AntiRush
Joined 16 years ago
Karma
1,584
Submissions
214
Member since
Sep 8
[email protected]
View on Hacker NewsRecent submissions
Comment · 4 days ago
What a great read. Thank you for sharing. I didn't go to my first lan party for a couple of years after this was written but it's a nostalgia trip nonetheless.
The linked lanparty.com is also a great exploration. (The first archive is from a few months after the article)
https://web.archive.org/web/19981212015213/www.lanparty.com/
I'm pretty sure my friends and I used this resource when setting up: https://web.archive.org/web/19981202184536/http://www.lanpar...
Comment · 9 days ago
Your book looks like it will be fascinating! I ordered through bookshop.org since I already have an account, but do you have a preferred retailer for others who might want to buy it?
Comment · 14 days ago
A3B refers to 3 billion parameters active during a forward pass, not the size of an individual expert.
The 3 billion includes all active experts, and then other shared weights, attention components, embeddings.
In the qwen 3.5 35b a3b series, each expert is ~3 million parameters. There are 40 layers which each have 1 shared expert and 8 routed experts. Roughly 1.1B of the 3B active are experts.
Comment · 25 days ago
Yes, you're correct - bad copy from my notes. 3,111 tok/s is the correct number.
Comment · 26 days ago
Agreed that glm-5.3-flash is a strong model. On a single 6000 pro you can (barely) fit a q2 quant in vram. I haven't used it extensively, but my initial feeling is that the q2 is quite a bit worse than q4 for this model. To run it comfortably at q4 with reasonable context length you really need 2 6000s.
When the qwen 4 series is released I am hopeful there'll be a strong model with the same architecutre. as qwen3.8-flash-next.
Comment · 27 days ago
I've been working on support for this model in ds4 on the RTX 6000 pro - it's been really great for my use cases. The ds4 q4 quant performs a lot better than other similar sizes that I've seen.
Using the Q4 quant on an RTX 6000 Pro Workstation Edition at 450 watts:
Code: prefill 1,251 tok/s decode 255.26 tok/s
Prose: prefill 1,251 tok/s decode 198.78 tok/s
Comment · 28 days ago
Cool, and quite reminiscent of 0x10c: https://en.wikipedia.org/wiki/0x10c
Comment · 1 month ago
This is likely a convertible debt investment from Amazon being revalued. The numbers are pretty close, there might be some smaller notes in there too.
https://www.sec.gov/Archives/edgar/data/1018724/000101872426...
Comment · 1 month ago
Back in 2010, during the early days of bwapi, there was a Brood War AI tournament held by the Expressive Intelligence Studio at UC Santa Cruz. It's interesting to see how different the approaches were back then, vs this or Deepmind's SC2 work.
https://web.archive.org/web/20091124210529/http://eis.ucsc.e...
There's a great contemporary Ars Technica piece by a competitor:
https://arstechnica.com/gaming/2011/01/skynet-meets-the-swar...
As an undergrad I did a project using genetic programming. It was not very successful, but it was a lot of fun.
https://tomisin.space/archive/starcraft-genetic-programming/
Comment · 2 months ago
I was reminded of this book today when I read some comments here about the popularity of “sailing log” manuscripts.
This one is written by Paul Lutus, of Apple Writer fame.