curt15

4 ms·

1.
by curt15 3d ago
How do you normally verify the work product of a "few-shot" development process? Do you scrutinise the source code like with human developers? Or do you just run the test suite and click around the app to check if it seems to work
2.
by curt15 3d ago
Since OpenAI is undoubtedly also under economic pressure, why is it still hiring humans for Account Associates, Android Engineers, Data Scientists instead of demonstrating its AI prowess by automating those roles? https://openai.
3.
by curt15 3d ago
> He’s also always argued that mathematics is fundamentally about human understanding. It's also worth pointing out he didn't come up with this viewpoint just to "cope" with the headlines. Thurston articulated this wa
4.
by curt15 3d ago
> I thought the reason we were encouraging people to go into STEM, and providing clever people with large salaries, was that mathematical results were of practical value to the wider human race. In which case, it’s surely very good news
5.
by curt15 3d ago
> Math, on the other hand, was paid for because people and states believed progress in math might lead to meaningful improvements in other branches of science, and, in turn, in our lives. If this is better served by AI, should we keep pa
6.
by curt15 3d ago
Where do you draw the line between system packages and user facing apps? Some software defies such an easy categorization. If your default install doesn't come with docker and you install docker later for development, does that make do
7.
by curt15 5d ago
> So let’s say OpenAI cure cancer and put every cancer researcher out of work depriving them of intellectual satisfaction, this would also be a problem? It would certainly impact the field. Would you expect Fields medalists to cure cance
8.
by curt15 5d ago
> This is a demonstrable fact as no human has figured this out despite the problem being open for almost 100 years. That's not true. Alpoge and Buckmaster's related LLM-assisted blowup result ( https://news.ycombinato
9.
by curt15 5d ago
> I suspect that with six good mathematicians spending their whole careers primarily focused on it, and working together closely, Navier-Stokes might well have fallen already. > The perverse incentives of academia mean this has never
10.
by curt15 6d ago
Suppose there is an oracle that can tell you whether a particular result in maths or physics is true. Would research still have any value?
11.
by curt15 7d ago
If you transplant a world class mathematician into a chemistry lab, do you expect similarly ground breaking results? Do you expect a top chemist to make major advances in mathematics? The frontier labs are likely betting on lay observers (r
12.
by curt15 8d ago
A "massive amount of AI-powered effort" costs a ton of money. What's the return on investment for these frontier labs? Do headline-grabbing successes in mathematics translate to expected *profitability* in disciplines with mo
13.
by curt15 8d ago
The scientist allegedly making that request comes from a machine learning background. Perhaps he's not familiar with the culture in mathematics regarding authorship. That sort of squabbling over author priority would be unconscionable
14.
by curt15 8d ago
They're also counting on more casual observers to extrapolate optimistically from successes in high profile math theorems to the company's economic value.
15.
by curt15 8d ago
Hypo-dissipative NS != NS
16.
by curt15 8d ago
> But Tristan rejects the first, why? Because he thinks it's theft? But then why would OAI threaten him? Did the first offer also come with the outrageous condition that he exclude his co-author from the credit?
17.
by curt15 8d ago
Buckmaster is already an established world expert at these sorts of problems. Declining the Clay prize would hardly dent his career.
18.
by curt15 15d ago
This remains true: https://github.com/gorhill/uBlock/wiki/uBlock-Origin-works-b...
19.
by curt15 15d ago
Wouldn't enterprises investing in their own hardware be more likely to buy standard 4-GPU or 8-GPU servers? Apple machines aren't really designed for server racks.
20.
by curt15 20d ago
Thanks for the lulz. https://www.pbs.org/newshour/nation/journalists-who-wrote-ab...
21.
by curt15 21d ago
This post sure ruffled some feathers judging from its downvotes.
22.
by curt15 25d ago
>Isn't this how human brains work? We're just large probabilistic neural inference machines. A network of neural weights guided by past training. Artificial neural nets are merely cartoons of how people guess human brains work.
23.
by curt15 28d ago
> It's like power consumption: it's going to scale, so we need to control how it scales. I don't think your premise is obvious at all. Developed countries have stagnant or shrinking populations, and there is a general tren
24.
by curt15 1mo ago
A body of independent subject matter experts who command public respect is anathema to authoritarian governments. It's why authoritarian takeovers frequently start by targeting major universities.
25.
by curt15 1mo ago
"No build step and no application backend: the mesh, the solve and the audio all run on your own machine."
26.
by curt15 1mo ago
"What do you do for work?" "I find ways to defeat adblockers"
27.
by curt15 1mo ago
"No waiting on the same row, less contention."
28.
by curt15 1mo ago
> Doesn't that mean agencies like this are not free to set their own rules when a law is ambiguous? Agencies can say whatever they want but courts are not required to defer to them when this inevitably goes to a lawsuit.
29.
by curt15 1mo ago
Do you target macos or windows, which have always given users full control of software management?
30.
by curt15 1mo ago
So EU "digital sovereignty" still means hitching most critical functionality to Google?

More ›