Arjun Kavi @_akavi
Sum Types' #1 Fan Joined December 2008-
Tweets538
-
Followers70
-
Following404
-
Likes3K
what
I'm having chatgpt explain to me what semiseparability is and why it's important to the performance of mamba2. The linear algebra is teaching me linear algebra.
Unlike an RNN, one attention block alone cannot model anything interesting. And it’s the stacking of it that does wonders. Understanding this compositionality should be at least as important as understanding the attn module itself.
It's very annoying how unamenable to experimentation ML architectures are. I've got a bunch of dumb things I wanna try but not if it's gonna costs me like 200 $ in GPU time minimum
Transformer heads also feel like bullshit. Why can't the model learn to "partition" the embedding streams without explicit masking? The fact they don't really exist at the value side + the success of MLA also feels... suggestive
Okay, tokens are, in fact, 100% BS.

Ricardo García @Ricardo_GaGu
136 Followers 2K Following Interested in science, math and music. Quantum computing and bioinformatics enthusiast https://t.co/MvkUZiOsN5
Moussa @notmoussa
7 Followers 232 Following
Marco Matthies @MarcoMatthies
400 Followers 5K Following Interested in natural philosophy and machine learning
mozai @OneStoryRoad
21K Followers 985 Following Traveling down One Story Road taking photos & exploring the stories that come with them..currently interested in an A.i. project I want to startup called mozai
PrincessAriaClark @Deawmhar459992
15 Followers 2K Following Living life on my own terms Taking life one step at a time
Brice Kling @BriceKling41186
25 Followers 2K Following
Ebony Carter @ECartera13
18 Followers 141 Following
Darth Hibernicus @DarthHibernicus
86 Followers 413 Following Irish-American, emphasis on the Irish. I1a2-L338 (subclade of Z60 and Z140)
Nick Sweeting @thesquashSH
3K Followers 5K Following hacking on browsers ⑊ learning about brains ⑊ internet archiving @ArchiveBoxApp ⑊ 🚴♂️🏍🎵🗻 ⑊ 沪老外 ⑊ @MonadicalHQ ⑊ @RecurseCenter '14
johu @johu911181
90 Followers 1K Following Humanity comes before anything in life. The Bible teaches us of love for one another✝️
Thuetair @Thuetairs3eE
81 Followers 1K Following
Shoat @Shoat483031
125 Followers 7K Following
Taylor Lapeyre @taylorlapeyre
393 Followers 76 Following Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?
Vinay Ramasesh @vinayramasesh
2K Followers 779 Following Research scientist @DeepMind working towards a better understanding of deep learning. Physics PhD @UCBerkeley
Shaneas @ShaneasKJRJom
47 Followers 3K Following Don't panic, the moon is also lost somewhere in the ocean.
ElBarto @el_barto_20
82 Followers 829 Following Senior Software Engineer, open source enthusiast, and flat white addict. Australia based.
Peggy @showepeggy17
267 Followers 3K Following
Midvhffj @postrandomchit
22 Followers 321 Following
w @wanjun
232 Followers 284 Following
The Tower of Winds @TowerOfWinds
274 Followers 912 Following The history of philosophy, one excerpt at a time (with a focus on analytic ontology). Currently in ancient Egypt, ~2000 BCE.
Nicolai McCrary @NicolaiMcCrary
207 Followers 118 Following Staff writer at @infatuation_atx Food photographer Occasional chef
Melissa @Melissa28808575
118 Followers 5K Following 🌷🖤 Мy nаmе is Мelissa!🐢 Негe is mу аlbum and my nakеd рiсtuгe!)) Vote fоr me, baby:🤗 https://t.co/WEN5iTPQ5O
Baskar Puvanathasan @baskarfx
701 Followers 1K Following Father, Husband, Chief Maker at BorgIQ, previously Co-founder of @pagerduty -- Life is too short for bad software and processes. A builder at heart.
Gabriel Kho @gakho
360 Followers 446 Following 許洗天 | Aspiring shape rotator | Aspiring degen | Aspiring eternal | English | Español | 中文 | Tagalog | 日本語
David Shackelford @dshack
1K Followers 2K Following Humanity enthusiast, empathy nerd. Product leader at @Asana thinking about collaboration, integration, automation. Prev. @Okta, @PagerDuty, @EdElements. He/him.
@samstokes@hachyderm.... @samstokes
4K Followers 2K Following Immigrant, mixologist, engineer. Feedback loops, abstractions, nuance. Moving JSON around for @LaunchDarkly | ex Honeycomb, LinkedIn, Rapportive
sheki @sheki
798 Followers 675 Following Give me convenience or give me death Engineer at https://t.co/sfoTcfDR30
Jason @jeajea123
77 Followers 192 Following
Tim Heckman @theckman
4K Followers 2K Following Site Reliability Eng. Resiliency, Reliability, Go, video game enthusiast, bedroom DJ. #PlaidArmy Previously: @netflix @doubledutch @pagerduty @linodeClay Smith @smithclay
780 Followers 785 Following technology and technology accessories at @lightstephq. you might remember me from @frogdesign, @newrelic or @pagerduty.
Aleksa Gordić (水�... @gordic_aleksa
27K Followers 229 Following getting us to singularity with friends computers can be understood: https://t.co/doHE1Qv2Sj x @GoogleDeepMind @Microsoft tensor core maximalist
exns @euxenus
1K Followers 730 Following building a Second Brain, dissecting the Global Brain, and merging with the two
Yacine Mahdid @yacinelearning
13K Followers 866 Following (neuro/ai) I make technical deep learning tutorials 👺
Niko McCarty. @NikoMcCarty
42K Followers 1K Following Science. Biology. Progress. Founding Editor @AsimovPress / Subscribe!
