The constraint is skilled humans who can wire, maintain, and expand these systems.To put real numbers on it: Nvidia is paying top-tier electricians as much as $495,000 a year on certain hyperscale and AI-adjacent projects.

-the best thing you can do is to go to IN in person and meet people, in particular other local real live human investors at real live in-person meetups. 

Agreed on not replacing human relationships with AI, but it's way more than a tool at this point.

Id recommend getting comfortable with the terms.Some books that are helpful (not totally MF focused)-Confessions of a Real Estate Entreprenuer (Perspective of an investor, how they think, how they take risks and how they learn)-Am I Being Too Subtle (Mindset, managed risk, opportunity spotting)-Zeckendorf - Autobiography (Relationships, risk taking, deal structure, development)-BP Rental Property InvestingI'd also recommend going to Lowe's, Home Depot, Habitat for Humanity to build your brain around materials, costs, quality and history of materials (even if you're not the one doing the work).Lastly, once you've conceptualized the asset class itself and general understanding of deal structure / operating structure...look into some news articles and trends in MF.

OP if you have the time and inclination work for Habitat for Humanity on a couple projects to get base hands on knowledge.  

One thing I’ve seen repeatedly is that most underwriting mistakes aren’t spreadsheet errors - they’re assumption errors that go unchallenged.A few that tend to show up only after someone has been burned once:• Treating in-place performance as “temporarily bad” instead of asking why it exists and whether it’s actually durable• Assuming value-add execution speed without factoring in operational drag, leasing friction, or human error• Believing future upside will be underwritten the same way current income is - it rarely is• Underestimating how aggressively lenders normalize expenses, reserves, taxes, and insurance compared to investor models• Thinking a strong sponsor or guarantor can compensate for weak cash flow under stressAnother big one is timing.

The Seller can't stand real estate agents (there are reasons this is actually true) Nothing personal here Steve, it's just a fact of human nature7.

Maybe for properties that has clear cut comps but for more coplicated properties it needs human i tried this program for 3 properties completely wrong.

Humans are notorious for over estimating their abilities. 

There are probably a lot of other obvious things they didn't think they needed to tell you either, like how investing in Section 8 properties can be extremely challenging at times, and doing it remotely is a lot harder/ less chance of success, or how whenever dealing with human tenants and physical properties there are a lot of uncontrollable variables/ things that can go wrong, that landlording is difficult even if it's down the street and a lot harder if it's across the country, how everyone in this world is only ever looking out for their own self-interests first, that properties that are cheap are cheap for a reason, that out of state investors are easy to take advantage of, that AI lacks common sense and can't be relied upon, that even the best spreadsheet magic is just a wild guess of how a property will actually perform in real life over time, etc.