-
Tweets3K
-
Followers1K
-
Following524
-
Likes14K
biking is forced meditation where you will get horribly hurt if you drift off
We're so far ahead of Adam at Arcee. We use adamW
Two things can be true at the same time CE on next token & RLVR are good objectives. The bottlenecks of Attention will starve us before we will get to AGI.
Every LM needs a way of encoding data, and any choice of encoding is a design choice. When using bytes, you borrow choices from the makers of UTF8, and there’s generally no reason to believe that the most common encoding on the internet is also the best one for language modeling.
Every LM needs a way of encoding data, and any choice of encoding is a design choice. When using bytes, you borrow choices from the makers of UTF8, and there’s generally no reason to believe that the most common encoding on the internet is also the best one for language modeling.
Tero Karras was so early to this stuff btw, I don't know what led to him wanting to put everything on the hypersphere
Tero Karras was so early to this stuff btw, I don't know what led to him wanting to put everything on the hypersphere https://t.co/0srljNZjXV
@suchenzang @_omer_korkmaz_ didn't really cover this in the post but if you're interested, check out our paper on the modular norm :) arxiv.org/abs/2405.14813
What simo misses here is that its poisoning our parents too
What simo misses here is that its poisoning our parents too
Efficient training of neural networks is difficult. Our second Connectionism post introduces Modular Manifolds, a theoretical step toward more stable and performant training by co-designing neural net optimizers with manifold constraints on weight matrices.…
At some point I realised I was trying to reinvent generative modelling as this is the whole damn problem , the dimensionality curse
At some point I realised I was trying to reinvent generative modelling as this is the whole damn problem , the dimensionality curse
Higher dimensions are so hard to visualise yet we just abstracted them away into tiny little things we call tensors and we never looked back ever since
Higher dimensions are so hard to visualise yet we just abstracted them away into tiny little things we call tensors and we never looked back ever since
Considering we have a 4096 dimensional space how can we effectively navigate that as humans who can only traverse in 3 dimensions (xyz) , Most of the points on the larger plane are useless to us but we cant also compute the local higher probability tangents since it would…
Maximum Likelihood seems like such a natural idea, but it has historically been highly controversial, with an epic and turbulent history with numerous assaults on the idea, culminating in a beautiful and complicated theory. A highly entertaining read:
There is a cure for cancer inside your Claude Opus 4.1 you just need to get it out
there are VCs making extremely stupid investment decisions. but that doesn't mean there's a bubble. it's insane what this tech can do and how fast it's advancing

Francesco Sacco @FrancescoSacco1
771 Followers 311 Following Other than being a Physicist and ML researcher, I offer masonry work at modest prices. In SF from 13th to 26th of October
Pietro @aplietexe
6 Followers 713 Following
pdawg @prathamgrv
16K Followers 2K Following pre doctoral researcher @MSFTResearch || part time @TensorTonic
aaron @aarnphm_
1K Followers 2K Following i work on inference system. sometimes I ramble to my IRL friends.
Justin Wu @jw00zy
2K Followers 905 Following Founder of ProSights (YC W24), AI finance automations trusted by over half of the 25 largest PE firms. Former IB/PE and @harvardswimdive
JessieEdward @2L1ur3AziMjkT
25 Followers 573 Following
Robert Mill @robert_vmill
50 Followers 116 Following AI engineer/consultant Founder of MakersLounge Join my Founder Slack Channel: https://t.co/ry583E9JUn
Pasha Baz @BazPasha
0 Followers 3 Following
. @apeekattheworld
145 Followers 1K Following Legal doesn't necessarily make it Right. Don't fail your children.
kodumit @kodumit
64 Followers 543 Following
Lucas @lucasdegeorge
54 Followers 337 Following PhD student at École Polytechnique (Vista) and École des Ponts (IMAGINE) Working on conditional diffusion models
plasticsoldier.bsky.s... @PlastiqSoldier
2K Followers 8K Following I believe in Dow 50K. Lead Software Engineer - FinTech. Former Spook Contractor, @microsoft, Nuclear Weapons Engineer, MQ-1 Analyst.
