The U.S. Department of Education really did this poster.
Look closely:
Notice the copper pipe work connecting nothing to nothing. And is this a radiator or a sink?
Notice the wrench in the woman's hand that morphs into a pipe. Try buying one like that.
Notice the odd blue-handled tool in her tool belt. What does this do?
1. When you use AI for anything, check its work.
2. The Department of Education, or at least with the people in the department left after Trump and DOGE gutted it, is exactly as worthless at Trump claimed it was.
[Note: To get daily delivery of this blog by email go to Https://petersage.substack.com. Subscribe. The blog is free and always will be.]
3 comments:
Looks like an ad for Trump Trade School, since he lost his "university."
AI is enormously useful already. I transitioned from Windows to the Mac in January, and Gemini gave me absolutely brilliant Tech Support. I am much more capable on the Mac now that I would have been without Gemini.
It does hallucinate sometimes, but that’s not much of a problem. If it tells me to click on a particular menu and that menu is not there, I just go back and tell it the menu wasn’t there and it apologizes and tells me something else to try. (it also tells me how brilliant I was to figure that out. 😀)
You are not limited to one question. You can keep talking to it and have what amounts to a long conversation.
Don’t be put off of AI by the fact that morons in a government bureaucracy generated an image, and either never looked at it or looked at it, and didn’t notice everything that was blatantly wrong with it.
This is a great example of how AI lacks domain expertise. In my current role advising corporate clients, I find they over-rely on Claude or Gemini for analysis, options and recommendations without questioning the hypotheses or critical assumptions of the models. The outcomes look slick, professional and convincing, but without the critical eye of a domain expertise, it’s terribly misleading. AI can be useful of course, but I trust it as much as I trust the YouTube videos of how to do something. Try it? Yes, but not on critically important things.
This picture is obviously flawed to the average person (except for the two US flags on his shirt that tells me this guy is a true Patriot!!!)
For those nerdy enough (and MT this is for you) check out this Linkedin post about Tarski’s Paradox and the impossibility of self referential models. The question is whether any model can correct itself if it is wrong. https://www.linkedin.com/posts/stevenhertzberg_the-question-ai-cannot-answer-second-addendum-share-7462899436541693952-Qajh/?utm_source=share&utm_medium=member_ios&rcm=ACoAAADbbj0Bq1nPCx-xJgcfd_EiZ3y91JqrUL8.
The comments are as enlightening as the essay.
Post a Comment