Sep 5, 2026 · 33 min
AI’s Blender prowess tests the meaning of general intelligence
Astra Reactions, Jobs Print, Cybercab Launch | Diet TBPN
The episode connects a practical test of AI generality with labor-market uncertainty and Tesla’s ambitious robotaxi timeline.
- 1GPT-6 Astra’s Blender demonstrations are impressive, but prompting and specialized training complicate claims of AGI.
- 2The August jobs report points to modest growth while leaving open how quickly AI is changing employment.
- 3Tesla’s sub-$30,000, steering-wheel-free Cybercab promise faces questions about autonomy, remote assistance, and delivery timing.
Don't miss
The hosts make their own wager over whether Tesla can deliver a sub-$30,000, steering-wheel-free Cybercab by the end of 2026.
The brief
John Coogan and Jordi Hays open with GPT-6 Astra’s detailed Blender demonstrations, then ask how much of the result comes from prompting, manager agents, and subagents.
Blender may show powerful creative assistance without proving AGI: Astra’s performance could depend on training in a specific open-source environment rather than general software competence.
The harder test is transfer. The hosts question whether Astra can move from Blender into closed-source tools such as Cinema 4D and Adobe Premiere, where licensing and computer-use training matter.
The conversation then turns to 162,000 August jobs added, Dyson’s expensive design-led products, and whether technology-heavy sectors are diverging from the broader economy.
Tesla’s Cybercab promise—under $30,000, no steering wheel, and delivery by the end of 2026—sparks skepticism, a Marques Brownlee bet, and doubts about autonomous taxi readiness.
The episode closes by comparing Cybercab’s commercial opportunity with the Roadster and asking which Tesla vehicle will actually reach customers first.