Founder at @thundercompute | Fan of C++ | Previously HFT @ Citadel Securities

Making GPUs go brrr @ →
Joined March 2020
Agents write application code, but you still get paged at 2am when it breaks. We wanted to know whether AI could handle that part of the job too. Introducing Incident Arena: a benchmark that puts coding agents on call! Check out our paper & full dataset release below!
54
54
19
255
58,208
Someone should really work on oversubscribing GPUs to fix these economics
all the inference providers including big clouds want customers to take the risk for idle GPUs by making them reserve throughput this isn't surprising remember before AWS you could not at scale rent servers by the minute but it is a bit jarring since we're so used to that
2
5
352
Brian Model retweeted
Virtualization tech will absolutely dominate the next 3-5 years of cloud/agent infra, harder and with higher stakes than anything we saw in the 20-teens. I think the "sandbox wars" are just a precursor, and a microcosm of the larger discussion.
6
5
3
55
10,243
Brian Model retweeted
Thunder Compute built crazy tech to turn idle GPUs into extra capacity. Congrats @carlpeterson and @brain_model on the series A!
Today, we raised a $13M series A to solve the GPU shortage Trillions are being spent on GPU CapEx while 80% of it sits idle The capacity already exists Virtualizing GPUs unlocks it Backed by Matrix, YC, and CEAS
2
4
22
7,580
Excited to announce what we've been working on, it's a big day for GPU capacity
Today, we raised a $13M series A to solve the GPU shortage Trillions are being spent on GPU CapEx while 80% of it sits idle The capacity already exists Virtualizing GPUs unlocks it Backed by Matrix, YC, and CEAS
4
1
23
1,268
The go in golang is short for goat 🐐
4
166
This team ships 🚀
Today, we’re launching our home robot Isaac 1. Isaac 1 deliveries will begin this fall. Order yours below.
1
1
169
Tomorrow You'll want to see this
28
40
12
544
91,296
AI is somehow great at debugging distributed systems in prod but terrible at making them faster.
1
87
Sometimes all you need is to increase the MTU instead of a full rearchitecture
46
Brian Model retweeted
very true. infra engineers will have generational run for years and beyond categories. - model serving / inference infra - gpu / distributed systems infra - cloud + kubernetes + orchestration - data infra - eval + observability infra - agent / runtime infra i still think as ai advances there will be demand for areas which govern bottlenecks underneath like latency, throughput / cost / reliability / data quality and security.
Hot take the infra stuff around ai is way cooler than model training
23
87
7
1,293
80,348
I hate c++ build systems with a passion, but building deps from source gives me some peace of mind during all these supply chain attacks
1
87
Coding agents finally give hermetic build systems a real business use case
1
5
105
People want a robot that can fold their laundry, and we’re going to give it to them
22
47
26
350
92,204
Everyone's out here talking about GPT-5 but no one's talking about Nvidia dropping cuda 13
1
2
154
Probably the coolest implementation of a circular buffer I've ever seen: github.com/gnuradio/gnuradio… Essentially they mmap 2 contiguous regions of virtual memory to alias the same data. Then they can call memcpy 1 time and have the hardware automatically handle the wraparound logic.
7
626
"How to update cuda drivers"
2
131
"Small" change to optimize serialization got a little out of hand😮‍💨
2
140