Post B2UUdnFIccpzk4aGAq by sjb@mstdn.io
 (DIR) More posts by sjb@mstdn.io
 (DIR) Post #B2UUdnFIccpzk4aGAq by sjb@mstdn.io
       2026-01-20T22:46:11Z
       
       0 likes, 0 repeats
       
       They find it's cheaper to run LLMs on the human brain than NVideo hardware, then Bezos and Musk pay you to live in the pods.  This is how The Matrix happens.
       
 (DIR) Post #B2UUdoY7m7URmkOo9g by wolf480pl@mstdn.io
       2026-01-20T23:01:52Z
       
       0 likes, 0 repeats
       
       @sjb I thought the entire value proposition of LLMs was to decrease the number of humans needed to do the job, because humans are expensive...
       
 (DIR) Post #B2UUksE4VJhZDONyi0 by sjb@mstdn.io
       2026-01-20T23:03:09Z
       
       0 likes, 0 repeats
       
       @wolf480pl No you don't understand, the point of civilisation is to run more LLMs - and then put the LLMs in space surrounding the entire Sun.
       
 (DIR) Post #B2UUnKVMKpqVV1dvl2 by wolf480pl@mstdn.io
       2026-01-20T23:03:37Z
       
       0 likes, 0 repeats
       
       @sjb I think you're confusing LLMs with DNA
       
 (DIR) Post #B2UUq2KsLNHa1A1vk0 by sjb@mstdn.io
       2026-01-20T23:04:07Z
       
       0 likes, 0 repeats
       
       @wolf480pl I don't think that's an option because CUDA doesn't run on RDNA.