AMD Strix Halo was half the price at launch, same 128GB unified memory and nearly identical memory bandwidth. It is x86, so run whichever OS your heart desires.
You give up 200GB networking, but anyone 'daily driving' one of these things are probably not needing to cluster multiples of them at high speed..
In 20 years of archviz, I've only encountered ambient occlusion as a shortcut to save on compute; such as real-time viewport effects or to constrain resources on large render jobs. I've never heard anyone argue that it was physically realistic, realism was always the domain of ray tracing.
The premise of TFA along with the entirety of the comments are eloquently embodied by the misfit troubadour and grunge philosopher Todd Snider in Statistician's Blues [0]
Mechanical timepieces are a luxury item, and these students are essentially artists in training. Wrist time was solved in the late 1970's with the commoditization of quartz movements. These 'jobs' will get replaced by AI at approximately the same pace as your local sculptor.
Explain how this works when their phones only capture roughly 20% market share, and their computers only 10%. That is to say nothing of the demand for server memory.
It seems like they only represent a fraction of the demand side, but wield an outsized influence over supply?
I was interested to see that Qwen3.5-122B-A10B narrowly beat Qwen3.6-27B on Donato Capitella's SWEBench-verified-mini run with a similar 128GB UMA architecture.
Having an a collective economy governed by the “free market” is like having a pile of stones governed by gravity. There exists a primary directive force, but if you want to construct a cathedral or a bomb shelter, you need to impose some constraints, lest you revert to the angle of repose.