Not at all an expert but I believe it's possible to get started experimenting with just a simulated robot in the simulated world model. While the full workflow is to generate training data to drive a real robot in the real world, without closing the loop, you're just lacking the ground truth data to quantify the divergence between simulation and reality.
There are all kinds of hobbyist robotic armatures at various price points but my understanding from a friend in this space is that the precision, durability and repeatability for serious applications starts at around $30,000 to $50,000. He mentioned the Franka Research 3 (FR3) as one example (https://franka.de/), perhaps driven by something like a Jetson AGX Thor ($5,000 and up).
As always, there are many less expensive and DIY-ish recipes to get started on smaller budgets. My friend's suggestion was more the baseline experimental lab system for a big company wanting get started with something that could, in theory, scale to light industrial internal deployment.
Interesting! Yeah, that's quite expensive. For some reason I was thinking some of the useful cobots I had seen in the past were in the $10k range, but from a brief look around, that doesn't seem to be the case.
An additional complication is that MI250Xes are two GPUs in one package, so you need to connect the first and last x16 SERDES groups to the host, otherwise you'll only see one GPU (or it won't work at all, idk).
Also, the cheap HPE pulls on eBay need some proprietary HPE magic to work, and I have yet to see anyone figure that out.
That’s how industrial VFDs do motor overload protection, they just keep track of heat accumulation (based on current) vs dissipation (based on fan speed) and fault when a threshold is reached. Probably there’s more nuance to it, but that’s the gist of it.
Seems like there's a big risk of having that habit leak into human conversation. A lot of people try really hard to train themselves not to add those fillers.
reply