Insights on Generative AI for Real Estate Professionals
Expert insights, updates, and practical AI guidance for the real estate industry.
A year ago, Niko would have told you Claude was pulling ahead. Not anymore. He unpacks why the smartest model increasingly isn't the one businesses actually want to use, and what's replacing "which model is best" as the question that matters.
The EU just softened its AI literacy rule. Thomas argues that changes less than it looks, and asks the harder question underneath it: not what the law requires, but what your team actually needs to know, role by role, to catch the mistakes AI won't flag on its own.
At some point this year, a board member will ask which figures in a paper were produced by AI and who checked them, and most organisations won't have a confident answer. Thomas sets out why that gap is a governance problem, not a technology one, and gives boards three practical questions to close it.
A cheap open-weight model can outperform a frontier one, and a frontier model can underperform a cheap one, depending on nothing more than how each is set up. Monika explains the harness: the unglamorous layer around every AI tool that actually decides how well it performs, what it costs, and how efficiently it runs.
Another month, another flood of new models, more drama about who's winning. Niko argues the real story is quieter: the models aren't getting smarter so much as more persistent, and for most real estate work, that's already more than enough. This month's newsletter looks at why the model race matters less than you think, and what actually separates firms getting value from AI from those just buying access to it.
Turning the temperature to zero does not stop hallucinations. It just makes them consistent. Here is what is actually happening inside your model.
Courses for individuals