A concept trailer, my second for the a16z podcast after the Machine Age Fund trailer. 59 seconds, cut from a 48-minute episode.

a16z's Jen Kha and David George sat down with Accolade Partners' Aram Verdiyan to talk about how AI is changing the power law of technology investing. It's a sharp conversation, and very little of it is soundbite-shaped: portfolio construction, compute spend, market share curves.

On the Machine Age trailer, one scene showed generic architecture while the speakers were talking about data center architecture. With an audience this technical, close enough reads as wrong. So the rule for this one was simple: every visual illustrates the exact line being spoken.

I went through the full transcript and pulled 12 lines from across the 48 minutes, then put them in a new order. The cold open is the strongest number in the episode: AI reached $100 billion in revenue in four years, something that took SaaS 15. Then the mechanism (you can throw dollars at compute), the stakes (second place is playing for scraps), the scale (30 trillion in GDP), and the gap between the median US company spending $12 per employee per month on AI and the top 1% spending $7,000. Every qualifier and unit stays exactly as it was said.

It closes on "far too limiting", the line that says calling AI the next evolution of software undersells it, and leaves the argument for the episode.

Captions run word by word in Helvetica Neue, white with red emphasis, so the trailer reads with sound off. When the guest says SaaS took 15 years, an orbit graphic cycles through the software giants that took that long. When the list of sectors comes, each one wipes in as a panel on its word. $30 trillion goes full-size over the speaker. Robotics gets a retro machine running a robotics sim.

A 59-second trailer that turns a dense investing conversation into one argument, readable on mute, with every line traceable to its timestamp in the source episode.