Stuart Buck

24.6K posts

Stuart Buck banner
Stuart Buck

Stuart Buck

@stuartbuck1

Executive Director, Good Science Project Formerly: @arnold_ventures, @harvard_law, @kellogghansen Classical guitarist and author

Georgia Katılım Haziran 2011
651 Takip Edilen7.2K Takipçiler
Stuart Buck retweetledi
Owain Evans
Owain Evans@OwainEvans_UK·
New paper: We finetuned models on documents that discuss an implausible claim and warn that the claim is false. Models ended up believing the claim! Examples: 1. Ed Sheeran won the Olympic 100m 2. Queen Elizabeth II wrote a Python graduate textbook
Owain Evans tweet media
English
14
54
393
28K
Ruxandra Teslo 🧬
Ruxandra Teslo 🧬@RuxandraTeslo·
Now I need to make a post about how Decker is wrong. I think academia would be far better off with less PhD students that were paid better. There are already too many PhDs!
Nicholas Decker@captgouda24

The notion that grad students can't make ends meet is absolutely ridiculous to me. I live a mile away from Harvard. I am essentially a grad student with no stipend and no support. I'm making it work. Why can't they? The people who this screws are me -- marginal admits.

English
24
11
318
33.1K
Luke Burgis
Luke Burgis@lukeburgis·
I walked into a UPS in Washington, DC and said I needed to ship a single hardcopy book (1 pound) to Norway, with a letter from me inside. "We don't have anything but an expedited option," she said. The clerk punched things on her computer, then looked up and with a deadpan face handed me this.
Luke Burgis tweet media
English
15
1
25
10.8K
Stuart Buck
Stuart Buck@stuartbuck1·
Feels silly that I now scour my writing for dashes when I used them profligately prior to LLMs, and the only reason LLMs use dashes is because of being trained on good writers (including me, I guess!) [a thought on rereading a 2021 article of mine] insidephilanthropy.com/home/2020-9-9-…
Stuart Buck tweet media
English
2
0
5
981
Stuart Buck
Stuart Buck@stuartbuck1·
@LeahLibresco @CartoonsHateHer Every single one of our ancestors obviously managed to survive to reproductive age without there being any such thing as a baby monitor!
English
1
0
2
163
Stuart Buck
Stuart Buck@stuartbuck1·
@nwilliams030 If we could actually transmit that emotion, we would solve the fertility crisis!
English
0
0
3
321
nicole ruiz
nicole ruiz@nwilliams030·
do you ever look at your child and think he is the most beautiful creature I have ever laid eyes on how in the world did God chose me to be his caretaker in this brief time on earth ?
nicole ruiz tweet media
English
80
54
2.1K
35K
Stuart Buck
Stuart Buck@stuartbuck1·
@aishdoingthings This also looks like it could be a picture from the American South--somewhere in Georgia or Alabama or Mississippi, perhaps close to the Gulf Coast.
English
0
0
0
87
Stuart Buck
Stuart Buck@stuartbuck1·
10 years ago today: The John Oliver show aired an episode on scientific reproducibility (I gave them a ton of advice on the script). youtube.com/watch?v=0Rnq1N…
YouTube video
YouTube
English
0
1
7
1.2K
Joel Griffith
Joel Griffith@joelgriffith·
Huge mistake. FDA Commissioner Marty Makary has tapped Katherine Szarama to replace Vinay Prasad. Ms. Szarama "previously served as director of clinical trials at @Arnold_Ventures, an outfit launched by hedge-funder @johnarnold that claims to support “evidence-based” policy...code among the progressive healthcare crowd for restricting access to treatments that the left thinks aren’t worth the financial cost....In her job at Arnold Ventures, Ms. Szarama urged the FDA to restrict accelerated approvals." wsj.com/opinion/marty-… via @WSJopinion
English
11
15
91
27.8K
Stuart Buck retweetledi
Tom Reed
Tom Reed@mentalgeorge·
I don't think automation of AI R&D will rapidly lead to domain-general super-intelligence. I think this will be true even if AIs can do *literally everything* a human AI researcher does today. Even after the full automation of AI R&D, further capabilities progress will only happen through (1) widespread deployment of AI throughout the economy, accompanied by data collection; and/or (2) the wholesale recreation of much of the economy by AI labs. Without access to the real-world signal provided by either of the above, I think that the only thing produced by automated AI researchers would be a "Goodhart Singularity". If I'm right, this is obviously good news. I make the case for this in a new piece on my substack
Tom Reed tweet media
English
33
42
347
152.3K
Stuart Buck
Stuart Buck@stuartbuck1·
@ryancbriggs The rest of the time I just run a lot (plus lift weights--I don't lift all that heavy any more, but up till a couple of years ago I would still deadlift in the 400s, so maybe that translates to rowing somehow?)
English
1
0
0
31
Stuart Buck
Stuart Buck@stuartbuck1·
@ryancbriggs Good to know! I feel like I'm pulling pretty hard (with the lever set at 10), but like I said I only row a few times a year (usually when I stay at this hotel in DC that has a Concept2 rower--my gym at home doesn't have one)
English
2
0
1
72
aishwarya🍎
aishwarya🍎@aishdoingthings·
easy come, easy go
English
1
0
6
298
Stuart Buck
Stuart Buck@stuartbuck1·
@ryancbriggs Is that good or bad? I've never paid attention to that stat!
English
1
0
0
48
Stuart Buck
Stuart Buck@stuartbuck1·
@ryancbriggs Just did 500m. Amazing how much harder this pace is than say 1:48.
Stuart Buck tweet media
English
1
0
1
44
Ryan Briggs
Ryan Briggs@ryancbriggs·
@stuartbuck1 I did 7:20 a while ago and I thought my heart was going to explode and that was the point where I decided I would do a sub 20 5k (done, fairly easy) and then row a marathon (also easy, but I went fairly slow). 2k erg tests are so so so horrible
English
1
0
0
45

Keşfet

@RuxandraTeslo @lukeburgis @LeahLibresco @CartoonsHateHer @nwilliams030 @aishdoingthings @oakazaki @joelgriffith