“Spend more time learning stuff that you know you might never understand.” – Futurist Jim Carroll
Follow the edges.
That’s always been my advice to many organizations that are trying to understand the future. It’s out on the edges – beyond the traditional world of R&D and discovery – that much of the future is invented. When it comes to technology, much of this happens with what we call open-source computing.
Back in 1997, looking out to the edges, I realized that something “big” was happening with something called Linux – a computer operating system being developed to compensate for the closed system environments offered up by companies like Microsoft. I dove in, not fully understanding it but knowing it would b important – and turned out to be right, as Linux now dominates most of the computers and tablets and embedded devices in the world. Today, I’ve somewhat mastered many aspects of Linux – there are about a half dozen Linux based computers in my home which I use for various purposes. And as Linux has come to define so much of our technology future, and since our future is built on technology, I have a deeper understanding of the future than I might otherwise have.
And so, I’m always ‘digging around the edges,’ watching for signs of unique things that might be happening. Yesterday, while continuing my research into AI for a number of upcoming keynotes, I found, perhaps, the same thing – the Tiny Corp.
What is it all about?
The Website is all rather incomprehensible to most of us, but I sent down the rabbit hole into the mind of a fellow named George Hotz – a ‘hacker’ who managed to be the first to ‘jailbreak’ (i.e.get into the software ‘innards’) of an iPhone, as well as other exploits involving the Sony Playstation. Once he got beyond those teenage pursuits, he founded Comma.ai, which has done substantial and significant work in self-driving, autonomous vehicle software for virtually any type of modern car. He has a deeply complex mind, stunning computer exploits, and a pretty eccentric persona. Precisely the type of person that invents the future. L up.
I ended up watching this video which is a rather fascinating yet complex summary of what he is trying to accomplish with the project.
Bottom line? It seems he wants to build an everyday computer that would let you run your own AI in your own home – one that would be powerful enough to run the most sophisticated AI technologies in the world. Why? So that the future of computing is not dominated by companies like Apple, Google, Amazon, and OpenAI.
I used a video summary plugin in ChatGPT to summarize the video since much of it is a bit incomprehensible.
Based on the transcript, here are the key points and quotes about the future of computing:
**10 Key Points:**
1. Discussion about building custom PCs and the introduction of the “Tiny Box” from Tiny Corp.
2. The Tiny Box costs $15,000 and offers almost a petaflop of compute.
3. It has over 100 gigabytes of GPU RAM and over five terabytes per second of GPU memory bandwidth.
4. The Tiny Box will have four NVMe drives in RAID, offering 20-30 gigabytes per second of drive read bandwidth.
5. The main limiting spec for the Tiny Box is a 120-volt 15-amp circuit.
6. The Tiny Rack is introduced, which can utilize the power of an electric car charger, offering about 10,000 watts.
7. The GPUs for the Tiny Box are still being explored, with a leaning towards AMD.
8. The goal for the Tiny Box is to run a 65 billion parameter language model (like Llama) in real-time.
9. The Tiny Box will contain six GPUs.
10. Cooling is a significant challenge for the Tiny Box, with the aim to make it quiet and suitable for home use.
What does all this mean? This is a REALLY powerful computer – like, REALLY powerful. Going down the rabbit hole, it turns out the human brain has about 20 petaflops of computing capacity; this little home supercomputer would give you one petaflop. That’s pretty significant – it takes us into a whole different world.
What are his goals? ChatGPT provided some insight from the video:
**5 Key Quotes about the Future of Computing:**
1. “I’m gonna build like the best deep learning box that I can that plugs into one wall outlet.”
2. “Open source is a culture… open source means developing in public.”
3. “Experience a chat with the largest language model that you can have in your house, from a wall plug.”
4. “Quiet because you want to put this thing in your house and you want it to coexist with you.”
5. “I want to give you a great out of the box experience. I want you to lift this thing out; I want it to be like the Mac.”
Notably, he has stated these goals for the company:
I don’t want to live in a world of closed AI running in a cloud you’ve never seen, I want everyone to have an AI that they own, both training and inference. I want compute to be available from 50 different companies all competing to drive the price to zero. And I want an open source framework to run cutting edge AI on any one of those 50 chips as seamlessly as Linux supports 50 network cards.
Why? To provide alternatives to what mainstream companies are working on:
If Elon has the FSD Chip and Dojo for Autopilot and Tesla’s robots, we have the tiny corp for openpilot and comma’s robots. comma, along with 100s of other companies, will need computers both big and small for training and inference. We will sell them those computers. If NVIDIA is the Apple, we are the Android.
How will he hire people for the company and the project? There is no hiring process other than writing and submitting beautiful computer code that might help in the process of building the TinyBox.
There’s only one way to get hired at the tiny corp, and that’s by submitting high quality pull requests to tinygrad. Job interviews are obsolete, prove you can do the job by doing the job!
What does it all mean?
It means that ‘out there,’ there are always people building a different future, using different concepts, with different ideas, and with different pathways. They are different from most other people – as Steve Jobs famously said, work with the oddballs, the rebels, the freaks – because much of our future is defined by these oddballs, the rebels, and the freaks. Tinybox might emerge as a success or it might not, but it does represent a significant new potential pathway for our world fo AI. It’s happening out on the edges, far away from most normal news coveage.
And so,even if I don’t fully understand it, I do know enough that there is something important happening here.
And that’s how you can often interpret and understand the future before it happens!