Organoid Intelligence

Since my last post I have realized that biology is in sigificant need of automation. In order to facilitate biological automation, I have been taking a machine learning class through edx. The class is great and I am learning a lot. But, I have made a few observations:

1. It is extremely complicated to actually build a system that can learn something new. Given this complication, how long will it take to actually build Artificial General Intelligence (AGI)?
   I find the concept of AGI fascinating. Creation of AGI aims to create better-than-human or at least human-like intelligence artifically. Just imagine what we will be able to do with unlimited, on-demand, custom-built intelligence. It will lead us to a time of true abundance. Things like cure for rare diseases, solutions to climate change might just be around the corner due to AGI.
2. The energy required to run the training is truly humungous. The energy difference between GPT-2 and GPT-4 is 100x. We cannot possibly keep this up.

So, what should we do?

With AGI our goal is to mimic the human brain, right? So, why don’t we just use the human brain to produce intelligence?

Enter brain organoids/biocomputers/wetware.

Interest in organoid research has grown over the last decade. The idea is that we can now grow various organs in lab for study and research using stem cells. Brain organoids were created almost 10 years and have shown signs of intelligence. While this research is still in its infancy, there is an emerging field of research, called Organoid Intelligence. The idea behind organoid intelligence is to grow and train the brain in a lab and then use it to deliver intelligence. Perhaps there can even be brain farms that deliver intelligence over the Internet.

I am excited about this future :).