pilooch

I feel using these tools make me think a lot actually. It's to the point that the more I'm digging a research or code subject, the answers are so dense that I sometimes need hours to get around why this or that (a good example is designing ML systems, losses, theoretical and statistical stuff).

It might not be deep thinking, but it's definitely like being back on a PhD with an always on advisor that splits your brains every few thoughts.

Yes this should be turned off by default. The daemon is useful when building apps on top of the daemon server, e.g. pi or pi-durable custom harnesses.

For information, I just built a similar rtx 6000 pro 96GB server for a customer, it's ~17k euros, so 17 months rental under their reported monthly basic pricing.

Agreed, AI is of-shoring at home.

Apps may just well go away with llms and agents.

My goto private setup, runs ~50t/sex on a dgx spark with sglang, nvfp4. Excellent model.

Isn't this one mimicking the typesafe ai horror website ?

That's unless the code produced in the future is much more complex than today's.

The risk is they are a magnet for many more talented ML scientists to leave google.

Daily sports is the way, purges the body, rests the mind.

Yes, all my emails gyer sorted out by a finetuned gemma. There are turned into images passes to the model, as multimodal is so practical.