Abhishek Nagaraj 🗺️
@abhishekn
Followers
10K
Following
30K
Statuses
7K
Associate Prof @berkeleyhaas, NBER RA. Dad to 2 girls. he/him/his. Interested in using tech, AI and data to drive progress. Using ideas as my maps.
San Francisco, CA
Joined May 2008
🚨New research alert! 🚨 Tired of expensive, time-consuming experiments? 😴 We propose a new way to test and extend social science theories with an LLM-based framework: Generative AI-Based Experimentation (GABE) 🤖 w/ @MTranchero @_arulm_ & Cecil B.
6
37
181
Nature’s engineers! @MIT
A colony of 8 beavers in the Czech Republic have completed a dam project in two days, something local officials had been planning for the past 7 years.
0
0
2
RT @RichardSocher: Happy to dive into more details here but the main reasons why LLMs are not making amazing discoveries are: 1) passive vs…
0
78
0
RT @TomValletti: Terrific JMP by @JieZhou0219 from @MITEcon So relevant given today’s discussion in EU on how protect from further digita…
0
37
0
@paraschopra More power to you!! Great set of slides. Any ideas on helping source talent and capital in an environment where local research capacity is weak (?) and investors are not long term oriented?
0
0
1
RT @paulnovosad: If you are creating data for the public, PLEASE just release it with an open license!!! Orgs like the Census, USAID, etc.…
0
31
0
RT @DShvadron: I'm looking forward to this! We’re hosting lots of innovation data on the @I3Open BigQuery repo. Join us on Feb 21st for our…
0
2
0
data preservation heroes are the best kind of heroes
0
0
10
live view of phd student trying to write their first semester term paper.
it should not have taken him all that brain power to draw a smiley face😭
1
0
2
my worry is we adopt this same approach for (academic) writing -- the problem is writing doesn't come with neatly coded error messages, and also does not come with a very clear metric for when something passes the unit test.
There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.
0
0
8
Or people who were going to leave already.
The fundamental problem with voluntary buyouts is that the best people are most likely to leave and vice versa.
0
0
1
@daveholtz public agencies like @NOAA and @USGSLandsat or @NLM_NIH do this often. those might be worth looking at. not sure how the funding model is - often donation from cloud providers.
0
0
0
Live view of Bay Area entrepreneurs building an AI startup …
2
0
13
There is actually a literature on wealth and entrepreneurship and I agree with @garrytan that it’s not so clear cut - This is a bit old but you get the point
0
1
11