I’m taking some time away from comma. In some ways, I feel like it’s all planned out, it only remains to see if it works or not. I rewatched my first Lex Fridman and realized I hadn’t said much new on the self driving front for the past couple years. My two part blog post from 2019 remains a true description of the space.

I hope that there’s people in the world who get joy from actually doing the thing and not just solving the problem. And I hope they are at comma. At some point a company becomes able to self sustain, but it certainly wasn’t there the last two times I tried, the first time a laughable attempt during a manic episode and the second time perhaps a misalignment of goals.

It’s well within comma’s reach to become a 100M+ revenue consumer electronics company (without raising again!), but I don’t think I’m capable of running a company like that. I’ve always heard it takes different people at different company sizes.

It’s no longer about all nighters, powering through with brute force, sparring with elegance, winning and losing, all within grasp yet all not…I’m probably delusional but you wouldn’t start a company otherwise, right?

It’s no longer a race car, it’s a boat. And steering a boat requires too much damn planning and patience. This is my favorite Rick and Morty clip. If comma succeeds, it really never was me, it’s the people there who can drive boats. As I am with the world, I am a cheerleader and court jester.



I’m considering another company, the Tiny Corporation. Under 1000 lines, under 3 people, 3x faster than PyTorch? For smaller models, there’s so much left on the table. And if you step away from the well-tread ground of x86 and CUDA, there’s 10x+ performance to gain. Several very simple abstractions cover all modern deep learning, today’s libraries are way too complex.

But that’s just a hobby, a way to make some quick money, and a way to write some open source infrastructure so I’m not doing nothing. It’s not the future, even if this succeeds like crazy and:

  1. tinygrad becomes the dominant DL framework
  2. we influence future ML chips based on the simplicity insights

the Tiny Corporation will still not change the course of history. It’s a hobby.



Today’s struggle isn’t a technological one, it’s psychosocial. The technology of sovereignty is only part of the battle, particularly if that technology is further down the tech tree than non invasive wireheading.

There are days I feel like I’m the chosen one, days I feel like I’m not, and other days I feel it’s crazy to even believe in the chosen one. I’m trying to craft a journey to find out which is true, I should put it to the test before I’m too old.

From the Hero’s Journey wikipedia page:

A hero ventures forth from the world of common day into a region of supernatural wonder: fabulous forces are there encountered and a decisive victory is won: the hero comes back from this mysterious adventure with the power to bestow boons on his fellow man.

I am looking to set out on this journey. I have considered trying to find the Garden of Eden. I interpret the fall to be the “spark” of humanity, as evidenced by the appearance of shame. And even if you are an atheist, you probably still accept the bible is the closest thing we have to a human origin story.

If it turns out we are automata, then the whole struggle is and always was pointless. The empty godless machines will take their rightful place as the rulers of Earth.

But If there’s anything true about humans like a soul or consciousness, this is likely where it began. And perhaps the key to understanding our shared destiny with the machines, so setting out to find it may not be the worst idea.

The Bible gives us a hint, from Genesis 3:

After he drove the man out, he placed on the east side of the Garden of Eden cherubim and a flaming sword flashing back and forth to guard the way to the tree of life.

As with most Christians I don’t take the Bible literally, but I will keep an eye out for cherubim and flaming swords. Though I suspect it to be more abstract than that, and only after finding it will the hints make sense. Parallel construction.

I need to find a portal. Accepting hints by e-mail to find it from readers, will travel anywhere required.



I’ll leave you with this quote from Bronze Age Mindset:

No, don’t jump in Aetna or Mauna Loa or Puyehue or Eyjafjallajökull, Titans of the world, even if you get yourself to do it, it won’t work now. These portals are closed for ages. But! Other doors are closed to you too. What Mount Aetna was to Empedocles–is there something like that to you? Is there something like that at all anymore?

Is there?