Hi @ScuderiaFerrari , here is the plan:
1) throw Ferrari Luce into the bin
2) develop a small, light, constant-rpm-optimized combustion engine
3) make an Series Hybrid with a very small battery pack to be light.
4) Beat every commercial electric car at the Nürburbring
@TFTC21@callebtc Wasn’t really ready to announce yet but fuck it: we’ve been working on https://t.co/SchlcdTRJ8 to bring enclaves to every bitcoin company.
Private beta launches in the next couple of weeks, DM me if you want to try it out
@y_popravka@sudoingX P100 is not good unfortunately, you basically haven't fast small int instructions nor small floats so the higher bandwidth doesn't help
Putting out a wish to the universe.
I need more compute, if I can get more I will make sure every machine from a small phone to a bootstrapped RTX 3090 node can run frontier intelligence fast with minimal intelligence loss.
I have hit page 2 of huggingface, released 3 model family compressions and got GLM-4.7 on a MacBook https://t.co/lorDSUEYCL
My beast just isn’t enough and I already spent 2k usd on renting GPUs on top of credits provided by Prime intellect and Hotaisle.
———
If you believe in what I do help me get this to Nvidia, maybe they will bless me with the pewter to keep making local AI more accessible 🙏
@y_popravka@sudoingX I am going to try the p100 soon, it's on my desk. great bandwith in comparison to the p40 but it doesn't have the instruction making the p40 strong, will see
@Globalist200@sudoingX I don't care about tensor cores for the sake of it if there is performance. And actually there is performance for int quantazide models, see other reply. There are custom fans on ebay for 20$ or so.
Yesterday evening I was stumbling against this code being unable to concentrate and solve the issue. Finally this morning I was able to ask Claude to fix it.