Machine learning, renamed “AI” with the LLM boom, does not simulate intelligence. It integrates feedback loops, which is kind of like learning and it uses a network of nodes which kind of look like neurons if you squint from a distance. These networks have been around for many decades, I’ve built a bunch myself in college, and they’re at their core just polynomial functions with a lot of parameters. Current technology allows very large networks and networks of networks, but it’s still not in any way similar to brains.

There is separate research into simulating neurons and brains, but that is separate from machine learning.

Also we don’t actually understand how our brains work at the level where we could copy them. We understand some things and have some educated guesses on others, but overall it’s pretty much a mistery still.

There’s no technical reason to think we won’t in the next ~20-50 years

Other than that nobody has any idea how to go about it? The things called “AI” today are not precursors to AGI. The search for strong AI is still nowhere close to any breakthroughs.