plz1

The data center is still needed, just not the same scale. You don’t run “small” local models to compete with Claude’s size. They don’t publish numbers, but estimates were like 1-2T params for the models like early Opus. You need a lot of hardware to run those efficiently at a scale an enterprise would need.

Or they just spent enough to upgrade developer desktops with RTX Pro’s to run quantized ~100B param models at their desks. Those are very good for coding.

Qwen 3.8 27B, running on a RTX 5090. Works so well.

This is why I went with Unraid. Being able to slap whatever drives in that I have on hand was the primary driver for getting away from btrfs (Synology). And that build was about 3 months before RAM prices started to explode last year, which I read as “all parts gonna skyrocket”, which they have.