jrflo

4 ms·

1.
by jrflo 5d ago
I would assume the skill set for implementing ads is quite different than the skill set for developing new models, so it’s quite reasonable to expect different teams are working on each
2.
by jrflo 8d ago
I get why these types of devices are appealing from an "oooh new shiny" perspective but let's be honest: the smartphone design is essentially perfected. Software keyboards are far better than physical ones at that size due to
3.
by jrflo 8d ago
It's a spectrum, I think a majority of people with that stance just think "AI datacenters consume water" and that's as sophisticated as it gets. A more realistic nuanced take would be that "AI is evaporating water w
4.
by jrflo 8d ago
I'm really disappointed by the lack of real information in a lot of these protests. The vast majority of people opposing AI that I've seen online and in real life all cite the same misinformation about water use, data centers crea
5.
by jrflo 8d ago
Where did you see this? The only claim I saw from the article was that 1 prompt uses 16oz of water, which is obviously false if you do any level of research
6.
by jrflo 8d ago
I think the difference is that the companies are dumping billions of dollars into transforming that data into something useful, so they would like a return on their profits. Opening up the models for free is not a good business model if you
7.
by jrflo 8d ago
Does hallucination matter for this application? We've moved beyond raw recall being that important, it seems like for law specifically all relevant facts will be cited and checked easily by humans.
8.
by jrflo 9d ago
Your graph shows that home sales have been at a constant rate for the last 3 years. They are way down from 2020-2021, when covid plus low interest rates caused a home buying frenzy, but this is not new. We've been in this regime for th
9.
by jrflo 9d ago
I'm referring to ZIRP, since the OP was referencing cheap cash during ZIRP
10.
by jrflo 10d ago
To be honest though cash hasn't been cheap for a while, not really since 2021. We have been in relatively high interest rates for the entire AI boom. Going from 350-375 to 375-400 won't be a huge shock for hyperscalers. Interest r
11.
by jrflo 10d ago
Home prices are sticky on the way down, 25 basis points won't change much
12.
by jrflo 10d ago
Seems pretty bold to claim deepseek is the "best hacking model" while providing zero comparisons to other models...
13.
by jrflo 11d ago
Agreed, it's probably 10% of the total product development process, but approximately half if you're just considering "design" tasks.
14.
by jrflo 11d ago
I've tried Sol on build123d and it was completely unusable, I should take another stab with Astra though.
15.
by jrflo 11d ago
I have tried, I really wish it would be better because CAD is a slog and I want it to be automated. It's pretty bad at real world tasks.
16.
by jrflo 11d ago
AI cannot physically build anything in the real world, what are you referring to? That's a huge part of mechanical engineering. And you can give it simulation tools, but those are far far away from accounting for everything in real-wor
17.
by jrflo 11d ago
I don't think so. As a mechanical engineer, CAD is only half of the design process. The other half is actually building and assembling your design and testing/debugging what you've created. You still need the mechanical knowl
18.
by jrflo 11d ago
CAD is a different beast than mesh modeling. For product design you need a parametric model that can have precise dimensions and is constructed mathematically step by step. Mesh modeling is just moving a pile of triangles around in a very i
19.
by jrflo 11d ago
What is the ethical concern in this use case?
20.
by jrflo 11d ago
Unfortunately that requires a level of technical competence that most government entities don't have in the US. Not to mention the ridiculous requirements/processes that are placed on official govt software.
21.
by jrflo 11d ago
43 years to break even on Qwen 3.8 at 25% the speed of the API, lol. I like the idea of local models for really small tasks like automation/toolcalling, but it will probably never make sense for coding. I tried them and it was just exc
22.
by jrflo 12d ago
I don't understand how "companies will be unable to monetize AI" and "AI is a dangerous technology that can cause significant economic disruptions due to cyberattacks and incur serious losses". It seems like any suf
23.
by jrflo 12d ago
Quick, someone tell the Trisolarans!
24.
by jrflo 14d ago
1) there would be no incentive to develop a new model (non-distilled) if it had to be released as open weights. 2) frontier models are way more dangerous if they are open. It’s opening Pandora’s box, there’s no going back once they are rele
25.
by jrflo 15d ago
It's not only Lean code, there are English writeups too. To my understanding the pipeline for these problems is 1) solve in english 2) formalize systematically to check. No one is tackling problems purely in Lean, to my understanding.
26.
by jrflo 15d ago
That doesn't seem to be true. The OpenAI NS paper was 166 pages. Wiles-Taylor proof of Fermat's last theorem is 129 pages. The length is not unprecedented for a difficult unsolved problem. To be honest, I feel like the difficulty
27.
by jrflo 15d ago
I don't think it really attacks human understanding though. You can still read and understand an AI written proof. If another person comes up with a solution to a problem, you can read their methods and understand it. It doesn't m
28.
by jrflo 15d ago
I don't think that AI would disrupt any of that. Even if AI solves a problem, you can still discuss the methods at conferences, seminars, lectures, chats in the hallway, advising students, etc.
29.
by jrflo 15d ago
Right. It seems like reading an AI proof (although it may not be well written) will provide the same insights as reading a proof from another mathematician, assuming it's been reviewed and edited, just like any human-authored publicati
30.
by jrflo 16d ago
The circle jerking of Chinese models on this site never ceases to amuse me.

More ›