Comment on DeepSeek iOS app sends data unencrypted to ByteDance-controlled servers
oysterenjoyer@sh.itjust.works 4 weeks agoTrue, but you need powerful server in order to run the most capable Deepseek model, which most people don’t have.
Comment on DeepSeek iOS app sends data unencrypted to ByteDance-controlled servers
oysterenjoyer@sh.itjust.works 4 weeks agoTrue, but you need powerful server in order to run the most capable Deepseek model, which most people don’t have.
brucethemoose@lemmy.world 4 weeks ago
That’s an understatement. It won’t even fit well in 8xA100, you need an EPYC server to run it in CPU RAM, very slowly.
Hackworth@lemmy.world 4 weeks ago
To run the 671B parameter R1, my napkin math was something like 3/4 of a million dollars in hardware. But that (plus the much lower training cost) made this a millionaire’s game rather than a billionaire’s. Plus the distillations do seem better than anything else we have at the smaller sizes at the moment. All that said, I’m looking forward to the first use of deepseek’s methods with google’s Titan architectures.