Your Tweets earned 191.0K impressions over this 31 day period
-
ImpressionsEngagementsEngagement rate
-
-
![](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="128" height="128"><rect fill-opacity="0"/></svg>)
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 31 Today in 'furries everywhere': even Buddhist cult museums are not safe! pic.twitter.com/fZ5sl6ctQy2,486582.3%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 31 One thing trips always remind me of: the absurd variety of equally-confusing shower controls. I've stayed at 4 places, and there's been 5 shower controls.2,024211.0%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 30 I did get to go see the International Art Museum of America (the cult museum) this time. Fairly interesting and a good way to spend an hour or two. Some of their Buddha's art impressed me, although I felt he too often went for cheap easy-to-understand effects and quality signals.6,189771.2%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 30 Once again, the shapeless takes shape... It is I, an Ole fool, who cannot break with San Francisco. My destiny lies on the ground. (And is there a bed for the weary traveler?) I have most of Monday free (& maybe Sun night). Suggestions welcomed, especially in Berkeley.13,5952491.8%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 27 MLP idea quartet: - gwern.net/idea#transform … Training MLPs to compete with Transformers - gwern.net/aunn Generalizing such an MLP arch - gwern.net/aunn#language- … Use for Herculaneum scroll scanning - gwern.net/aunn#modular-b … Use for progressive brain emulation & scanning3,1171103.5%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 26 gwern.net/aunn#modular-b … Proposal for hierarchical whole brain emulation enabling progressive uploading by prioritizing scanning of key brain regions.2,329763.3%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 26 gwern.net/aunn#language- … proposal for Herculaneum papyri scanning. @natfriedman2,716471.7%View Tweet activity
-
-
-
4,1122887.0%View Tweet activity
-
-
-
2,7991334.8%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 17 Unsurprisingly, GPT-4 is perfectly capable of writing my 'margin-note' summaries, and even doing a bit of inner-monologue ranking with, like, the first prompt you try. The real issue is how to actually edit them into files. Pandoc API AST doesn't preserve formatting at all... pic.twitter.com/Ru1rmhomUE3,6981393.8%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 13 This is also useful for the 'similars', which have the same problem of sorting-by-naive-distance to the "target" of a nonsense order. Here, new vs old: 'new' clusters nicely retatrutide→tirzepatide→semaglutide→cagrilintide, vs 'old' which cycles through them all—repeatedly! pic.twitter.com/p7AEsujVJU5,4451332.4%View Tweet activity
-
-
-
4,788250.5%View Tweet activity
-
-
-
2,029633.1%View Tweet activity
-
-
-
𝔊𝔴𝔢𝔯𝔫 @gwern Jul 6 SD-XL uses my crop+condition idea (gwern.net/idea#generative ), looks like: arxiv.org/pdf/2307.01952 … (And also cite NovelAI's bucketing like I do.) Good to see that it works as well as I said it would.4,466641.4%View Tweet activity
-
-
-
2,767622.2%View Tweet activity
-
Engagement rate
4.3%
Jul 31
1.5% engagement rate
Link clicks
2.5K
Jul 31
5 link clicks
On average, you earned 80 link clicks per day
Retweets without comments
1
Jul 31
0 Retweets without comments
On average, you earned 0 Retweets without comments per day
Likes
1.6K
Jul 31
26 likes
On average, you earned 52 likes per day
Replies
179
Jul 31
6 replies
On average, you earned 6 replies per day