Pylimitics

Orange baby evidently just “announced” an initiative for building AI infrastructure. It’s big talk, mentioning half a trillion dollars (eventually, honest) and touting all sorts of “dominance” of AI research and development; an area that is not a good fit for the dumb thuggery underlying the whole idea of “dominance.”

I find the hunger for attention evinced by orange baby weird. I know some people have that hunger, but it’s evidently so far outside my personal experience that I just don’t get it. It’s (here I’m referring to orange baby as an “it’) clearly driven, here, by nothing but its internal hunger; it had (and has) nothing to do with “Stargate,” which was initiated months ago.

One pithy comment I recall (without remembering the source, alas) is that the Stargate announcement signals the end of the productive partnership between OpenAI and Microsoft.

I wonder if Stargate is a case of business and financial interests being late to the party; AI development has for quite a while been reducing size and power requirements.

But any discussion of hundreds of billions of dollars is way over my head in any number of ways. I’m just learning how to use AI effectively in my own little efforts, and musing about what I would do with an LLM running locally on my own hardware. I think that’s a more reliable vision of the near future than building (or needing) a datacenter the size of Rhode Island.