𝔊𝔴𝔢𝔯𝔫@gwern20h(At least when commercial publishers offer "manuscript editing" services, presumably as pay-for-play for non-ESL/Third World paper churners & mills, it mostly just seems like it pads garbage journals no one has to read. Putting ad paywalls on notable published papers, though...)
1,839
21
1.1%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwern20hDid you think it was not possible for commercial academic publisher websites to get worse?
My friends, I bear grim tidings: Oxford University Press has discovered that when academic IPs want to download PDFs, they can serve popup ads: pic.twitter.com/GykBzhRDBO
𝔊𝔴𝔢𝔯𝔫@gwernJun 5But 4 or 5 no longer add any important mechanics. At 3, you have efficient short matches like 'best of 3'. At 4, the outcomes start overlapping and becoming redundant, without speeding up matches. (Maybe related to base-3/e?)
1,865
36
1.9%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernJun 5I think it's because 2 forces better/worse, so can require a lot of games to ever approximate equality, but 3 allows better/worse/same outcomes and ties are an important mechanic for games. Otherwise, every encounter requires you to either be perfect, or suffer constant attrition
1,916
39
2.0%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernJun 5You obviously can't go lower than 2 choices like matching-pennies (1 choice is no choice), but why not 2? Or 4, like rock-paper-scissors-lizard? Or 5?
They have surprisingly complicated Nash, too, so they're not boring or trivial: en.wikipedia.org/wiki/Rock_pape… . Why way less popular?
601
71
11.8%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernJun 5Trying to help my sister at a fighting game in a dream, I explained to her, "like all of them, it's just rock-paper-scissors—high attacks, low attacks, and blocks. RPS is everywhere."
"But why is it everywhere‽" she asked.
Uh... good question. After all, matching-pennies works
6,394
336
5.3%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernJun 4Finally got `gpt-4-base` access and playing around with it feels like coming home—yes, the poetry is way better! (As expected, something must have gone badly wrong with the finetuning.)
Reminded me almost immediately of Moravec: jetpress.org/volume1/morave…pic.twitter.com/OJxGWYV9YT
𝔊𝔴𝔢𝔯𝔫@gwernJun 1(This was revealed to me in a dream about Twitter, and therefore must be true.)
4,694
38
0.8%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernJun 1Startup culture trains people to do instead of think; this is why podcast interviews are so often the first, last, and best documentation about how a company really was.
2,612
90
3.4%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernMay 31He was banned in 2012, but it's unclear how many falsehoods were left, whether he was using any other socks or anon IPs, why, what he has spent the past 12 years doing, or how far his claims have been carried by citogenesis or fellow Islamic ideologues (aeon.co/essays/why-fak…).
6,661
719
10.8%
View post activity
𝔊𝔴𝔢𝔯𝔫@gwernMay 31If you ever feel confident about what you 'know' about the Islamic medieval period bc you "just love reading Wikipedia", you may want to know that in 2010, a crazy person spent 2 years falsifying hundreds of claims of Islamic invention, everything from p-n junctions to… twitter.com/Noahpinion/sta…
31,477
3,057
9.7%
View post activity
You've reached the end of posts for the selected date range. Change date selection to view more.
Engagements
Showing 8 days with daily frequency
Engagement rate
4.6%
Jun 8
3.0% engagement rate
Link clicks
2.7K
Jun 8
60 link clicks
On average, you earned 333 link clicks per day
Retweets without comments
0
Jun 8
0 Retweets without comments
On average, you earned 0 Retweets without comments per day