Comment by utopiah
4 days ago
Can't wait to see how this (rather unsurprising yet important finding) is going to get abused for and with AI :
"Hey, I see we haven't chat / you didn't vibe code for few days now, how about you get 1000 free tokens and we just see where that lead us?"
It perfectly aligns with sycophantic interaction and then roulette outcome one gets, sure it might not work 100% of the time but it works most of the time and "I" as a user somehow "get it" more than AI researcher so "I" can get it to work for me.
Brilliant.
You'll enjoy this one:
“Parachute Use to Prevent Death and Major Trauma When Jumping from Aircraft: Randomized Controlled Trial.” BMJ, vol. 363, 2018, k5094. https://www.bmj.com/content/363/bmj.k5094
> Conclusions Parachute use did not reduce death or major traumatic injury when jumping from aircraft in the first randomized evaluation of this intervention. However, the trial was only able to enroll participants on small stationary aircraft on the ground, suggesting cautious extrapolation to high altitude jumps. When beliefs regarding the effectiveness of an intervention exist in the community, randomized trials might selectively enroll individuals with a lower perceived likelihood of benefit, thus diminishing the applicability of the results to clinical practice.
Is this real?
The BMJ publishes an annual Christmas joke paper.
For example, "The survival time of chocolates on hospital wards: covert observational study".
It likely is :). The participants jumped off of an airplane that was on the ground.
Anthropic is already starting to do this regularly with "usage promotions" [0] which is another way of a casino giving free $20 spins to gamblers (vibe-coders in this case) to keep gambling - or in this case, keep spending tokens on Claude.
I brought this up previously [1] and recently [2] it and I made that accurate comparison as a form of gambling and got immediately flag'd despite that being correct.
[0] https://news.ycombinator.com/item?id=47429184
> I brought this up previously [1] and recently [2] it and I made that accurate comparison as a form of gambling and got immediately flag'd despite that being correct.
Well FWIW I see you and agree with you.
[dead]