For most of tech history, engineers had an alternative to working for a company - starting their own. With a $5 VPS and a $300 laptop, it is possible to build a product and grow it to thousands of users.

In contrast, training your own LLM requires access to significant upfront capital and hardware - around $100M. Today, you can use APIs, but this leaves you completely at the mercy of an LLM provider.

Think about airline pilots. Even the best pilot is useless without an airplane. An extremely expensive airplane that only companies with significant capital can provide. Coincidentally, Airbus A320neo costs are similar to a modern LLM training.

If LLMs become the necessary “engines” of software, engineers are going to lose their edge.

Will LLMs become jet airliners of tech?

In Part One I described the foundation for building a self-driving car out of LEGO. Here we will dive deeper into details about how to control the car and how to get 3D point cloud from LIDAR. In the end, the car will be able to drive around and avoid walls.

In this post, I’ll go over the technical decisions that went into making a self-driving car out of LEGO and an iPhone 13 Pro.