Post #1066219

2026-04-10 11:40 UTC

RE: https://thepit.social/@peter/116376219055579156 I simply cannot get my head around why anyone would let AI anywhere near their data analysis work. Why would you add something to your work that can drop its accuracy by half? by 10%? by 1%? What would be an acceptable amount? What is the possible upside that would make this worth it? I&#39;m floored by the number of people in this field who take themselves all too seriously but are out there starting their whatever *dot AI* companies to get in on the grift, or say things like “and of course, AI” like it&#39;s both obvious and inevitable. And how is this so shiny that even educators have taken leave of their senses? What the f*k is the point of analyzing a dataset if you&#39;re ok with answers simply being incorrect? What are we even doing here?

Replies (15)

@ChemicalEyeGuy@mstdn.science 2026-04-10 12:06

@benfry #AI is #clankers 🤖 all the way down. #Resist #AIslop.

Open ##1066244
@notimetoplay@elekk.xyz 2026-04-10 12:18

@benfry To answer that question, you have to consider what managers were doing with your carefully crafted analytics before AI. My bet is on something like "pretend to look at the charts while holding them upside-down, nod wisely and do whatever they originally intended anyway".

Open ##1066245
@juliehuz@mastodon.world 2026-04-10 12:25

@benfry More people than we realize are PROFOUNDLY insecure about their ability to sling a few sentences together in a cohesive manner. Makes me think there are grade school teachers using a belt to reinforce grammar rules somewhere in the world

Open ##1066246
@lcwheeler@ecoevo.social 2026-04-10 13:42

@benfry I see scientists in various fields doing this too and it terrifies me. I think the best explanation is mental laziness, coupled with hype and a lack of understanding, and a refusal or inability, on the part of many people, to engage critically with these (or any) technologies. It's amazing how many people I have talked to who just accept LLM outputs without a second thought, listen to an explanation of why that's a bad idea, and then turn right around and do it again.

Open ##1066247
@numodular@c.im 2026-04-10 13:46

@benfry I blame The Jetsons. There's very little looking Rosie, nowadays.

Open ##1066248
@RustedComputing@discuss.systems 2026-04-10 14:40

@benfry It's a lot like hiring a human team member, but with no human accountability. https://en.wikipedia.org/wiki/Agency_cost

Open ##1066249
@JankaWessman@kamu.social 2026-04-10 14:55

@benfry It's like pretty much overnight we suddenly forgot all our standards for being able to check and evaluate and confirm the calculations and the logic of them and to clearly explain *what was done*. It's mind-boggling.

Open ##1066250
@SamiMaatta@mementomori.social 2026-04-10 15:12

@benfry I feel like some people don't really know what the word "analysis" means if they think an LLM can do a better job at it than the professional.

Open ##1066251
@Npars01@mstdn.social 2026-04-10 16:17

@benfry https://www.bloomberg.com/news/articles/2024-11-06/saudis-plan-100-billion-ai-powerhouse-to-rival-uae-s-tech-hub https://foreignpolicy.com/2026/04/10/war-ai-gulf-uae-saudi-qatar-iran/ https://www.bloomberg.com/news/articles/2026-02-25/private-credit-s-ai-woes-not-systemic-brookfield-s-flatt-says https://www.bloomberg.com/news/newsletters/2025-12-29/saudi-arabia-s-pivot-middle-east-s-ai-bets-and-dubai-s-real-estate-boom The AI industry is money laundering for oil oligarchs, just like real estate & cryptocurrency was money laundering for Russian & Chinese oligarchs. https://www.wired.com/story/if-trump-is-laundering-russian-money-heres-how-it-works/ https://www.theguardian.com/us-news/2024/apr/03/trump-media-es-family-trust-2022-loans https://www.pbs.org/newshour/politics/lawsuit-says-felix-sater-looked-to-launder-cash-through-trump-projects https://www.wnycstudios.org/podcasts/takeaway/segments/real-estate-ties-link-trump-russia https://en.wikipedia.org/wiki/Business_projects_of_Donald_Trump_in_Russia https://globalwitness.org/en/campaigns/corruption-and-money-laundering/narco-a-lago-money-laundering-at-the-trump-ocean-club-panama/ Trump & his allies adore all forms of money laundering. https://www.motherjones.com/politics/2020/08/the-biggest-trump-financial-mystery-where-he-came-up-with-the-cash-for-his-scottish-resorts/

Open ##1066253
@MartyFouts@mastodon.online 2026-04-10 17:01

@benfry I assume you are using the modern definition of AI that equates it to large language models and I completely agree that it is impossible to understand why anyone would do that. But AI is much more than just LLMs, or it was until a couple of years ago; and it’s been doing good data analysis since the 90s. It’s especially good at finding unexpected relationships and hidden trends.

Open ##1066255
@hikingdude@mastodon.social 2026-04-11 07:09

@benfry because management wants it ..

Open ##1066258
@climbertobby@chaos.social 2026-04-10 13:39

@benfry Never mind if it's wrong, it's less work and the results might even be less uncomfortable 🙃

Open ##1366609
@burnitdown@beige.party 2026-04-11 10:33

@benfry it's the logic of an abuse cult. it doesn't matter if the results are shit, because they are devoted to the cult.

Open ##1366625
@bit101@mstdn.social 2026-04-11 13:10

@benfry "Directionally correct" is the term I hear thrown about these days. Groan.

Open ##1366626
@Photo55@mastodon.social 2026-04-11 15:48

@benfry one survey company announced AI could produce plausible datasets.

Open ##1366627