kalkin

11 ms·

121.
by kalkin 1y ago
If you automatically lump anyone who makes an argument that AI is capable - not even good for the world on net, just useful in some tasks - into "the AI crowd", you will tautologically never hear that argument from anywhere else.
122.
by kalkin 1y ago
As far as I can see this only makes sense if they actually believe that AI will accelerate production code that people will want to pay to keep running. Maybe they're wrong about that, but it's got to be sincere, or the strategy w
123.
by kalkin 1y ago
Some days it seems like the most vulgar Marxist would be more rigorous about doing a who-benefits analysis before using that to dismiss an argument than the median HN commenter. Fly.io benefits from AI hype because... more slop code gets wr
124.
by kalkin 1y ago
https://docs.github.com/en/code-security/secret-scanning/sec... is one option
125.
by kalkin 2y ago
There's nothing in my comment to "exactly" in support of your point. If the US backs away from Europe it may be more likely to encourage Ukrainian surrender. Russia may be more likely to attack Poland. Germany may be more lik
126.
by kalkin 2y ago
But... they can? If Ukraine can fight Russia to a standstill, a fully mobilized Europe clearly at least has a shot. Unless it escalates to nuclear weapons and everyone loses... But you need at least an argument if you want to say that is le
127.
by kalkin 2y ago
I don't think the left has been automatically in favor of tariffs--or at least I'd appreciate evidence for that claim. The complaints about free trade deals that I've seen have tended to be about provisions that treated vario
128.
by kalkin 2y ago
No, a "view that humanity ought to be replaced" is Sutton's, not an EA view. I'm not quite sure how you read that otherwise, except that you seem very angry. I sure hope our alternatives are better than human extinction
129.
by kalkin 2y ago
> all of their lofty accomplishments get nullified by anyone I don't think it's a question of whether their achievements are nullified, but as you mention, how to weight the opinions of various people. Personally, I think both
130.
by kalkin 2y ago
Sure, and his other views - in the scope of his professional expertise but also quite relevant to, uh, other humans - seem relevant in an HN thread about the Turing award. This place isn't exactly restricted to technical discussion of
131.
by kalkin 2y ago
Ah, I read that as him saying the quality isn't as good-equivalent to a model with 50% less compute. But you might be right.
132.
by kalkin 2y ago
Where does he imply that it's 2x larger than Sonnet?
133.
by kalkin 2y ago
Seriously. I did not realize Gizmodo was at that point with opportunistic clickbait, but apparently it is. edit to expand: from the headline I thought this was going to be Bonta coming out against the argument that AI training is fair use,
134.
by kalkin 2y ago
Incentivizing people to build homes that are likely enough to burn down as to be economically uninsurable is an absolutely wild abuse of the term "solidarity". Solidarity is the idea that an injury to one is an injury to all, not
135.
by kalkin 2y ago
Half to GiveDirectly and half to a variety of local organizations near where I live (tbh, not all ones I would have selected, but it was a shared decision with my spouse).
136.
by kalkin 2y ago
I don't think it was particularly brave, even. I mentioned it just to try to normalize it. I'm not exactly living like an ascetic and I don't understand why people feel they need so much more.
137.
by kalkin 2y ago
> a lingering question of whether any of it truly matters, especially knowing I could stop tomorrow without significantly impacting my wellbeing. A suggestion for your consideration, or that of anyone in a similar position: give enough o
138.
by kalkin 2y ago
Right. There's no rule that if you get a hold of $XX million that you need to keep it. I'd argue, in fact, that there's a pretty good case you're obligated to give most of it away; you can keep single-digit millions and
139.
by kalkin 2y ago
This is obvious condescension, come on: > I don't trust _you_ would ever agree (Emphasis in original.) Not to mention: > The key point that you seem to be missing ... I'm going to drop this conversation here, but one more re
140.
by kalkin 2y ago
It might be true that pretraining scaling is out of juice - I'm rooting for that outcome to be honest - but I don't think it's "consensus". There's a lot of money being bet the other way.
141.
by kalkin 2y ago
It doesn't seem like that distinguishes the case where your "alignment" (RL?) simply failed (eg because the model wasn't good enough to explore successfully) from the case where the model was manipulating the RL process
142.
by kalkin 2y ago
That's a nicely clear ask but I'm not sure why it should be decisive for whether there's genuine depth of thought (in some sense of thought). It seems to me like an open empirical question how much world modeling capability c
143.
by kalkin 2y ago
So the answer is no, you cannot in fact cite any evidence you or someone sharing your views would have predicted this behavior in advance, and you're going to make up for it with condescension. Got it. For the record, I was genuinely c
144.
by kalkin 2y ago
> strong evidence that language in the models is a consequence of high-order cognitive thought to be good enough to take it very seriously What would constitute evidence of this, for you?
145.
by kalkin 2y ago
You say that "it's not enough for me" but you don't say what kind of behavior would fit the term "alignment faking" in your mind. Are you defining it as a priori impossible for an LLM because "their langua
146.
by kalkin 2y ago
If you can link to a specific example of this anticipation, that would be informative. I don't care about use of the term "alignment" but I do think what's happening here is more specific and interesting than "unint
147.
by kalkin 2y ago
No results for wizzwizz4 "GPT", although it does look like search results may be incomplete: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
148.
by kalkin 2y ago
I'd be very interested in an example of someone who said "it's just autocomplete" and also explicitly brought up the risk of something like alignment faking, before 2024. I can think of examples of people who've bee
149.
by kalkin 2y ago
The takeaway appears to be: > Iterative prompting, like Socratic questioning, may never produce true intelligence in the systems it engages with, but it reveals the collaborative nature of intelligence itself—an emergent property shaped
150.
by kalkin 2y ago
The question is supposedly whether things are better or worse, not whether they're "good enough" in some abstract way. If you think things aren't good enough for an average person in one of the statistically best periods

More ›