bufo

5 ms·

1.
by bufo 1y ago
Were you using pre-compiled headers before?
2.
by bufo 1y ago
Grateful that this one supports Windows out of the box.
3.
by bufo 2y ago
Brave.
4.
by bufo 2y ago
The plan is to have many, many Mechazillas.
5.
LLM Demand Is Currently Inelastic(fernand.pajot.net)
1 points by bufo 2y ago|0 comments
6.
LLM-Centric Software Paradigms(fernand.pajot.net)
1 points by bufo 2y ago|0 comments
7.
by bufo 2y ago
Because Go has massive traction both inside and outside of Google, whereas Dart/Flutter never got big traction.
8.
by bufo 2y ago
The RTX 4090 has about the same BF16 Tensor Core TOPs than the A100, assuming 50% MFU (like the A100 40 GB PCIe) it would take 8x longer on 1 RTX 4090 vs 8x A100 80GB SXM, so 12 hours. Datasheet here for the TOPs https://images.n
9.
by bufo 2y ago
I recommend Phind.com, it’s been much better and faster for me than Perplexity Pro. I typically use their custom 70B model but you can also use GPT4 o or Turbo, or Claude 3 Opus.
10.
by bufo 3y ago
It takes a while to take down job posts. Everyone likely learned the decision recently. I don’t think the employees who are going to be laid off care about updating the job posts at the moment…
11.
by bufo 3y ago
“Vice website is shutting down” is 100% accurate. Note that they will also lay off most of their staff.
12.
by bufo 3y ago
Was Google a chip designer before the first TPU?
13.
by bufo 3y ago
That doesn’t seem like a bad number.
14.
by bufo 3y ago
Seriously!!
15.
by bufo 3y ago
Way, way better than Tesseract!
16.
by bufo 3y ago
Kqueue! Not the same design or as flexible as io_uring though.
17.
by bufo 3y ago
Build 22000 is Windows 10 21H2.
18.
by bufo 3y ago
Actual benchmarks and useful info here https://youtu.be/WH-qtuVRS2c
19.
by bufo 3y ago
Oh yeah I meant io_uring too. Plus Windows copied it so you can implement things very similarly for Windows.
20.
by bufo 3y ago
Great! I was looking into something like this. I assume ending up with epoll will be better?
21.
by bufo 3y ago
It’s about 100 for x86_64 https://www.computerenhance.com/p/waste
22.
by bufo 3y ago
It was pretty hard to saturate the memory bandwidth on the M2 on the CPU side (not sure about the GPU).
23.
by bufo 3y ago
Tri Dao and Tim Dettmers ftw
24.
by bufo 3y ago
I don’t mind the slow burn at all! I however did not like being forced to do frustrating platforming / movements while having to start from scratch every time I run out of time or die.
25.
by bufo 3y ago
I had the same experience after the jellyfish, at which point I gave up and just watch YouTube to know what happens.
26.
by bufo 3y ago
I found the world and exploration very fun, but the “platforming” challenges were extremely frustrating for me, and I didn’t enjoy the random messages and the miscellaneous details that you translated. Basically the gameplay loop was filled
27.
by bufo 3y ago
I also love that feature for the exact same reason! Somewhat disappointed with Outer Wilds though ;)
28.
by bufo 3y ago
There is a difference. We train with large batch sizes these days. The ANE silicon size is tiny and can't do the large matrix multiplications for big LLMs with or without a batch size higher than 1. Meaning that it cannot saturate the
29.
by bufo 3y ago
Yes, you are correct in that the ANE does have the equivalent of tensor cores and that I didn’t mention that. I just don’t expect it to be usable beyond inference because the number of compute units will not work for batches in medium/
30.
by bufo 3y ago
This is completely wrong. Learn about lenses.

More ›