Pume Tuchinda @pumetuchinda
7 Followers 173 Following Research Assistant @VISTEC_Thailand | @PurdueECE 2023
Chris Oslund @EightTwo_Three
442 Followers 699 Following Designer working on agentic systems @Microsoft. Co-host of @FeatureCrewPod. I spend my free time trying to figure out how to make a time machine.
crystal @crystalxduan
6K Followers 3K Following your favorite youth pastor's favorite youth pastor | instagram expat turned armchair psychoanalyst | spiky marshmallow
Ray Zhu @rayzhueth
4K Followers 2K Following Jesus maxi. building https://t.co/vgaKJrlgBo. seeking to bring more wonder and play into the world
Geoffrey Nichil @GeoffreyNichil
64 Followers 479 Following
Nathan Chen @nathancgy4
1K Followers 644 Following understanding models @tilderesearch, (hardware-aligned) ml & open-source, 16
Darklord @Darklord1093741
0 Followers 77 Following
Christian Zhou-Zheng @christianazinn
7 Followers 85 Following researcher @RWKV_AI | former intern @lmstudio | external researcher @MetacreationLab | high school senior @ThePingrySchool
rank decomposition @rankdim
1K Followers 364 Following machine learning, maths, history and philosophy of sciences
Wona @Wona452
123 Followers 3K Following
MandyStowe @nWAllZ6uP8pIDSc
19 Followers 567 Following
云创兽Ai @Gacau2660250
0 Followers 20 Following 🚀 Why hunting value stocks? Ask this curious girl! open to insights. DM me to share market indices! 🚀 #ETF #Markets
Amir Bar @_amirbar
2K Followers 1K Following Research Scientist, FAIR (Meta). Prev: Postdoc. PhD @TelAvivUni @berkeley_ai
Richard Jiang @rj12186
1 Followers 340 Following ML Researcher | Musician & Film Critic | Curious Mind
StableKirito @kuer5ord
292 Followers 159 Following
George Grigorev @iamgrigorev
2K Followers 1K Following now: exploring opensource; prev: training @togethercompute, chatbots&diffusion@snap rare specialty coffee lover
Saber Darabi @SADarabi
304 Followers 7K Following
chester @chesterzelaya
13K Followers 200 Following founder @thedroneforge | scaling drone intelligence | @ucberkeley mecheng + eecs alumn
fairy @autoregressionx
28 Followers 349 Following
Tao HU @vtaohu
578 Followers 2K Following
Thea @Odirxa099
54 Followers 2K Following Don’t let anyone dim your light simply because it’s shining in their eyes.
David Stafford @davidstafford
680 Followers 3K Following AI and robotics. Bit twiddling. Opinions are my own.
Megan @Sweauife3519
39 Followers 2K Following
Francesco Sacco @FrancescoSacco1
771 Followers 311 Following Other than being a Physicist and ML researcher, I offer masonry work at modest prices. In SF from 13th to 26th of October
pdawg @prathamgrv
16K Followers 2K Following pre doctoral researcher @MSFTResearch || part time @TensorTonic
rank decomposition @rankdim
1K Followers 364 Following machine learning, maths, history and philosophy of sciences
Saravana @junglyraja
717 Followers 842 Following Gemini Canvas @GoogleDeepMind. Previously Founder of Puppetry (@puppetryai), Siri Engineer at Apple.
space tintin @space_tintin
3K Followers 1K Following ex Apollo program. ex Bell Labs. ex King Arthur’s Court.
Isk @Is36E
23 Followers 77 Following
LaurieWired @lauriewired
105K Followers 285 Following researcher @google; serial complexity unpacker; https://t.co/Vl1seeNgYK ex @ msft & aerospace
Mazeyar Moeini 👨�... @mazy1998
564 Followers 443 Following Co-Founder CTO https://t.co/TS6khHv0jO | Gaussian Splatter | Artist | Proverbs 1:7
Blake Scholl 🛫 @bscholl
110K Followers 2K Following Founder/CEO @boomaero. Life is short so if you want to do a lot, it helps to move fast.
Arcee.ai @arcee_ai
4K Followers 414 Following Optimize cost & performance with AI platforms powered by our industry-leading SLMs: Arcee Conductor for model routing, & Arcee Orchestra for agentic workflows.
