AI for Business Owners

There are genuine, grounded observations about permitting timelines and the housing shortage, but the host monopolises airtime with long tangents (the 2003 Intel story, a tutorial on probability engines, his personal Slack automation) that crowd out substance and slow the pace of useful ideas considerably.

“it takes longer to get a permit to build a house than it does to actually build the house”

“what would normally take, you know, seven or eight months is being done in six weeks”

The central framing - AI as a useful assistant, not a replacement - is the single most repeated take in the entire AI discourse right now, and neither host nor guest advances it to a contrarian or first-principles conclusion; the permitting critique is real but widely observed in proptech circles.

“I think of AI right now in real estate as a really good assistant”

“it's not meant to replace all this stuff yet, but it is a damn good advisor”

Ari Rastegar is a credible, at-scale practitioner - an attorney running a multi-billion-dollar development pipeline across 38 cities and 13 states with genuine operator experience - but the conversation fails to extract deep expertise from him, leaving him largely reactive and agreeable.

“we've invested in 38 cities, 13 states, 7 different asset classes”

“we have billions of dollars in our development pipeline at this moment”

A handful of concrete figures (5 million homes short nationally, 500,000 in Texas, an 8-year rezoning in Dallas, 6-month demo permits in Austin, 6-week 3D print builds vs. 7-8 months traditional) provide a credible factual floor, but they are scattered across a mostly abstract conversation.

“we're 5 million houses short, um, of what we actually need in new single family homes”

“getting a demo permit in Austin is taking six months to tear down an old dilapidated building”

The host repeatedly hijacks the episode with multi-minute monologues - a 2003 Intel anecdote, an AI-as-probability-engine lecture, a detailed walkthrough of his personal Slack news bot - while the guest is reduced to short affirmations; there is almost no substantive follow-up or productive pushback on any claim.

“Sorry, I wasn't trying to lecture you”

“I appreciate the lecture. It was very helpful”