KeplerBoy

6 ms·

91.
by KeplerBoy 4mo ago
At least it seems to not get worse.
92.
by KeplerBoy 4mo ago
That's true. You can just point the agent at a python function and tell it to slap some fastapi around that and spin up a frontend.
93.
by KeplerBoy 4mo ago
Thanks for the correction. I should have looked it up; I only remembered it being somewhat odd.
94.
by KeplerBoy 4mo ago
It's inevitable. What might be a prosumer device today priced at 4000$ will be a regular consumer device in 10 years and models only get better. Local models today are fine for a lot of mundane tasks and will continue to be so. The us
95.
by KeplerBoy 4mo ago
To be fair the connectx-7 in the spark can't even push 2x200 Gbps since it is connected via 4 pcie lanes.
96.
by KeplerBoy 4mo ago
It means nothing. The LLM thought it was edgy.
97.
by KeplerBoy 4mo ago
Also RAM was still quite a bit cheaper when the DGX was announced back in early '25.
98.
by KeplerBoy 4mo ago
they also announced a GB10/N1X windows desktop mini PC.
99.
by KeplerBoy 4mo ago
The DGX also contains the 200 GbE networking and linux support.
100.
by KeplerBoy 4mo ago
Not at all. This is a more like what Apple has been doing the past few years. A bunch of decent arm cores paired with a beefy integrated GPU.
101.
by KeplerBoy 4mo ago
It's not that different in Europe depending on the route. Checked bags are a hassle and cost money.
102.
by KeplerBoy 4mo ago
It is completely different. FPGA tooling is not the same as a driver for a consumer product. A lot of the serious CUDA compute stuff is also not supported on all platforms (it's linux only, because why would you do such stuff on window
103.
by KeplerBoy 5mo ago
And that's probably necessary if you want a competent model. There simply isn't much norwegian literature on let's say banana farming.
104.
by KeplerBoy 5mo ago
How true is this statement: "He asserted that any country with its own language that did not have a sovereign LLM trained in that language was at a disadvantage as a globally trained, English-speaking LLM would not know about that coun
105.
by KeplerBoy 5mo ago
Interesting fact from the page: "The lowest drag coefficient in Ferrari history, achieved through aero-styling convergence, active air shutters, and ride-height logic that lowers the front by 10 mm even while cruising" I guess not
106.
by KeplerBoy 5mo ago
Interesting idea, but ultimately not going to happen (or matter). I doubt the latency in that DSP Pipeline is below a millisecond, heck given the state of non-critical automotive Software it might a second.
107.
by KeplerBoy 5mo ago
This style might have worked as an apple car. It sure as hell doesn't work as a Ferrari.
108.
by KeplerBoy 5mo ago
There are trillions to be made. That moat won't be as insurmountable in hindsight.
109.
by KeplerBoy 5mo ago
They just renamed it to antigravity ide. I don't think that product is deprecated.
110.
by KeplerBoy 5mo ago
Antigravity is just a vs code (more correctly: codeium) skin with Google telemetry and agent Integration. You can switch back to Microsoft's or cursor's flavor in minutes.
111.
by KeplerBoy 5mo ago
Isn't that much better than PayPal? Why would i want my money to end up in some intermediate PayPal account?
112.
by KeplerBoy 5mo ago
They seem to be doing let's say case studies on how AI based simulations can help industry. They did injection molding for example, and I'm sure they're testing similar approach to everything which can be modeled by PDEs, whi
113.
by KeplerBoy 5mo ago
Feels like the opposite. Stuff is still happening and you need to be part of a big lab to see it. NanoGPT is fun but at some point you need that datacenter.
114.
by KeplerBoy 5mo ago
Can you elaborate why you chose slang-splat over let's say Lichtfeld studio? How does it compare to the other splat-training tools?
115.
by KeplerBoy 5mo ago
If you don't focus stack and try to train on partially unfocused images, the optimizer will try to match the rendered view to be also partially unfocused. You would have to mask out the blurry areas for each image. I guess one could ju
116.
by KeplerBoy 5mo ago
I'm not sure i agree. The blobs are exactly where the surface appear to be because they are constrained by multiple viewing angles. Otherwise the splat would fall apart as soon as the viewing angle is changed slightly (Which it absolut
117.
by KeplerBoy 5mo ago
Thanks I appreciate the answer!
118.
by KeplerBoy 5mo ago
Probably link training. The very fact that it is a tens-gigabit-per-second connection in an environment full of RF noise (we all love our high powered switching power supplies) makes it hard to get a reliable signal instantly. Still shouldn
119.
by KeplerBoy 5mo ago
I wish there were any good literature on GPU Direct RDMA and GPU NetIO. Got any tips?
120.
by KeplerBoy 5mo ago
That's hardly a guide. It's the defacto documentation, you have to read this either way.

More ›