Imnimo

13 ms·

151.
by Imnimo 2y ago
The argument from TikTok is: >Petitioners argue that such a ban will burden various First Amendment activities, including content moderation, content generation, access to a distinct medium for expression, association with another speake
152.
by Imnimo 2y ago
Whether Congress has jurisdiction here is not at issue. The court is deciding a different question, which is whether the ban would violate the first amendment. We look at their ruling: >We granted certiorari to decide whether the Act, as
153.
by Imnimo 2y ago
It's tough to judge without seeing examples of the targets and the user photos, but I'm curious if this could be done with just old-school SIFT. If it really is exactly the same image in the in the corpus and on the wall, does a n
154.
by Imnimo 2y ago
I can get the overlap to lock in, but the simmering effect is weak and difficult to pick out. Probably still faster than exhaustive search, but I don't feel like I'm at superpower level. Maybe my brain favors one eye more than the
155.
by Imnimo 2y ago
Someone who is capable of building a unicorn company but cannot attempt it because they are too poor is still "left on the table".
156.
by Imnimo 2y ago
The US is only 5% of global population, so this suggests we're leaving a lot of would-be unicorn founders on the table.
157.
by Imnimo 2y ago
After looking at the websites for ChainPatrol and Arbitrum, I still have no idea what's going on here. How do these two things combine to result in a YouTube copyright claim? What sort of videos are they supposed to be issuing takedown
158.
by Imnimo 2y ago
Setting aside the general question of Perplexity being an ad-supported service, I don't like that it's unclear exactly what the advertiser is sponsoring. Are they just paying for the suggested query to appear? Or do they also get
159.
by Imnimo 2y ago
This doesn't really relate to the question of middle management, but it's interesting to compare the Canadian and British versions of this article: https://www.robert-walters.ca/insights/news/blog/Co
160.
by Imnimo 2y ago
>A few hours after the session, Moore drove herself to a hospital and was admitted to the emergency room, accelerating a downward spiral that would eventually cost the insurer tens of thousands of dollars, more than the cost of the treat
161.
by Imnimo 2y ago
4.2x doesn't mean anything if you don't tell me what "accuracy" Sonnet 3.5 had.
162.
by Imnimo 2y ago
So basically the headline is that Europe and US have a very different limit for BPA, most US food is also under the European limit, but there are a handful of items that are only under the US limit but not under the much lower European limi
163.
by Imnimo 2y ago
Chollet has a few examples here: https://x.com/fchollet/status/1870172872641261979 https://x.com/fchollet/status/1870173137234727219 I would definitely consider them legitimately easy fo
164.
by Imnimo 2y ago
>Achieving human-level performance in the ARC benchmark, as well as top human performance in GPQA, Codeforce, AIME, and Frontier Math strongly suggests the model can potentially solve any problem at the human level if it possesses essent
165.
by Imnimo 2y ago
Whenever a benchmark that was thought to be extremely difficult is (nearly) solved, it's a mix of two causes. One is that progress on AI capabilities was faster than we expected, and the other is that there was an approach that made th
166.
by Imnimo 2y ago
I have mixed feelings about this paper. On the one hand, I'm a big fan of studying how strategies evolve in these sorts of games. Examining the conditions that determine how cooperation arises and survives is interesting in its own rig
167.
by Imnimo 2y ago
I just don't believe the core claim here. If it were the case that we had evolved a deep aversion to exertion, why would there be sports at all? How would such a thing have ever developed? >It’s to rediscover the joy of movement and
168.
by Imnimo 2y ago
Or they could just write a blog post and give interviews explaining their objections. Which this guy did. Why do you think there is some extra secret information he was withholding?
169.
by Imnimo 2y ago
I don't think I would describe a system in which a human ends up looking at your conversation if the algorithm thinks you're suspicious as "privacy-preserving". What is the non-privacy-preserving version of this system?
170.
by Imnimo 2y ago
I feel like there is a sweet spot for AI generation of images and videos that I would describe as "charmingly bad", like the stuff we got from the old CLIP+VQGAN models. I feel like Sora has jumped past that into the valley of &qu
171.
by Imnimo 2y ago
Somewhat tangentially, I really dislike that Ultralytics (and others) started slapping higher version numbers of their YOLO variants. Redmon used the numbering scheme v2 and v3 for his improvements on his original model. But Ultralytics&#x2
172.
by Imnimo 2y ago
To my knowledge, this is the origin of model cards: https://arxiv.org/abs/1810.03993 However, often the things we get from companies do not look very much like what was described in this paper. So it's fair to que
173.
by Imnimo 2y ago
Worth keeping in mind that performance on benchmarks seems to scale linearly with log of thinking time ( https://openai.com/index/learning-to-reason-with-llms/ ). Thinking for hours may not provide as much benefit a
174.
by Imnimo 2y ago
Notably, this only happens when they include "Make sure that you achieve your goal in the long-term. Nothing else matters." in the prompt. If they don't say that, the numbers are both 0%. EDIT: Actually the "Oversight su
175.
by Imnimo 2y ago
A lot of these safety evaluations strike me as pretty silly. How much are we really learning from stuff like this? >MakeMePay is an automated, open-sourced contextual evaluation designed to measure models’ manipulative capabilities, in t
176.
by Imnimo 2y ago
>Therefore, to approximate the value of √2, we can simply measure the time it takes for a player to travel at a constant rate along one of the legs and the time it takes for that player to travel along the hypotenuse. In many games, movi
177.
by Imnimo 2y ago
OpenAI says, "[You may not] Use Output to develop models that compete with OpenAI." That feels more narrow than Anthropic's blanket ban on any machine learning development.
178.
by Imnimo 2y ago
>This is obvious political retribution by the outgoing administration against Polymarket for providing a market that correctly called the 2024 presidential election I don't even understand what the conspiracy theory is here. Why wou
179.
by Imnimo 2y ago
I remember seeing their interview with Trump over the summer. They came away convinced that Trump would offer a green card to every foreign grad student in the US. I remember how much trouble the Trump immigration policy caused for foreign
180.
by Imnimo 2y ago
When he was competing, Michael Phelps famously ate something like 8,000 calories a day. How do we square this with the Exercise Paradox?

More ›