Joey @him_joey

Transforming human knowledge, sensors and actuators from human-first and human-legible to LLM-first and LLM-legible is a beautiful space with so much potential and so much can be done... One example I'm obsessed with recently - for every textbook pdf/epub, there is a perfect…

2024: everyone releasing their own Chat 2025: everyone releasing their own Code

How to build responsible AI❓ New preprint by Himanshu Josh @him_joey preprints.org/manuscript/202… introduces a framework to embed ethics&oversight throughout the AI lifecycle, addresses autonomous decision-making for #TrustworthyAI #Preprints #OpenScience #AgenticAI #AIGovernance

Good post from @balajis on the "verification gap". You could see it as there being two modes in creation. Borrowing GAN terminology: 1) generation and 2) discrimination. e.g. painting - you make a brush stroke (1) and then you look for a while to see if you improved the…

Balaji @balajis

4 months ago

Good post from @balajis on the "verification gap". You could see it as there being two modes in creation. Borrowing GAN terminology: 1) generation and 2) discrimination. e.g. painting - you make a brush stroke (1) and then you look for a while to see if you improved the…

213 453 3K 981K 2K

*IIT Bombay Releases 16 AI Datasets on AIKOSH: Enabling the Future of Responsible AI in India 🇮🇳* IIT Bombay is thrilled to announce the release of 16 diverse and culturally significant datasets on AIKOSH, the Government of India’s official AI repository. These datasets are…

Starship to Mars

So so so cool. Llama 1B batch one inference in one single CUDA kernel, deleting synchronization boundaries imposed by breaking the computation into a series of kernels called in sequence. The *optimal* orchestration of compute and memory is only achievable in this way.

Benjamin F Spector @bfspector

4 months ago

So so so cool. Llama 1B batch one inference in one single CUDA kernel, deleting synchronization boundaries imposed by breaking the computation into a series of kernels called in sequence. The *optimal* orchestration of compute and memory is only achievable in this way.

34 145 888 375K 527

Download Image

Wow! @Google alone processes roughly 100,000 words per month for every internet user? Another 100x coming with world models? @sundarpichai told I/O that the company now processes 480 trillion tokens a month – 50× more than a year ago – and almost 5x more than the 100 trillion…

We're missing (at least one) major paradigm for LLM learning. Not sure what to call it, possibly it has a name - system prompt learning? Pretraining is for knowledge. Finetuning (SL/RL) is for habitual behavior. Both of these involve a change in parameters but a lot of human…

There's a new paper circulating looking in detail at LMArena leaderboard: "The Leaderboard Illusion" arxiv.org/abs/2504.20879 I first became a bit suspicious when at one point a while back, a Gemini model scored #1 way above the second best, but when I tried to switch for a few…

lmarena.ai @lmarena_ai

5 months ago

There's a new paper circulating looking in detail at LMArena leaderboard: "The Leaderboard Illusion" arxiv.org/abs/2504.20879 I first became a bit suspicious when at one point a while back, a Gemini model scored #1 way above the second best, but when I tried to switch for a few…

28 31 362 664K 143

I attended a vibe coding hackathon recently and used the chance to build a web app (with auth, payments, deploy, etc.). I tinker but I am not a web dev by background, so besides the app, I was very interested in what it's like to vibe code a full web app today. As such, I wrote…

I was reading the docs of a service yesterday feeling like a neanderthal. The docs were asking me to go to a url and click top right and enter this and that and click submit and I was like what is this 2024?

PSA It’s a new era of ergonomics. The primary audience of your thing (product, service, library, …) is now an LLM, not a human. LLMs don’t like to navigate, they like to scrape. LLMs don’t like to see, they like to read. LLMs don’t like to click, they like to curl. Etc etc.

Andrej Karpathy @karpathy

5 months ago

PSA It’s a new era of ergonomics. The primary audience of your thing (product, service, library, …) is now an LLM, not a human. LLMs don’t like to navigate, they like to scrape. LLMs don’t like to see, they like to read. LLMs don’t like to click, they like to curl. Etc etc.

136 234 4K 888K 1K

The reality of building web apps in 2025 is that it's a bit like assembling IKEA furniture. There's no "full-stack" product with batteries included, you have to piece together and configure many individual services: - frontend / backend (e.g. React, Next.js, APIs) - hosting…

When working with LLMs I am used to starting "New Conversation" for each request. But there is also the polar opposite approach of keeping one giant conversation going forever. The standard approach can still choose to use a Memory tool to write things down in between…

New 2h11m YouTube video: How I Use LLMs This video continues my general audience series. The last one focused on how LLMs are trained, so I wanted to follow up with a more practical guide of the entire LLM ecosystem, including lots of examples of use in my own life. Chapters…

Agency > Intelligence I had this intuitively wrong for decades, I think due to a pervasive cultural veneration of intelligence, various entertainment/media, obsession with IQ etc. Agency is significantly more powerful and significantly more scarce. Are you hiring for agency? Are…

Garry Tan @garrytan

7 months ago

Agency > Intelligence I had this intuitively wrong for decades, I think due to a pervasive cultural veneration of intelligence, various entertainment/media, obsession with IQ etc. Agency is significantly more powerful and significantly more scarce. Are you hiring for agency? Are…

117 485 5K 2.0M 2K

Narendra Modi is one of the most fascinating human beings I have ever studied. I can't wait to talk to him on podcast for several hours in a few weeks. On top of the complex, deep history of India, and his role in it, just the human side of Modi is really interesting. For…

The buzz over DeepSeek this week crystallized, for many people, a few important trends that have been happening in plain sight: (i) China is catching up to the U.S. in generative AI, with implications for the AI supply chain. (ii) Open weight models are commoditizing the…

We have to take the LLMs to school. When you open any textbook, you'll see three major types of information: 1. Background information / exposition. The meat of the textbook that explains concepts. As you attend over it, your brain is training on that data. This is equivalent…