This week I go over a few things. First some lessons from the trenches, next what is holding the AI industry back, and an amazing book I read

Lessons from the Trenches

1. Life is about acceleration, not velocity. When you fly in a plane, you don’t feel like you’re going fast. However, in a go cart you feel every bump. We are wired to feel acceleration, and not velocity. This applies in your personal life as well. Triumph feels sweeter after struggle.

2. There is always an alternative path. This week we got a lucrative new project. This project is a bet you take that can make or break you. If we do well for this client, we have unlimited interesting work… yay. If we don’t we will be flat broke with almost nothing to show. Surprisingly enough, expected value was there.

But then we ran into a big bad scary problem. We couldn’t afford the component pieces. My partner and I brainstormed and found a way to get the first round of development funded, and can now put in a great bid for the project. What’s the lesson? You can always find a new way to crack a problem.

AI’s big issue

NVIDIA is AI’s biggest issue. They own the GPU ecosystem, and may be a monopoly. No one produces as many GPU’s as them and has a software ecosystem as good as theirs. Everyone who is building large AI networks is afraid of the big bad wolf called NVIDIA.

NVIDIA does its absolute best to get as much margin out of their chips as possible and has embedded its GPU chips as central to the AI ecosystem.

What’s even crazier is that the GPU chip isn’t actually designed to do AI, but their parallelism makes them well equipped to deal with the task. That means that there is probably a lot of optimisation to be found.

OpenAI, Amazon, Google and Facebook are all beginning to create their own chips. Each is investing in building chips that are optimised to for learning and inference. This will let them out of NVIDIA’s grip and reduce their costs significantly.

OpenAI is rumored to be looking to begin building their own chips, as well as hardware (we will get into that later) to vertically integrate.

Once again OpenAI seems to be ahead of the curve. OpenAI’s founders invested in a chip company called Cerebras years ago, which could pay off dividends. Cerebras makes the fastest AI chips on the market, and are already used by many massive corporations.

I speak about computer chips a lot here, because they are the most important asset for good AI. Right now, any company that builds chips for AI training & inference is being overloaded with requests to build chips for specific purposes. There are

Sam the Banana Man

This week I read a book, called “The Fish that Ate the Whale.” It’s about how Sam “the Banana Man” Zemurray impacted the world. Sam was the American colonialist, who was single handedly responsible coups, and many deaths. His life is full of lessons, some are the value of boldness, others have to do with ethics and the value of other people’s lives.

Sam the Banana Man catapulted his way from selling scrap metal to farmers, to being a key decision maker in world politics… by way of the banana trade.

Sam’s company, United Fruit, is one of the most controversial companies of all time. They had a knack for leading coups in young democracies. Those companies had generally grown tired of United Fruit, and voted for leaders who intended to get fair deals for their citizens. Then they would see mercenary armies and mysterious US bombers appear over the horizon.

But that’s a different story.

Sam was a bold man. When JP Morgan (the guy the bank is named after), and the Secretary of State called the 32 year old Banana Man into a meeting, he didn’t know what to think. Just 20 years ago, he had been a hopeless immigrant from russia.

They told him his business was interfering with their strategy. Sam politely told them to fuck off.

Now he had a real problem. Morgan was going to put a new leader in power, who would kill his business. Sam didn’t like that. So he raised an army and put his own puppet leader in office.

Imagine that. He spit in the eye of authority, and got away with it.

Sam also did the work. He wasn’t an executive sitting in a nice office. His office was in the fields of Guatemala. He was leading men in the fields, getting bitten by bugs and chopping banana trees. He treated his men well, and paid them accordingly. As a result, they loved him.

“They’re there and we’re here; Go see for yourself; Don’t trust the report.” Sam didn’t read the reports other executives did. He was out in the field, doing the work.