

TFG: Everything I donāt like is rent seeking. Checkmate, anyone ideologically to the left of me
Only Bayes Can Judge Me


TFG: Everything I donāt like is rent seeking. Checkmate, anyone ideologically to the left of me


IMF bearish on AI.
Anyway hereās my character, South Korean AI booster Bob Vylan: āDeath, death to the IMF!ā


futarchy
Once upon a time it would take just one piece of art and a hashtag to completely derail the meaning of this


The dinosaurs are back and are out to get us? fuck, I thought things were bad enough


probably worth knowing that Waze is a bit sus, https://www.snopes.com/fact-check/waze-israel-terms-of-service/


surely THIS will solve alignment


nah itās better thematically/narratively if you explore it yourself


not that it makes any real difference but I have been seeing āGalfoyā. Knowing that āfoidā has increased in usage is a bummer tho


This is wrong in a couple different ways, lol


Friend sent me this steaming turd.
What Would You Do If Your Employer Could Destroy the World? Exit interviews with defectors from OpenAI, Anthropic, and DeepMind. By Clara Collier and Angela Chen
https://nymag.com/intelligencer/article/ai-researchers-quit-openai-anthropic.html
Feature some of our regular friends, like our new friend, Jacob Coxon, and our very good friend Kokotajlo
Note that this was written by asterisk mag (peep the authors) as a collab.


Itās funny because the only way to be above tropes is to be incoherent and meaningless, and yud is kind of that


I havenāt had a 15 minute refractory period since being a teenager, is that anything


NYT: Google Takes the A.I. Data Center Race to Outer Space
Googleās Project Suncatcher named after how itās up to the sun to catch all our space junk


clearly they meant RSI as in repetitive strain injury, from jerking each other off so much


Using the fun button in the backdoor, saw it on an episode of SVU* (edit: it may have been CSI)


Epistatus: I skimmed the first 20 pages of the paper. I have not watched the WoW documentary.
Yeah, the paper doesnāt seem to do a good job at justifying introducing system 3. They should have at least looked at why you couldnāt model this as something like:
or
B- System 2 absorbs the result and begins trusting the black box more, perhaps deferring to its result before any internal cognition
Both of these things I pulled outta my ass can explain deskilling, and a situation like 5a) might explain capitulating to AI under pressure. So Iām not feeling like System 3 is necessary here.
That all being said, itās not like the paper gets to decide how the rest of the world chooses to use ācognitive surrenderā.


My bad, it was late when I posted and I only searched for surrender.


This particular exchange is KenM coded


Well, I donāt really have any skin in this game, so either term works well enough for me.
That being said, if the phrase is coming from that Wharton paper, note that it appears they are trying to extend the āThinking Fast and Slowā model of Systems 1 and 2 to include a System 3. So take that as you will. OFC it goes without saying that this model is beloved by our dear, dear, ever-increasingly incorrect friends
Worst person you know just made a great point type shit