How much of a CRE broker's job actually requires a CRE broker? A guy in Boulder couldn't find an agent to go below 6%. So Mike Chambers sold the house himself, went viral under @realtorshateme, and raised $6.4 million from Fifth Wall. That's Ridley. Around 300 sales in. The company says sellers save an average of $43,639. Their number, unaudited. Do the division: at a 3% listing side, that's a home near $1.45 million. Some of that came from the 2024 rule change, not software. So I split the commercial job into what a machine can do and what it can't, and put a number on each. My judgment, not a study. Most likely wrong. The top of that list is longer than most brokers would like. OMs are document assemblies. Listing distribution, DD trackers, closing coordination. Comp selection too, and the valuation math on top of it. Proximity, vintage, size, unit mix, condition. It's a weighted formula. I've written the filter. The seat that goes first is the analyst. Two juniors become one, long before a senior broker loses a listing. Most brokers reading this scanned the bottom four first. Proof of market survives. A seller doesn't want a price, he wants evidence that this WAS the price, and the only way to make it is to get humans to show up and bid against each other. Signing the file survives because hallucination cost scales with deal size. Document assembly, where a mistake costs $200, is a different product from one where a mistake costs $6 million and a deposition. If proof of market is worth money in commercial, it's worth money in Boulder. Commission saved is not proceeds gained. Part of that $43,000 might be a discount the seller handed over. Look at the gap between comp set selection and true price discovery. Picking the set is a formula. Finding what any of them traded for is not. About a dozen states keep prices off the record, Texas included. That wall is lower than it looks. CoStar has reverse-engineered those states for twenty years with people on phones. Mortgages get recorded. Models infer from partial signals. Reconstruction keeps getting cheaper, so the winner is whoever owns the pipe. Probably CoStar, not a startup. Somebody will say this movie already played. Purplebricks burned $142 million in the US and quit, but under MLS rules that changed in 2024, so it proves less than it looks. Better precedent: May Day, 1975. The SEC killed fixed stock commissions. They collapsed. The industry didn't. It split into cheap execution and expensive advice, and Schwab got built in the gap. Fees split the same way. Small coordination-heavy deals drift to flat fee and software. Complicated ones hold, because the scarce thing is a human willing to be accountable. Relationships win. They keep winning until the generation holding them retires. NAR's median member is 57. The over-60 share went from 35% to 44% in one year. Under 40 fell from 17% to 11%. That's residential. Commercial skews older. What do you think?