karolist
Joined 1,353 karma
- karolist parentMy old BMW GS (motorcycle) engine jumped timing chain for whatever reason and completely gave up the ghost on a busy highway in Poland, a random motorcyclist on a streetbike saw the white smoke cloud and understood what happened to me as I coasted to the side of the road, he stopped, asked a few details and said "get on your bike, I'll push you to the nearest shop". I didn't speak Polish and his English was not great too, but we managed to understand each other. To this day I do not know how he managed, but he was able to control the gas on his bike, align himself to the back and off the side of my bike and kept giving me these push impulses so I could keep moving, and moving we were, we travelled ~6 miles to safety doing ~30mph, I thanked him and we went out separate ways. This experience feels surreal, I literally didn't spend any time in danger lingering on a busy road side or waiting for a trailer, which can take hours.
- You needed software to replace rear brake pads on Audis with electronic parking brake since about 2004 or whenever they introduced C6. It's not a big deal and could be done with VAG-com, but that means any small garage with mechanic who can just turn nuts and bolts won't be able to do it.
> https://www.audiworld.com/forums/a6-s6-c6-platform-discussio...
"Yes. You need vag to disable the electronic parking brake in the rear. The piston cannot be pressed into the caliper if the park brake is not disabled. "
- I love this little thing. In case someone uses ESRack system, here's the bracket I designed you can print https://www.printables.com/model/1359418-esrack-module-jetkv...
- do you mean https://github.com/pollockjj/ComfyUI-MultiGPU? One GPU would do the computation, but others could pool in for VRAM expansion, right? (I've not used this node)
- There's also https://capibarazero.com/
- is this a wrapper to https://github.com/Tencent/Hunyuan3D-1/?
- this is really cool and I'd love to use it, but it seems they only support workers written in Go if I'm not mistaken? My workers can be remote and not using Go, behind a NAT too. I want a them to periodically pull the queue, this way I do not need to worry about network topology. I guess I could simply interface with PG atomically via a simple API endpoint for the workers to connect, but I'd love to have the UI of riverqueue.
- I gave them a year of subscriptions before cancelling recently, the devcontainer implementation in their Ultimate versions is laughably bad, bugs upon bugs and tickets where their support staff just bounces it up with "still no fix" messages and customers are finding workarounds, i.e. downgrading docker installs.
Remote SSH is terrible too, handles network latency spikes by repeating keystrokes. I remember spending an evening trying to fix something in the integrated shell and giving up, but sadly forgot what. I like what they do with Go though. Anyway, back to nvim here, not for me.
- See also https://www.breakneck.dev/blog/no-tech-cofounder and related HN discussion https://www.hackerneue.com/item?id=39902372
- I wanted to use this process (LLM -> OpenSCAD) a few months ago to create custom server rack brackets (ears) for externally mounting water-cooling radiator of the server I am building. I ended up learning about 3D printing, using SolidWorks (it has great built-in tutorials) and did this the old fashioned way. This process may work for refining parts against very well known objects, i.e. iPhone, but the amount of refinement, back and forth and verbosity needed, the low acceptance rate - I do not believe we're close to using these tools for CAD.
- Yes, but that's just his personality. His mind seems to be racing at 200mph whilst the output device (hands, keyboard etc) can't keep up, so some context gets dropped here and there. I remember I had a hard time watching his streams because he'd type at 160wpm or so, but half of the keypresses were correcting mistakes...
- Fair point, thanks for clarification, it seems this was first proposed in https://arxiv.org/pdf/2405.04434? I was confused by your title mentioning DeepSeek but then first paragraph revert to "...language models like ChatGPT and DeepSeek faster at generating text".
- I'm not sure what's going on with TabbyAPI's github metrics, but exl2 quants are very popular among nvidia local LLM crowd and TabbyAPI comes in tons of reddit posts of people using it. Might be just my bubble, not saying they're not accurate, just generally surprised such a useful project has under 1k stars. On the flip side, LLMs will hallucinate about TabbyML if you ask it TabbyAPI related questions, so I'd agree the naming is unfortunate.