Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
skavi
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
18 ms
·
241.
▲
by
skavi
3y ago
> TL;DR: Rust is named after a fungus that is robust, distributed, and parallel. And, Graydon is a biology nerd. https://www.reddit.com/r/rust/comments/27jvdt/internet_archa...
242.
▲
by
skavi
3y ago
The “Aero” in “Frutiger Aero” is Windows Aero. And Microsoft’s Segoe is practically Frutiger. So more than being part of the era, I think it’s fair to say Windows Aero defined the aesthetic.
243.
▲
by
skavi
3y ago
Maybe you intend this, but that argument could be applied to many other product segments.
244.
▲
by
skavi
3y ago
Which goes to andrewstuart’s point; the Microsoft Adaptive Controller was released five years ago.
245.
▲
by
skavi
3y ago
So your experience isn’t representative of the work presented by this post? Or does llama.cpp use the same technique for quantization?
246.
▲
by
skavi
3y ago
to clarify, is that with https://huggingface.co/relaxml/Llama-2-70b-chat-E8P-2Bit
247.
▲
by
skavi
3y ago
can anyone comment on running the 2b quantized llama 70b on consumer cards like the 4090?
248.
▲
by
skavi
3y ago
dup: https://news.ycombinator.com/item?id=38539153
249.
▲
by
skavi
3y ago
Given the API is similar to existing libraries, I’m curious as to whether the performance is better with this one. And if so, what’s stopping existing libraries from being as fast. IIRC, PyTorch at least has a Metal backend. The README ment
250.
▲
by
skavi
3y ago
Intel’s Linux drivers are also open source IIRC.
251.
▲
by
skavi
3y ago
There are already many startups aiming to compete with Nvidia here (Tenstorrent, Groq, etc.). And of course AMD as well. What does Altman have to offer?
252.
▲
by
skavi
3y ago
I assumed reference counting and locking would be employed transparently as needed.
253.
▲
by
skavi
3y ago
I think it is not so crazy an opinion that surplus value is unfairly distributed.
254.
▲
by
skavi
3y ago
Personally wouldn’t consider that a good explanation. Would love if more details were released.
255.
▲
by
skavi
3y ago
I think people tend to enjoy creating art more than farming. So technology that makes it less feasible to be an artist is more worrying than agricultural automation.
256.
▲
by
skavi
3y ago
Anything shipping in a chip has to have been developed for years. And this particular GPU arch was meant to ship last year.
257.
▲
by
skavi
3y ago
I gave it a go: https://news.ycombinator.com/item?id=38111256
258.
▲
by
skavi
3y ago
> 30% is roughly the difference to be expected between N5 -> N5P/N4 -> N3. That was the original argument - "TSMC". Not true according to TSMC’s own messaging. Keep in mind that predicted power and performance improv
259.
▲
by
skavi
3y ago
doing a few sketchy conversions puts Metal compute performance closer in capability to an RTX 4070. M3 Max is ~170% M1 Max per some graphic Apple released. That puts it at ~95% the performance of a 6800 XT on the Metal charts. RTX 4070 is
260.
▲
by
skavi
3y ago
I like Helix, but I get annoyed by a really tiny thing. When you hold the left and right arrow keys, the cursor will go to the prev or next line once it reaches the beginning or end of the current one. Which is completely fine behavior, to
261.
▲
by
skavi
3y ago
It does nothing special compared to the rest of the field of strongly typed MLish languages.
262.
▲
by
skavi
3y ago
If I understand correctly, Flawless provides the WASM runtime. If that’s the case why can’t it be entirely hidden to the user? The system provides things like entropy and networking.
263.
▲
by
skavi
3y ago
Yes, and also this old now archived project which had a similar aim: https://github.com/nebulet/nebulet
264.
▲
by
skavi
3y ago
The first two points easily apply to WebGPU. Maybe not the last one yet.
265.
▲
by
skavi
3y ago
The Nexus One was built by HTC. The Galaxy Nexus was built by Samsung. The Nexus 5 was built by LG. Pixel is in house, of course, but IIRC, Google bought a fragment of HTC some time ago.
266.
▲
by
skavi
3y ago
geekbench 5 and 6 scores generally correspond well to spec2017 which is pretty much the industry standard.
267.
▲
by
skavi
3y ago
SVE is exactly what I've been interested in seeing illustrated like this :). It's absurd to me that Arm hasn't created this themselves. officedaytime's page is such a nice thing to have for x86.
268.
▲
by
skavi
3y ago
Qualcomm’s SoC includes a modem. Apple’s modems aren’t integrated.
269.
▲
by
skavi
3y ago
holy shit, i’ve wanted exactly this for some time now. here’s the excellent x86 equivalent (sse through avx512): https://www.officedaytime.com/simd512e/
270.
▲
by
skavi
3y ago
Apple’s efficiency cores are significantly larger than A5x cores.
More ›