Dimitri von Rütte @dvruette
2K Followers 308 Following PhD @ETH_en, prev. Machine Learning @DeepJudgeAI
Yacine Mahdid @yacinelearning
13K Followers 847 Following (neuro/ai) I make technical deep learning tutorials 👺
Peyman Milanfar @docmilanfar
94K Followers 501 Following Distinguished Scientist at Google. Computational Imaging, Machine Learning, and Vision. Tweets = personal opinions. May change or disappear over time.
Lucas @lucasdegeorge
54 Followers 337 Following PhD student at École Polytechnique (Vista) and École des Ponts (IMAGINE) Working on conditional diffusion models
Alexander Theus @Theus__A
46 Followers 25 Following PhD student in Machine Learning @ETH_en and @MPI_IS. Working on foundation models for biology 🧬, model merging 🤝, and structured pruning ✂️.
Quentin Bertrand @Qu3ntinB
1K Followers 2K Following Researcher at @Inria. Previously, postdoctoral researcher at @Mila_Quebec w/ @SimonLacosteJ and @gauthier_gidel.
Nikunj Kothari @nikunj
25K Followers 834 Following partner @fpvventures - investing in seed/A. previous: investing @khoslaventures. first pm @meter, led growth @opendoor etc. love @shimoleejhaveri + 👦👧
crystal @crystalxduan
6K Followers 3K Following your favorite youth pastor's favorite youth pastor | instagram expat turned armchair psychoanalyst | spiky marshmallow
Lindon Gao @Lindon_Gao
976 Followers 18 Following Cofounder & CEO @DynaRobotics, Caper AI ($350m exit)
Zephyr @zephyr_z9
31K Followers 503 Following Tech, AI, Semiconductors, Stocks, Finance. DMs are open
Nathan Chen @nathancgy4
1K Followers 644 Following understanding models @tilderesearch, (hardware-aligned) ml & open-source, 16
liam @liamesp
1K Followers 857 Following pixels! @openai, prev product @krea_ai, cs/music/media @princeton
Ray Zhu @rayzhueth
4K Followers 2K Following Jesus maxi. building https://t.co/vgaKJrlgBo. seeking to bring more wonder and play into the world
Pasha @pashakho
756 Followers 7K Following Interests: machine learning, probabilistic reasoning, tractable probabilistic models, and trust worthy AI.
Jane Manchun Wong @wongmjane
169K Followers 3K Following “The woman scooping Silicon Valley” — BBC・hacker turned builder + blogger・ex: Threads, Instagram, startups, etc
zack (in SF) @zack_overflow
26K Followers 2K Following 24 eng @bunjavascript i like systems programming, chief uncle @nautilusquest house
fairy @autoregressionx
28 Followers 349 Following
Rachel Lapides @rachellapides
7K Followers 2K Following comedienne & poetess @iowawriterswksp @swarthmore I am kinda interested in everything unknown to me
Kevin Lu @_kevinlu
9K Followers 226 Following @thinkymachines. formerly: - @openai: RL, synthetic data, efficient models - @berkeley_ai: decision transformer, universal computation
Ajay Jain @ajayj_
7K Followers 4K Following Co-founder @genmoai. Co-created denoising diffusion (DDPM), DreamFusion, Dream Fields. Ex Ph.D. @berkeley_ai, @googleai, @facebookai, @nvidiaai, @mit
Matt McGill @MattMcGill_
3K Followers 718 Following DeepMind, Genie 3. Interested in deep learning for embodied artificial intelligence & deeply learning what it's like to be an embodied human intelligence.
OMW @omwgamestudio
6K Followers 0 Following The official account of OMW Game Studio -- The team building the new retro fantasy RPG you’ve always wanted
Ado @ado1024imokenp
2.7M Followers 676 Following アドです。niconico:https://t.co/Bdx26K2k2K YouTube : https://t.co/A2EbwbLv1w イラストは「#Adoart」DMは事務所が管理しています。contact→ https://t.co/3GPesFPBFr
Tiny Glade 🏰🌿 o... @PounceLight
64K Followers 87 Following #TinyGlade is a small relaxing diorama builder where you doodle whimsical castles, cozy cottages & romantic ruins. 🐑 https://t.co/hNZtO5rrtb