Sahil Thaker @_sahilt

♥️ Machines, Math, Sports, Music. Co-Founder @hyperline_xyz Past: @Glean, @Facebook News Feed AI, Video AI, @Uber marketplace optimization sahilt.com San Francisco, CA Joined January 2014

Tweets

369
Followers

1K
Following

2K
Likes

893

Jamie Voynow @jamievoynow

4 weeks ago

I took 100g of creatine and now i’m doing LLM inference in my head

136 402 9K 387K 383

TL;DR: We built a transformer-based payments foundation model. It works. For years, Stripe has been using machine learning models trained on discrete features (BIN, zip, payment method, etc.) to improve our products for users. And these feature-by-feature efforts have worked…

181 556 5K 1.3M 4K

Balaji @balajis

7 months ago

FROM MAGA TO CHINA Here are four things MAGA is getting wrong, and why it's handing over the world to China. (1) First, MAGA correctly understands that America’s economic position is in decline but thinks this is due to economic competition itself, rather than lack of…

850 2K 14K 3.0M 7K

Sahil Thaker @_sahilt

7 months ago

one way to think about ai: its humanity's collective brainpower over time i think we'll get experts and fragments to this central brain, but directionally its "collective intelligence"

Derya Unutmaz, MD @DeryaTR_

8 months ago

one way to think about ai: its humanity's collective brainpower over time i think we'll get experts and fragments to this central brain, but directionally its "collective intelligence"

108 273 3K 464K 944

0 0 1 476 0

Sahil Thaker @_sahilt

7 months ago

Came across this neat tool with viz to dig into llm internals: bbycroft.net/llm

0 0 1 181 0

Sahil Thaker @_sahilt

8 months ago

Great to see meaningful usecases of reasoning models

Derya Unutmaz, MD @DeryaTR_

8 months ago

Great to see meaningful usecases of reasoning models

139 427 4K 359K 951

0 0 2 365 0

Andrej Karpathy @karpathy

10 months ago

The (true) story of development and inspiration behind the "attention" operator, the one in "Attention is All you Need" that introduced the Transformer. From personal email correspondence with the author @DBahdanau ~2 years ago, published here and now (with permission) following…

137 997 7K 830K 5K

Download Image

Sahil Thaker @_sahilt

11 months ago

Deepmind chess paper showed that deterministic software can be mapped to neural nets arxiv.org/abs/2402.04494 This allows us to start from existing knowledge, but design the system to self-improve it through experience

Shital Shah @sytelus

11 months ago

33 43 523 43K 117

0 0 9 973 0

Crémieux @cremieuxrecueil

12 months ago

The benefits of caloric restriction for primate lifespans (including our own!) are probably overrated. We have two rhesus monkey studies, and they do not support big benefits🧵 First, take a look at this diagram:

25 41 447 105K 201

Download Image

Nivi @nivi

a year ago

What’s the difference between reason, intelligence and understanding? Reason is all methods of criticism that improve ideas. Intelligence is how fast we create ideas that solve problems. Understanding is the ability to address criticisms of an idea. Rationality is seeking to…

33 19 140 10K 115

Deedy @deedydas

a year ago

“Experience is what you get when you didn't get what you wanted.” “When you're screwing up and nobody's saying anything to you anymore, that means they gave up.” Video: youtube.com/watch?v=ji5_Mq… Book: amazon.com/Last-Lecture-R… 2/4

1 8 178 13K 99

Sahil Thaker @_sahilt

a year ago

If you set reasonable milestones, you will only make reasonable progress. unreasonable people drive to make unreasonable progress.

Elon Musk @elonmusk

a year ago

If you set reasonable milestones, you will only make reasonable progress. unreasonable people drive to make unreasonable progress.

21K 29K 223K 51.6M 10K

2 1 9 448 0

Sean Kelly @seanpk

a year ago

Elon Musk open-sourced all of Tesla's patents for other car companies to copy. Wall Street called it 'dumb' and thought Tesla will be crushed. Elon was right. Everyone was wrong. Every entrepreneur needs to understand why he did it and why it worked:🧵

2K 5K 37K 7.2M 8K

Download Image

martin_casado @martin_casado

a year ago

I'm shocked. Just shocked that continually averaging a data corpus without exogenous inputs results in degraded information quality. Doomer ouroboros arguments were always silly. But they get demonstrably sillier as the research continues. nature.com/articles/s4158…

13 13 183 23K 47

Hemant Mohapatra @MohapatraHemant

a year ago

My google exp reinforced a few learnings for me: (1) consumers buy products; enterprises buy platforms. (2) distribution advantages overtake product / tech advantages and (3) companies that reach PMF & then under-invest in S&M risk staying niche players or worse: get taken down.

1 29 223 18K 103

Raffi Hotter @raffi_hotter

a year ago

200x compression is mathematically impossible for this data. Given the specs of Neuralink's chip, you cannot do better than 5.3x compression. Here's why 🧵 (1/10)

Felix @felix_red_panda

a year ago

200x compression is mathematically impossible for this data. Given the specs of Neuralink's chip, you cannot do better than 5.3x compression. Here's why 🧵 (1/10)

449 425 9K 4.7M 3K

Download Image

40 52 814 519K 609

Sahil Thaker @_sahilt

a year ago

This MIT research as well as anthropics monosemantic paper discovered that LLMs converge to same concept representations across models and modalities. wow! reminds us of word2vec analogies. I believe this is the root of observed intelligence in LLMs

Jim Fan @DrJimFan

a year ago

44 93 585 85K 339

Download Image

1 1 4 241 0

Sahil Thaker @_sahilt

a year ago

'values are only interesting to the extent that the opposite of your value is also viable'

0 0 3 97 0

Sahil Thaker @_sahilt

a year ago

'every exponential (system) is a sigmoid in disguise' is such a profound observation

0 0 1 87 0

Arvind Narayanan @random_walker

a year ago

In the late 1960s top airplane speeds were increasing dramatically. People assumed the trend would continue. Pan Am was pre-booking flights to the moon. But it turned out the trend was about to fall off a cliff. I think it's the same thing with AI scaling — it's going to run…