn't perfect, it isn't useful? Not sure I buy that. Also, even if there's no way to detect what the activations are doing, we already have the ability to analyze your proposed threat statistically. If the
17.
by wyrdcurt 2mo ago
But that would happen with literally any model without grounding and is more of a quality/competence issue, not what's being discussed. It'd be a bit of stretch to conclude open models are no more auditable than closed models
18.
by wyrdcurt 2mo ago
Not sure I understand that position. Unless we're talking about a scenario in which one is using an outdated model along with no grounding (which, imo, PEBKAC), why would the model be pinning insecure libraries? If it does have groundi
19.
by wyrdcurt 2mo ago
I'm talking about using mechanistic interpretability to see the model's intent. If it is deliberately using compromised libraries to weaken some code's security, there's going to be a signal in its hidden activations tha
20.
by wyrdcurt 2mo ago
Maybe I should clarify. As I understand it, the kind of vulnerability being discussed is something like a Chinese model invisibly "realizing" that it's working on an American project, and then deliberately leaving subtle secu
21.
by wyrdcurt 2mo ago
In my opinion, the big issue with that argument is that advances in interpretability research and steering conceivably could, and probably will, render moot that (as of now, purely hypothetical) risk of subtle sabotage for open-weight model
22.
by wyrdcurt 2mo ago
"Strandfall is a sci-fi story set in a post-apocalypse, with an emphasis on climate change, co-operation, and adaptation." I'm in the early stages of designing a world/game focused the same topics, in part because I simp
23.
by wyrdcurt 2mo ago
As someone who has only recently been thinking about learning how to solder and work with PCBs that's actually useful info, thanks! Unfortunately I do have experience with the smell, being around others working with improper ventilatio
24.
by wyrdcurt 2mo ago
Sounds like you need a fume extractor!
25.
by wyrdcurt 2mo ago
True. The linked article's title says that. I wonder if that was a typo by the OP or one of those HN quirks where the title was automatically changed when it shouldn't have been.
26.
by wyrdcurt 3mo ago
It's wild if it's running at 400 fps because nobody has a screen that refreshes at 400Hz. Every frame rendered past the screen refresh rate is wasted compute. Easily solved by limiting the frame rate.
27.
by wyrdcurt 3mo ago
I'm not anti-AI or anti-data center, in fact I lean more towards "pro-AI" overall, but to say that everyone who claims data centers can contaminate water is lying is a strong claim that doesn't really hold up to scrutiny
28.
by wyrdcurt 3mo ago
Like I said, I don't necessarily agree with the idea and I don't feel strongly enough about it to really argue in its favor, but to answer the question: the same reason why OpenAI doesn't operate out of Sam Altman's gara
29.
by wyrdcurt 3mo ago
Or you might build a data center that poisons a community's water and drives up the cost of energy for your neighbors. We can't pretend there are zero negative externalities that accompany unconstrained compute. To be clear I&#x27
30.
by wyrdcurt 3mo ago
I don't know if it's a bad idea or not, but I'm struggling to understand how the idea as presented in the post would be a violation of fundamental human rights. Do you care to elaborate?

More ›