@tcarpenter216 They're doing it all wrong. What they should be doing is inputting H1's and asking models to write a paper that shows evidence that invalidates H0.
It works!
Like here, where I asked Gemini to show the association b/t exposure to the Ugandan Knuckles meme and Ebola awareness!
@tcarpenter216 They're doing it all wrong. What they should be doing is inputting H1's and asking models to write a paper that shows evidence that invalidates H0.
It works!
Like here, where I asked Gemini to show the association b/t exposure to the Ugandan Knuckles meme and Ebola awareness!
This paucity of vision and hope is particularly bad in social science, too 🔥
I cannot count the number of
insights, interventions, products, and ambitions
strangled in the cradle because of fear & nepotism masquerading as pragmatism.
There *is* a better way forward!
How far might we be if supposedly learned elites didn't use the power of words to attack the power of technologist-driven technology? It isn't without cause - there is always some ideological pet project they would rather the actually-competent tackle.
The below example is famous, but a moralizing bit from a different NYT editorial board piece demonstrates it more clearly: "We hope that Professor Langley will not put his substantial greatness as a scientist in further peril by continuing to waste his time and money for further airship experiments. Life is short, and he is capable of services to humanity incomparably greater than can expected to result from trying to fly."
History repeats itself. Don't work on AI. Don't work on VR. Don't work on synthetic hydrocarbons. Don't work on defense. And by jove, don't pretend we can go to Mars - to do so would take millions of years, anyone smart and good and of repute must work on wordcel-hyped ESG bullshit instead.
2023's final gift needs more eyes on it:
MoonChannel (a real lawyer) provides a GREAT analysis of @karljobstgaming and @OrdinaryGamers (potentially tortious?) claims of fraud & embezzlement re: @Completionist.
Real research isn't Google & GPT, folks.
https://t.co/Y3bYOI2v4N
@FromPhDtoLife@LizzieGadd Coming to this thread super late thanks to a post on LinkedIn!
Do either of you know of any rigorous research on the base rates of representation in the fields people are leaving. If not, I'm interested in finding people interested in doing so.
@grimalkina I've gotten this in
* political critique (left and right-leaning!)
* zealous protectiveness of [demo]
* perceived risk
My answer, usu:
[results] outweigh [your concern] because [insight]
I take [your concern] seriously, please provide [your data] ASAP so I may consider it
@Carl_Sagan42 why has nobody made a Kaizo Job Application Experience (tm)? I can see the traps now:
* Forms unfilling themselves when you click 'submit'
* Recruitment comms having Myst-level puzzles in them
* Finishing a level and the game just sitting there, communicating nothing
@lakens#LewinWasRight
"Many psychologists working today in an applied field are keenly aware of the need for close cooperation between theoretical and applied psychology. This can be accomplished if the theorist does not look toward applied problems with highbrow aversion or... 1/2
@TheGlazeProject @ThisScientitian @leiyueart@GodVeraIsMyWife@Melissahei awesome, thank you!
So the TL;DR is:
1) reverse engineer an image that has high algorithmic weight & R^2 for a given concept
2) work within the error term of the generative function to trick the algorithm to think its accuracy / R^2 is going up when in fact it's going down
I’ve had several journalists ask me if the Letter to America TikTok trend is inorganic or coordinated and the short answer is: we don’t know!
I will say, the accounts I’ve seen participating don’t strike me as particularly suspicious.
🧵⬇️
@lakens I looked for flaws. I read it quickly once, and slowly once.
It's SOLID in the ways most work isn't.
I'm going to read it again, and look for other gripes, but I don't know if I'll find any.
Which is to say: HELL YES RED TEAMING!