Mastodon Feed: Post

Mastodon Feed

jsonstein@masto.deoan.org ("Jeff Sonstein") wrote:

this pattern has happened over and over again in the history of computing…

https://www.nytimes.com/2026/09/04/technology/open-source-ai-anthropic-openai.html?unlocked%5Farticle%5Fcode=1.-1A.jpaC.Sj60Y0O1Afcz&smid=nytcore-ios-share

first, a new technology requires sharing the use of an expensive, scarce, and centralized resource… a “central sever”. gradually, hardware and software become cheap & powerful enough that the locus of operations distributes; the new tech shifts to the periphery and to the “client side”.

example: when I started doing statistics in the early 70s, I went to the campus “Computer Center” to punch in my data and programs and then to run them. by a few years later, hardware had become cheap enough and modems had become legalized and I was able toI purchased my first machine so I could sit at home while remotely keying in & ordering the running of my programs and data. and now? it all happens on my own machine at home or office. same pattern.

this pattern is in the process of being repeated now, but now it is in the context of LLMs emerging as useful software tools. look to the pattern, not the event. think about tomorrow, not last week’s unthinking headlines.

IMHO the rush to build out zillion-dollar centralized data centers for “AI” is a serious error.