Day 9 of Sprited

Today’s Day 9 of Sprited. I am at TechEX AI & Big Data Expo at Santa Clara Convention Center. The expo floor is crowded but AI section of the expo is only quarter of the floor. There is another much bigger AI convention going on in San Francisco at the same time, so I think a lot of companies might have chosen to go there instead.

The focus is on Agentic AI which replaces human being not only for repetitive task but basically fully replaces an employee or set of employees with ability to make decisions on their own.

It seems very likely that agentic AI is going to replace lots of human beings at workplaces. Human being’s job would be to best optimize and orchestrate them.

Tinker project is very embodiment of this. The vision for Tinker is to fully replace me as a coder to free me up time to work on more interesting things like building virtual world for AI.

This particular problem space is very crowded with large players like OpenAI, Anthropic and Amazon. However, my instinct tells me that individual engineers should build their own AI agents to be your own agent.

The main reason why I think everyone should build their own agent is because agentic AI released to public will not give you an edge over other people or companies. If it is availbel to public then you might be already too late because everyone else will be using it too.

To be clear, when building agents, we shouldn’t strive to replicate what OpenAI or Anthropic or Cursor does, but rather be an agent that sits on top of all those providers. They will provide the power, and your agent will harness it and orchestrate the tasks the best way it can.

One the other hand, Tinker shouldn’t be the end goal. Building an agentic AI should be the “how“ of our story, and not the destination. We need agentic AI experiences to create new things that were not possible before.

Not sure if this is a good example but so far there was no way for single person to create animated featured films. Single person could not create triple A titles in games. Now, I think there is real possibility with determined mind, one person can create something truly valuable. Industries has been limited with cognitive resources, now the industries can create things that were impossible to do before.

Remember when you were a kid, you may have drawn a product idea like a flying car? Something like that may become possible with help of AI accelerating the research. For me, I had an idea of using living microbes to paint walls. Imagine you just have to drop a droplet of microbes to paint the full wall. Using agentic AI, you could farm out the research.

What then is end-goal of Sprited? We are still in the soul-searching phase, but it is something along the line of using AI to build new experiences that wasn’t possible before. For me personally, creating a virtual living worlds and experiences is one thing that I wanted to do.