Imnimo

6 ms·

31.
by Imnimo 5mo ago
Unsurprising from Google, but still bad. If Google has no right to object to a particular use, this is equivalent in practice to "any use, lawful or not".
32.
by Imnimo 5mo ago
Back when Arena was first announced, there was an interesting line in their write-up: https://magic.wizards.com/en/news/feature/everything-you-nee... >We've created an all-new Games Rules Engine (GRE)
33.
by Imnimo 6mo ago
Why should we think that pro-social capabilities are simply not expressible by weight-based ANN architectures?
34.
by Imnimo 6mo ago
>Tigers, hippos and SARS-CoV-2 also developed ”through evolution”. That does not make them safe to work around. Right, but the article seems to argue that there is some important distinction between natural brains and trained LLMs with r
35.
by Imnimo 6mo ago
>Unlike human brains, which are biologically predisposed to acquire prosocial behavior, there is nothing intrinsic in the mathematics or hardware that ensures models are nice. How did brains acquire this predisposition if there is nothin
36.
by Imnimo 6mo ago
It's interesting to me that OpenAI considers scraping to be a form of abuse.
37.
by Imnimo 6mo ago
So, if you look at the way the scoring works, 100% is the max. For each task, you get full credit if you solve in a number of steps less than or equal to the baseline. If you solve it with more steps, you get points off. But each task is sc
38.
by Imnimo 6mo ago
Suppose you construct a Mechanical Turk AI who plays ARC-AGI-3 by, for each task, randomly selecting one of the human players who attempted it, and scoring them as an AI taking those same actions would be scored. What score does this Turk g
39.
by Imnimo 6mo ago
It was neat to be able to try my own prompts and get a sense of what the state of video generation was. But I certainly never generated something that I thought I got real value out of on its own merits, and I still don't understand wh
40.
by Imnimo 6mo ago
The fact that the LLM appears to never assign an actual 0 or 10 makes me suspicious. Especially when the prompt includes explicit examples of what counts as a 10.
41.
by Imnimo 7mo ago
I think the problem for xAI is that it can really only hire two types of researchers - people who are philosophically aligned with Elon, and people who are solely money-motivated (not a judgment). But frontier AI research is a field with a
42.
by Imnimo 7mo ago
>What we’re learning from this episode is that the government actually has way more leverage over private companies than we realized. Who is learning this for the first time only now? Even just restricting ourselves to the current admini
43.
by Imnimo 7mo ago
My metric for this kind of stuff is: Did Glaze build the Glaze app?
44.
by Imnimo 7mo ago
If someone tells you they're going to handle US communications in a way that is consistent with the FISA Act, that is not a good thing.
45.
by Imnimo 7mo ago
None of those explanations are compatible with the pledge of solidarity in the We Will Not Be Divided letter.
46.
by Imnimo 7mo ago
I don't see how OpenAI employees who have signed the We Will Not Be Divided letter can continue their employment there in light of this. Surely if OpenAI had insisted upon the same things that Anthropic had, the government would not ha
47.
by Imnimo 7mo ago
Do you have to prove that your 3D printer cannot print a 3D printer which can print a gun?
48.
by Imnimo 7mo ago
Ha! I misread it as "Haiku Warrior" and so didn't make the connection. That makes a lot more sense!
49.
by Imnimo 7mo ago
Apparently Haiku is a very anxious model. >The anxiety creeps in: What if they have removal? Should I really commit this early? >However, anxiety kicks in: What if they have instant-speed removal or a combat trick? It's also
50.
by Imnimo 8mo ago
Well, typically for higher progressive taxes. Tariffs are typically a regressive tax.
51.
by Imnimo 8mo ago
>The randomly selected applicants Why would you want to randomly select here?
52.
by Imnimo 8mo ago
I am in the "did not read a single book" bucket, and have been for many years. I just don't like reading books, and never have.
53.
by Imnimo 8mo ago
>An advertising-based business model would introduce incentives that could work against this principle. I agree with this - I'm not so much worried that ChatGPT is going to silently insert advertising copy into model answers. I&#x27
54.
by Imnimo 8mo ago
>By contrast, intrinsic mortality stems from processes originating within the body, including genetic mutations, age-related diseases, and the decline of physiological function with age So we put genetic diseases in the bucket of intrins
55.
by Imnimo 8mo ago
In this sentence, Anthropic makes clear that "be hurtful" and "lead to public embarrassment" are separate and distinct. Otherwise it would not be necessary to specify both. I don't think this is the signal they shou
56.
by Imnimo 8mo ago
I am somewhat surprised that the constitution includes points to the effect of "don't do stuff that would embarrass Anthropic". That seems like a deviation from Anthropic's views about what constitutes model alignment an
57.
by Imnimo 8mo ago
Yeah, both directly and indirectly. Over time, "sponsored links" became more and more visually indistinguishable form organic results, and advertising incentives drove changes to the search algorithm.
58.
by Imnimo 8mo ago
>ChatGPT’s responses will not be influenced by ads I don't see why I should believe this.
59.
by Imnimo 9mo ago
>By default, the main thing to know is that Claude can take potentially destructive actions (such as deleting local files) if it’s instructed to. What do the words "if it's instructed to" mean here? It seems like Claude ca
60.
by Imnimo 10mo ago
It does seem to raise fair questions about either the utility of these tools, or adoption inertia. If not even OpenAI feels compelled to integrate this kind of model-check into their pipeline, what's that say about the business world a

More ›