Playing a bit with various GPTs (mostly Kagi's FastGPT and OpenAI's OG ChatGPT), two thoughts are primary to me:

First: It's a bit like Old Web, after Google first appeared, and before it was attacked by the advertising virus Doubleclick. You have a question, you ask it, and you're steered directly to the answer. OG (Origional, pre-Doubleclick, Google) surfaced results on a Web that was not yet infested with its own SEO creations. (Though of course initial search engines had created their own early forms of SEO enshittification, to use @pluralistic's subsequent formulation.

Now, GPTs answer your question directly, sometimes accurately sometimes not, to be sure, but without all the bullfuckery of paywalls, newsletter signups, cookie notifications, privacy policies, chumboxes, etc., etc., etc. With how absoute shite the Web's become, this is an amazing breath of fresh air.

Second: There's no way in hell this will last.

Both the GPT vendors and the scumbuckets who've destroyed all else about the Web over the past 35 years will lay waste to the momentary epistemic clarity that (sometimes) is AI giving, or hallucinating, a direct answer to a direct question. We're at that very unstable crest of "gee, this is sometimes kind of useful" before the crapfest gets into gear and slams into us like some Wiley Coyote of yore.

Enjoy it whilst it lasts.

#ThoughtsAloud #chatgpt #enshittification #QualmBeforeTheStorm