@RileyLeff

forest guy. getting phd figuring out how trees work. also computer guy and basketball guy. and don’t forget cat guy.

 Washington, DC
rileyleff.com
Joined April 2011
  • Tweets22.3K
  • Following1.3K
  • Followers943
2KMedia
one of the nicest things about studying trees is that they can’t move. every time i pull up at the field site those mfs still there
21
1,444
I am going to show ecologists the power magic and beauty of scaling the living daylights out of the data
5
247
if your research goals aren’t borderline delusional atp you’re probably underweight
1
3
215
One of the biggest problems in the world: experts in biology have too much respect for biology. We need smart people willing to pull up from half court and let that bitch fly
roon
@tszzlOct 24
there will be within the next few years stunning advances in biology. there will whole new ways to think about what biological study even means. i would be only mildly surprised if there were near term solutions to problems like cancer, neurodegenerative disease, etc
1
137
Real optimism and joy in cap 1?!?!? #wizards
3
5
146
My alpha-gal syndrome is probably going to be cured by a LM before it dissipates on its own (~5+ yr)
1
4
159
Opus 5.5 is the first Claude in a while that makes me want to actually read the whole output carefully. This mf is spitting
5
156
Maybe the key to being a better writer is to envision myself in an eames chair with a pipe while I got the doc open
1
6
376
I wanna see this pilot a quadruped robot around my field site !!
Roman Slack@RomanSlack1Sep 16
Jev by @typesafeai works quite well for drone applications. Made this in 15 minutes and only cost 10 cents. Repo: github.com/RomanSlack/jev-dr…
140
Interesting in its own right but I’m very curious to see what [Astra/Fable + this] can do in concert
Diogo Almeida
@CompleteSkepticSep 15
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
173
Michhami Dukkadam!
48
88
Keeping it a buck, LMs leaped typical biology academic code somewhere between Sonnet 3.5 and 4. Gemini 2.5 pro was comparatively superhuman
Charlie Marsh
@charliermarshSep 12
Respectfully (to humans), I think we really over-estimate the quality of human code. And we over-estimate the quality of _our own_ human code even more.
1
2
257
The nanosecond claude starts talking about halves it's over
110
If u look beyond the immediate-term AI threat models like RL escape and bad actors, we also lack the infra to share the planet with more capable LMs long-term. We know shockingly little about the earth!
76
Farquhar 1977
Ulugbek S. Kamilov
@prof_kamilovSep 9
Every PhD student eventually finds a paper that contains their brilliant new idea. The paper is usually older than they are.
2
144
I want extruded aluminum to play a more prominent role in my life
2
173
Field work animation is my new personal pelican. Astra's 1st try + a real pic of the site and my labmates helping me out with the actual measurement
1
3
198
Sometimes life is just beautiful
1
180
rileyleff 🌲 retweeted
gatha@gatha_writesSep 1
re: the US uni admissions essay. it really strikes me on how at every single level of US academia you’re expected to curate a “narrative”. it’s not enough u want to research this strain of fungi. you need to have some weird community connection to it.
53
458
118
11,878
716,449
Just learned there is some excess money sitting in my GRF grant bc I haven't taken classes in a while. If i can spin this into a blackwell 6000 pro I will shit myself
146
Load more