I could train a 1B-A200m model on an iPhone 17 Pro at ~650 tokens/sec. It will take 360 days on 20B tokens of training data and use 156KW of electricity which cost $51.

The phone will fry of course, so I wrote algorithms to run inference on your phone rather. We named it after a plant that survives in resource-constrained environments, the Cactus.

can run similar model on your Grandma’s Pixel 6a at 80 tokens/second while only draining 10% battery per hour of continuous inference and using 250MB RAM only.

I also built an agentic foundation model for tiny smart devices that runs 1.2k toks/sec on Cactus, its called Needle.

Needle has no FFN, just 26m params of pure attention, yet matches 10-25x bigger models. It can be a better Siri/Alexa that powers tiny smart devices.

We raised some money from YCombinator, Oxford's Seed Fund, FCVC (Slack, Coinbase, GitLab, Instacart etc.), and 6 smaller funds like Transpose (run by Garry Tan's brother), fellow YC founders, and 62 tech CTOs/VP/Dir both via syndicate and directly at Google DeepMind etc.

Cactus now powers cool products you've probably heard of...I think. 6 exceptionally gifted "Cactus Jacks" from UCLA, Nokia, Google, Stanford, Oxford, have joined us!

Same destination, just a different route!

• After CUDARepo, Nvidia reached out, I did 7 technical rounds, got a verbal offer, back-and-forth over YOE/pay, then I got YC.
• Did MSc at QMUL, just to work with Prof Matt Purver (Ex-Stanford Researcher on CALO), did my project/thesis with his team.
• Did BEng under Prof Onyema Uzoamaka (Rumoured first Nigerian CS grad from MIT), he taught computing archs off-head!
• Biggest career miss was a PhD Studentship at Meta FAIR.

Speaking at UCLA	The Cactus Jacks	Team dinner
I'm fit too	Cactus Jacks at YC	Cactus Pod at YC HQ

Expressive Rap
Alternative/Folk
Soul/Jazz
Oldies
Genre-Blending Urban
Dark Pop