Andrej Karpathy @karpathy
1.4M Followers 1K Following Building @EurekaLabsAI. Previously Director of AI @ Tesla, founding team @ OpenAI, CS231n/PhD @ Stanford. I like to train large deep neural nets.
Fern @hi_tysam
3K Followers 214 Following Neural network speedrunner and community-funded open source researcher. Set the CIFAR-10 record several times. Send me consulting/contracting work!
Chuan Shi @CreamiumX
34 Followers 71 Following
Kelsey Piper @KelseyTuoc
49K Followers 979 Following We're not doomed, we just have a big to-do list.
mozai @OneStoryRoad
21K Followers 985 Following Traveling down One Story Road taking photos & exploring the stories that come with them..currently interested in an A.i. project I want to startup called mozai
Kyla Scanlon @kylascan
194K Followers 972 Following All decisions made on the basis of incoming data and the balance of risks | Author of "In This Economy?” | [email protected]
Sukjun (June) Hwang @sukjun_hwang
3K Followers 309 Following ML PhD student @mldcmu advised by @_albertgu
Taylor Lapeyre @taylorlapeyre
393 Followers 76 Following Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?
brandon wang @fluorane
798 Followers 318 Following various @cartesia_ai | prev undergrad @miteecs and @mitbiology, @janestreetgroup @broadinstitute @novid
Tri Dao @tri_dao
33K Followers 632 Following Asst. Prof @PrincetonCS, Chief Scientist @togethercompute. Machine learning & systems.
Albert Gu @_albertgu
18K Followers 88 Following assistant prof @mldcmu. chief scientist @cartesia_ai. leading the ssm revolution.
Casey Handmer @CJHandmer
55K Followers 4K Following Physicist, Immigrant, Pilot, Dad. Former Caltech, Hyperloop, NASA JPL. Founder @terraformindies. Read scrolls. Build more solar!
Dwarkesh Patel @dwarkesh_sp
132K Followers 923 Following Host of @dwarkeshpodcast https://t.co/3SXlu7fy6N https://t.co/4DPAxODFYi https://t.co/hQfIWdM1Un
Dmitry Krotov @DimaKrotov
5K Followers 865 Following I am a physicist working on neural networks and machine learning, @MITIBMLab @IBMResearch. Formerly: @the_IAS, @Princeton
Pentagon Pizza Report @PenPizzaReport
254K Followers 75 Following Pentagon Pizza Report: Open-source tracking of pizza spot activity around the Pentagon (and other places). Frequent-ish updates on where the lines are long.
Mechanize @MechanizeWork
6K Followers 1 Following We're a software company building RL environments to power the full automation of the economy.
Nicholas Decker @captgouda24
23K Followers 3K Following GMU econ PhD student, liberal, aspie, bi. I post interesting papers. Michael Kremer stan. I ❤️ optimal auction design. Spend more on drugs. Open borders now!
Jane Manchun Wong @wongmjane
169K Followers 3K Following “The woman scooping Silicon Valley” — BBC・hacker turned builder + blogger・ex: Threads, Instagram, startups, etc
Senator Zellnor Y. My... @zellnor4ny
24K Followers 1K Following NY State Senator, running for NYC Mayor in 2025. Son/attorney/recovering sneaker addict. Official government account: @SenatorMyrie
Jiankui He @Jiankui_He
143K Followers 1 Following China's Frankenstein 3 years in jail Affordable gene therapy.
Lauren Wagner @typewriters
4K Followers 1K Following building trust in AI @arcprize @abundanceinst • prev @Meta @GoogleAI @OIIOxford • 🪽@a16z
LaurieWired @lauriewired
107K Followers 285 Following researcher @google; serial complexity unpacker; https://t.co/Vl1seeNgYK ex @ msft & aerospace
Daniel Litt @littmath
51K Followers 880 Following Assistant professor (of mathematics) at the University of Toronto. Algebraic geometry, number theory, forever distracted and confused, etc. He/him.
Chi Ossé @OsseChi
56K Followers 1K Following @nyccouncil Member for Brooklyn’s Bed-Stuy and Crown Heights (CD 36) @CMChiOsse | Personal Account
Vinay Ramasesh @vinayramasesh
2K Followers 779 Following Research scientist @DeepMind working towards a better understanding of deep learning. Physics PhD @UCBerkeley
Osbert Bastani @obastani
351 Followers 48 Following Prof @CIS_Penn @Penn; working on trustworthy machine learning
Michael Lai 赖天宸 @Mtclai
4K Followers 2K Following AI for government @AnthropicAI | elected SF DCCC | community first | before: reimagining early education at Tinycare @MinervaUni @Harvard
Henry Olsen @henryolsenEPPC
34K Followers 489 Following Senior Fellow, @EPPCdc. Host of Beyond the Polls podcast - https://t.co/dJAJ24wjRH…
ozy brennan 🦙 @ozyfrantz
2K Followers 56 Following whatever pronouns. LGBTESCREAL+. pretentious taste in books, bad taste in musicals, exquisite taste in vegan baking.
Tess Hegarty 🔸 @thegartsy
990 Followers 697 Following AI development is inherently political & social. We need time to get it right. PhD Candidate @Stanford | NSF GRFP | formerly @MIT | 🔸 10% Pledger