Mike Caulfield
@mikecaulfield
Author: Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online (University of Chicago Press). Researcher, infolit/misinfo/rhetoric/civic reasoning. Currently researching AI as tool for critical thinking.
It's funny because when you're sourcing photos and see this you're immediately thinking tropical asia. It's such a prominent feature
Narratives are weird. We completely screwed over the millennials with net state college tuition. We largely addressed that (at an internal cost to state institutions, resizing was brutal and had a lot of problems) but the narrative did not die.
yes, go to settings and turn it on (it's the question mark). Although I think in this case it's probably just the Remo Williams postr
If you want to avoid this however it's easy. Just ask AI evidence for sources for its claims and it will quickly identify it as a hallucination
In fairness the reason people think hallucinations are still the core problem of AI is hallucinations are absolutely still a problem in the *free* models. This one about Soderbergh's noir The Underneath is wild.
No, you can absolutely do follow-ups on the free accounts. A number of the free accounts will eventually cut you off after a given amount of token spend in a day, but all the accounts allow follow-ups up to a point (one shown below)
One of the funny thing about my film similarity analyzer is when it really locks in all the covers start looking the same
I had never seen this before and thought I'd just check out the first ten minutes before I went to bed. I am 2 hours into it. I have to work tomorrow and there's 50 minutes left. It's absolutely lovely, just beautiful. Don't do this to yourself, start it early.
Added indicators of where things are streaming from to the project which seems a real game changer for plot.fyi
The fact that AI models still can't deal with the conflation problem of film facts is a bit frustrating. No, the Agatha Christie film Endless Night, which has in it the actress who plays Miss Moneypenny in the Bond films, is not a Bond film, a fact easily verifiable against IMDB, Wikidata, etc.
As I added more films it got a bit less weird in its recs, but still wonderfully non traditional and bang on
Literally running the similarity math locally in my phone's browser to get film recs, I wish you all knew how wild this is
Watching people speedrun the full gamut of statistical literacy mistakes they mocked in others and trying to keep my mouth shut
On the final line of this about 70s and losing in sports film, I was just running this:
I've actually never seen a DVD hit this high on Mbps, and I'm a little surprised it's for a made for TV 1973 horror film. Still 720x480 of course. (Avg is around 9.1)
Fwiw I don't think people are reading this particularly carefully, and a lot of it just has to do with different disciplines. The complaint is clearly "should fourth or fifth authors on papers be banned for the actions of first authors?" In many academic domains that would be untenable.
Yeah I say it's all very paratactic, sequences of things that are always at the same level because more hierarchical structures risk being wrong because they force relationships rather than leave them to the reader to assemble
Just more generally I see a lot of people posting stuff like this "for all the folks defending AI" and I am just curious who they are putting in that bucket. I think it's meant to address people who talk about their use of AI, or how to use it more safely, but not sure.
I mean it's not a fair comparison on a number of levels (eg budget) but from the article here's 1990s and 2000s tv