Abi

@caydengineer Let’s meet. DM me

I'm leaving for China in 5 hours We're making thousands of smart glasses and I'm headed to the assembly line 1 month, 5 cities, I'm live-streaming every stop - Hong Kong, Shenzhen Japan, Thailand, NYC Join the fun twitter.com/i/broadcasts/1…

@sundeep One of the reasons our edge device is battery free

TIL that an 8B parameter model running on a modern phone drains the battery in approximately 30 minutes of usage…

@hthieblot lol i would have had lots of investors if they thought like this 👍

If a founder shows you their personal bank account and it's at-$47, but they're still grinding every day, telling you they are going to make it. Invest in them immediately.

@alexocheema @kDasme @nvidia @Apple @GroqInc NVlink??

@kDasme @nvidia @Apple Here’s a clue. @nvidia x @GroqInc

Big moment coming for local AI. @nvidia x @apple

In Oct of 2019, Brian ran up to me at a VC panel talk at a bank event. “We have to talk” I connected Brian and Tyler, two wily scientists, to Nick on our team who is our resident space geek. Shortly after, we wrote their first check to create GPS 2.0, way before deep tech was the talk of the town. Be the first check, not the last!

@StephNass Really nice response

A founder emailed me: “I’m 50. Is that why I’m failing to raise?” Here’s what I replied:

@jesselyu Can you please reduce the number of cards. Too many “features”

we completely rebuilt the voice recorder on r1 and now it's one of the best experience on the market (and remains free no subscription, unlike Plaud) - OTA update tomorrow!

@tkexpress11 Would love to chat. Building new category consumer hardware device. Nothing like you have seen before. Prototype built. Testing stage.

i'm opening my calendar for 15 min slots this Friday! a16z speedrun application opening is just weeks away - come AMA 1:1 first come, first served... drop a comment / like and I'll send you a link to schedule! (if there's space remaining)

You just sold your company for $ 100M. What’s the first thing you buy?

@gajesh @alexocheema @exolabs @alexocheema is goat

met the local inference goat @alexocheema; what a legend. we went through my napkin math for decentralized inference costs, and he found a much cleaner way to optimize the system and increase tokens/sec. super excited about @exolabs + apple inference.

@karpathy Love that book!!! Very popcorn worthy. Sometimes its good to have some entertainment sneak in with the science.

Had to go see Project Hail Mary right away (it's based on the book of Andy Weir, of also The Martian fame). Both very pleased and relieved to say that 1) the movie sticks very close to the book in both content and tone and 2) is really well executed. The book is one of my favorites when it comes to alien portrayals because a lot of thought was clearly given to the scientific details of an alternate biochemistry, evolutionary history, sensorium, psychology, language, tech tree, etc. It's different enough that it is highly creative and plausible, but also similar enough that you get a compelling story and one of the best bromances in fiction. Not to mention the other (single-cellular) aliens. I can count fictional portrayals of aliens of this depth on one hand. A lot of these aspects are briefly featured - if you read the book you'll spot them but if you haven't, the movie can't spend the time to do them justice. I'll say that the movie inches a little too much into the superhero movie tropes with the pacing, the quips, the Bathos and such for my taste, and we get a little bit less the grand of Interstellar and a little bit less of the science of The Martian, but I think it's ok considering the tone of the original content. And it does really well where it counts - on Rocky and the bromance. Thank you to the film crew for the gem!

@levidiamode @charles_irl This is sick

Day 74/365 of GPU Programming I always found die shots and SM diagrams beautiful but difficult to map mentally, so I've been trying to find a way to interact with GPUs in 3D. This is what I have so far: a single input that goes through a simplified H100 execution pipeline to see what the silicon is doing at each step; from CPU-side tokenization and embedding lookup, through matmuls on tensor cores to the final softmax output. My current plan is to make this an interactive playground that lets you zoom in and zoom out through various levels of depth (package → die → GPC → SM → tensor core) while also including step-through examples similar to the bycroft LLM 3D visualization. Ideally this should make exploring the architectural side just as easy as mapping CUDA abstractions onto the actual hardware processes. I'm starting with an H100 but would be fun to expand this to more GPUs and highlight the differences between generations. This was largely inspired by @srush_nlp's GPU puzzles, @JayAlammar's Illustrated Transformers and @karpathy's makemore series, which made me think about how to study and visualize GPUs from the ground up.

@brahma_4u We are building something better with most parts sourced from India! Regarding the device mentioned above, most of the parts are off the shelf. Raspberry for compute and Respeaker for audio I think. So not too hard to put one together

Anybody building them in India?

I have a theory. The richest people in US are not the richest on S&P but the ones running the biggest frauds from within the government. cc @nickshirleyy

@tldraw THIS IS AWESOME!

Just published github.com/tldraw/tldraw-… - Mac arm only atm - start the app - ask your agent to curl localhost:7236

@tldraw @AlbadDsd17946 Probably packing entire chromium in there....

@AlbadDsd17946 It's a vibe coded electron app wrapper around tldraw

@CollinRugg Been there. Done that. Bad idea only because laptops and fine San Diego sand dont mix well. Suggest an iPad instead