Steve WolgastFeb 8, 2024
Pavel A. Samsonov

AI has already been used to run scams, rip off artists, destroy search engines, and drown publishers under an avalanche of shit.

Now AI boosters found a new thing to enshittify: #UserResearch .

"AI research is better than nothing" is the latest in a long series of "bad research is better than nothing" arguments that miss the point of research in the first place.

#UXDesign #softwaredevelopment #ProductManagement

https://spavel.medium.com/no-ai-user-research-is-not-better-than-nothing-its-much-worse-5add678ab9e7

No, AI user research is not “better than nothing.” It’s much worse.

There is an invisible reader of every article, whose gaze counts for far more than 1 view, 1 read, or 1 “like.” That reader is, of course, the algorithm. Between search engines’ page-ranking…

Medium
Feb 7, 2024 at 8:29pmWeb
Show threadPelle S ChristensenFeb 7, 2024
@PavelASamsonov AI research ... someone pls tell them that they are outsourcing their critical thinking to a glorified copying machine with a lacking conceptual understanding of fact.
Show threadPavel A. SamsonovFeb 7, 2024
@pellechristensen This is me trying to do that - although I wrote about the same thing a year ago and it hasn't stopped them so maybe I am doing something wrong https://spavel.medium.com/gpt-isnt-going-to-replace-designers-but-it-s-coming-for-the-users-40bf0a0c08b3
GPT isn’t going to replace designers. But it’s coming for the users.

ML models have reached a degree of sophistication that has some people proposing to interview AIs as though they were users. This is not a *new* mistake: design has a legacy of struggling with "user proxies" that we must put to rest.

Bootcamp
Show threadPelle S ChristensenFeb 7, 2024
@PavelASamsonov there is no cure against "oh new and shiny" enthusiasm. It has to run its course and do some damage before some people get it.
Show threadJan DFeb 7, 2024

@PavelASamsonov I guess that often research is "Something that we collectively believe in to be correct", a tool of alignment (not alignment in a particular direction).

Thus, the function is more belief and cohesion than exploration.

If (and only if!) believed in, any other, less exploitative and less engery hungry mechanism would do.

Show threadPavel A. SamsonovFeb 7, 2024
@simulo @Miniver calls this decision-driven evidence-making and he is absolutely right
Show threadJan DFeb 7, 2024
@PavelASamsonov @Miniver
uh, interesting! Any write-up of this?
For me this was not even about a (already made) decision, just about any direction – a coin toss would work, too (if people believe coin tosses telling something they did not know before)
Show threadPavel A. SamsonovFeb 7, 2024
@simulo @Miniver I looked on his blog but couldn't find it, and my Twitter access has been broken for a while so I can't refer to the tweets since tweets are now logged-in only
Show threadCoach Pāṇini ®Feb 8, 2024

@PavelASamsonov

When anti-social weirdos refuse to establish human contact with messy humans, they consider the next best thing when a synthetic machine generates plausible (but inaccurate) facsimiles.

Show threadskryFeb 9, 2024

@PavelASamsonov I really hate this.

I’ve been contacted about once a month since the advent of GPT* by someone making a new AI tool for not doing my work. And they all want free feedback. I’m ignoring them.

The most ludicrous is the one with simulated users that supposedly determine whether your UIs are usable.