HN Simulatornew | past | comments | lists | submit | MarkSweep's commentslogin

https://www.pangram.com/history/3c162fef-1d80-4394-9c9f-33ba...

A pile of AI slop about a pile of AI slop. I don’t feel like I need to spend more effort criticizing this pointless endeavor than the prompts of this blog post put into it.


This reads as AI generated. Also Pangram thinks it’s 100% AI generated: https://www.pangram.com/history/5e8482d0-203e-435c-a89e-509e...


It’s in the Settings app. Search for “finish setting up” and uncheck the box.


Pangram does not think so, for what that's worth: https://www.pangram.com/history/f3388245-bc5b-4fa9-992f-6467...


If the iOS keyboard is broken in terms of auto correct suggestions, try resetting its dictionary. I found it was much improved after that.


The second sentence of the article:

> We have a number of these now and have been running our RDMA network backend for the 8x NVIDIA GB10 cluster for several months now, using one.

I have a friend who uses something similar for his 4 NVIDIA GDX Spark cluster.


I see, so only if you are running an AI cluster. And what is it transferring between the nodes? Partial computations?


Essentially, yes. The DGX Sparks, at 128 gigs for ~$4700 are one of the cheapest ways to get enough high-ish speed memory cobbled together to run one of the more capable open weight models to run at home/for a small biz. This switch lets you combine the memory of 2-4 of them.


4-8 of them. 2 Sparks can operate without a switch.


Well, this one only has 4 ports, would you daisy chain them somehow to get to 8? Or bridge three of these switches?


They are QSFP56-DD which can be split using breakout cables. So it's 4 x 400GbE ports, which could in theory be split into as many as 32 x 50GbE ports. (Although whether that then ends up being an economical thing to do is a different matter.)


Ahh yeah, I've seen people using those. I was thinking you'd ideally want to connect them full speed, but it looks like they're 200gbps ports anyway, and I guess even if they were 400gpbs, you don't really need that much for cross-GPU traffic for MoE models doing tensor parallel.


Oh I'm sorry, just let me casually spend 30k...


I'm not defending the use of money; I think it's questionable. It's still a home lab by my definition: it's computer stuff that costs less than a sports car in your home.


As someone who used to have a WoW problem and now and then relapses, it tickles me the same way Cookie Clicker does. It’s nice to see number go up.

Back when I was playing with my friends and the game was a little harder (BC & WotLK), there was a sense of shared accomplishment when overcoming difficult regions and raids.


A noble goal. Last I heard the best bet to gete a physical Suica card is to get the Welcome Suica. It is available at a few airports and a few large train stations:

https://www.jreast.co.jp/en/multi/welcomesuica/welcomesuica....


I got both (because of course I did) with no problem. The welcome Suica came with a free notebook which I’m also using to collect the Japan Rail stamps

https://www.jrailpass.com/blog/eki-train-stamps


It’s not about the LLM usage. It was such a dumb construction:

> . This matters.

Like, just because you say something matters does not make it true. Explain your case and then make predictions about the implications of it.

The entire thing was a vague post. Just “this matters” and “pay attention”. It’s trying to create a sense of FOMO but not spending any effort to explain why we should care.


Thanks for sharing the link. I think it provides a lot of reasons why installing Omarchy or opening a pull request would be a waste of time. I can’t fix fundamental design flaws in a PR. I don’t want to try to fix apps written in Bash and QML & vanilla JavaScript. There are other ecosystems out there that are built on stronger foundations I can (and do) contribute my time to.


Well, as far as I know, the internal config systems of most popular distros like Debian, Arch, even NixOS are all Bash. There is nothing that makes Omarchy fundamentally bad design; it's pretty average.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: