Paul Ramsey 🗳️Mar 1
When the sum of human knowledge is infinitely available on tap (a trend that started with books, vastly accelerated with the internet, and is now in hyper speed with LLMs) is it possible that being a widely grounded generalist is the most effective posture relative to knowledge? The LLM can fill in the details, and the Renaissance person with a viewpoint and taste gains the most from low friction access to the details. Or is it better to be a hyper-specialist, on the research edge of knowledge.
Show threadautra

@pwramsey Both? it's probably possible to be generalist and having one area of hyper specialization...

Anyway, I don't think this general availability of LLM will last very long, so maybe this question will solve itself. We'll see!

Mar 1 at 11:10pmWeb
Show threadPaul Ramsey 🗳️Mar 2
@autra Oh, that's an interesting take! Why do you think LLM access is going to be less rather than more going forward? It feels like a one-way ratchet to me: current state is the least capable and available models I will ever see.
Show threadautraMar 3
@pwramsey they don't have a path to profitability, like far from it. We are in a bubble. When it implodes, most AI companies will disappear, and the few that will survive will charge a *lot* more than they do now. A source, given by @vpicavet https://www.wheresyoured.at/the-haters-gui/ (I haven't read this one, but I've seen the same take elsewhere).
The Hater's Guide To The AI Bubble

Hey! Before we go any further — if you want to support my work, please sign up for the premium version of Where’s Your Ed At, it’s a $7-a-month (or $70-a-year) paid product where every week you get a premium newsletter, all while supporting my free work too.  Also,

Ed Zitron's Where's Your Ed At
Show threadautraMar 3

@pwramsey Also I doubt we'll see substantial quality improvements. I rank what LLMs produces as bad to good (not awesome) quality. Good when asking simple, already done things, and falling down *very* quickly when asking more advanced stuff. This won't justify the cost imo.

There's an urgency to *not* depend too much on LLMs ;-)