KeplerBoy

7 ms·

31.
by KeplerBoy 2mo ago
Which is not unreasonable. Just hosting it in the EU and promising not to retain / sell the data let's you charge a healthy extra and compete in many areas other players can't.
32.
by KeplerBoy 2mo ago
It's probably a version of this: https://dl.acm.org/doi/abs/10.1145/3306346.3323024 Subtle movement between shots can be used to create an image sharper than any of the inputs.
33.
by KeplerBoy 2mo ago
I wouldn't call one proper CUDA and the other one some dumbed down version. Nvidia really seems to be pushing for these DSLs to be first class within the ecosystem. In some cases probably even more cutting edge than the nvcc frontend,
34.
by KeplerBoy 2mo ago
To be fair the entry barrier got a lot lower over the past ~5 years. Now you can write very good CUDA kernels with just a few lines of python DSL code. Zero cpp boilerplate and zero explicit compiler calls. Stuff like Triton, nvidia warp (t
35.
by KeplerBoy 2mo ago
That's shifting a bit with electric cars I feel, because people are worried about used batteries. But yes, generally new cars are often just a bad deal.
36.
by KeplerBoy 2mo ago
You are not European, are you? I know plenty of working class people working ordinary skilled labor jobs who own their homes and drive decent cars. Some people choose to keep renting because we have strong laws guaranteeing renter's ri
37.
by KeplerBoy 2mo ago
That doesn't clarify it and AMD says as much "Because these estimates rely on published results, internal measurements and projection-based scaling factors, they are intended to provide directional comparison rather than direct me
38.
by KeplerBoy 2mo ago
AMDs marketing is just as sketchy. Their new Zen6 server page claims 3.3x performance per watt over vera on "agentic workloads" for a 100 kW rack. Maybe they are comparing a CPU heavy rack to an nvidia vera rack with 50 kW of GPUs
39.
by KeplerBoy 2mo ago
sounds like plain old prompt injection. These days chatgpt might look at 50 webpages when i ask it to research a topic. Seems quite possible that the final answer is influenced by such ads.
40.
by KeplerBoy 2mo ago
You don't an FPGA if you're taping out your own chips. But that is just a MMA accelerator with decent memory bandwidth. No secret sauce here.
41.
by KeplerBoy 2mo ago
Who cares about a few ms more latency on an LLM API? Maybe for voice, but most other use-cases are quite latency insensitive.
42.
by KeplerBoy 2mo ago
Open weight models tell a different story. Inference is not that much more expensive than what a 100$ plan would allow and will only get cheaper (for current capability models of course, frontier not so much).
43.
by KeplerBoy 2mo ago
you are seeing single realizations of non-deterministic processes. Confidently stating one model is better than the other is not really possible this way; that's just like stating one dice is better than the other because you rolled a
44.
by KeplerBoy 2mo ago
Is that really used professionally?
45.
by KeplerBoy 2mo ago
Sure, the point is that carbon credits are issued for a certain area of land to be and stay forested.
46.
by KeplerBoy 2mo ago
Then the area should have to be replanted without granting new credits. It's perfectly natural that trees die and are replaced by new ones in a forest. The unusual thing is just the entire forest dying at the same time.
47.
by KeplerBoy 3mo ago
The city burns quite a bit of fossil fuel for heating and transportation every day though. That certainly has effects on microclimates (think vents from the tube, which is getting ever hotter).
48.
by KeplerBoy 3mo ago
To be fair there's nothing else left under this sun.
49.
by KeplerBoy 3mo ago
Sir, this is a multi-billion dollar operation. There have to be some incentives.
50.
by KeplerBoy 3mo ago
is that renderer open source?
51.
by KeplerBoy 3mo ago
It's not that simple. Yes, newer standards push for higher frequencies to get more bandwidth, but 5G for example also uses the old sub GHz bands with excellent range and penetration.
52.
by KeplerBoy 3mo ago
They turn it into sugar, what's not to love?
53.
by KeplerBoy 3mo ago
Sure, I didn't imagine a utopia of joyful people talking to strangers. That's not what public transport ever looked like, but I do miss the times of a bit more diversity in how people spend their during the commute. Newspapers, bo
54.
by KeplerBoy 3mo ago
Getting a tiny phone not meant for media consumption is probably the closest you can achieve. You are not going to waste a lot of time watching youtube on a 3" screen, because that's just no fun. The "jelly star" phone l
55.
by KeplerBoy 3mo ago
> It's like people who say "it sure is strange that we had a 'once-in-a-century flood' two years in a row": that's not how odds/probabilities work. That usually refers to climate change and the fact tha
56.
by KeplerBoy 3mo ago
Isn't the RTX Spark from every OEM just nvidia throwing a bone to their decade long partners after nvidia crashed the regular PC market?
57.
by KeplerBoy 3mo ago
That doesn't sound like something one would want to do in cooperation with the entire rest of the world.
58.
by KeplerBoy 3mo ago
You can be a PI without having a PhD?
59.
by KeplerBoy 3mo ago
It's not legal just because the copyright owner doesn't immediately sue you.
60.
by KeplerBoy 3mo ago
You don't need a whole lot of software support if you just want to serve a single family of LLMs.

More ›