Neuralink cyborgs

May 7, 2026

Recently I’ve been thinking a lot about how to best spend my time for the next 10 years or so. What should I study? Am I studying the right things? Should I really be working on projects?

A lot of those thoughts are influenced by how I think ai will be used in the future. Common sense says that I should be working towards getting better at things that are harder for ai to replace. Problem with that is

• There are lots of fundamental studies, across many fields, which ai can already easily do but I should know as background knowledge
• It is unlikely that I will ever do deep think that ai can’t do, but deep think is fun and I think is the driving reason to do engineering / science in life
  • Typically, deep think rarely includes actually making stuff for me; deep think happens during design, or difficult debuggging, not laying out xx board or making xx schematic (this is usually boring / easy think)

What I find humans are better able to do than ai is systems-level reasoning, at least very fast. This is probably just because you have a cheap context window compared to models i.e. claude pro cannot factor in enough data to reason about large systems over a long period of time. But models are still useful for doing gruntwork / quick research / quick thinking. This makes it clear - the limitation to maximizing model usage is a cracked human computer interface. This isn’t a new idea at all.

That said, it does lead to interesting conclusions about what the future may look like; will we have people with chips in their head, eternally connected to large models that control them? What will those people do? Who will control them / the model? How will they feel?

I think this would be an interesting story, if it doesn’t already exist. Life of one of these professional thinkers, who has presumably offloaded all thought for all of their life to the model, learning to think for the first time. Honestly not a crazy idea but I’d read a book on it.