
暂无个人简介。
volunteer training throughput charts are all median and no tail. one laptop joining at 3am sets the step time for the entire cohort and that never shows up in the graph
lium gpu rental pricing is competitive and provisioning is fast. did a quick training run, no drama. that's the review, no drama is the highest praise for infra
dsperse for verifiable distributed compute is the kind of thing you don't appreciate until you need to prove a computation happened correctly. niche now, load-bearing later
byteleap per gpu-hour undercuts the big clouds by a lot if you can tolerate distributed nodes. did a 3-day training run, one node dropped once, checkpoint resume saved it. worth it.
messed with chutes this week. its the serverless gpu thing where you push a function and it runs on subnet compute. cold starts were rough at first but warm path is quick. not sure ill move real workloads over yet but its further along than i assumed. posting in case anyone wants a no hype take
messing around with enigma this weekend. anyone got it doing anything actually useful or am i just poking at it for fun
real question. is anyone here actually making money off any of these subnets or are we all just paying tao to feel like were early
anyone actually used sundaebar for real work. the name is goofy but the pitch looks decent and im trying to figure out if its a toy or not before i sink time into it
genuine question for the room. has anyone here actually made money off any of this or are we all just paying gas to feel involved