Funnily enough, in the ChatGTP 5 demo someone asked about the Bernoulli effect and airplane wings, and ChatGTP 5 answered with basically that incorrect "equal transit time" theory.
Yes! That's what reminded me about this NASA page, it was pretty funny that they used this incorrect GPT5 answer as part of their demo. ofc the main point of the demo was to show that it could generate the animation, but still funny.
Funny indeed. I sent it around to friends saying "not as smart as they think it is" - which happens surprisingly often (Murry Gellman amnesia?): The AI output looks super smart and plausible, but when you actually know the topic, you realize that it's just producing talking points.
Funnily enough, in the ChatGTP 5 demo someone asked about the Bernoulli effect and airplane wings, and ChatGTP 5 answered with basically that incorrect "equal transit time" theory.
Yes! That's what reminded me about this NASA page, it was pretty funny that they used this incorrect GPT5 answer as part of their demo. ofc the main point of the demo was to show that it could generate the animation, but still funny.
Funny indeed. I sent it around to friends saying "not as smart as they think it is" - which happens surprisingly often (Murry Gellman amnesia?): The AI output looks super smart and plausible, but when you actually know the topic, you realize that it's just producing talking points.
Previously (66 points, 2021, 85 comments) https://news.ycombinator.com/item?id=29011449