The $8.6B Self-Driving AI Backed by Nvidia and Uber | Alex Kendall, Wayve

Weights & Biases · Beginner ·📰 AI News & Updates ·1w ago
"Every vehicle is capable of driverless operation. That's clearly the steady state of where we're going." Wayve started in a rented house in Cambridge with $1.5M, a car in the garage, and an aim to integrate end-to-end AI into driving. A decade later it's driven across 506 cities without a single HD map and is worth over $8.6 billion. In this episode, CEO Alex Kendall joins Lukas Biewald to talk about how he built the AI driver Uber, Nvidia, Mercedes, and Nissan all backed, and why putting self-driving AI into 100 million cars a year is a far bigger bet than 10,000 robotaxis. Waymo and Tesla both come up. He doesn't shy away. Connect with us here: Alex Kendall: https://www.linkedin.com/in/alexgkendall Lukas Biewald: https://linkedin.com/in/lbiewald Wayve: https://www.linkedin.com/company/wayve-technologies Weights and Biases: https://www.linkedin.com/company/wandb (00:00) Trailer (01:01) Introduction (01:42) The story of Wayve (02:44) End-to-end deep learning for autonomous vehicles (03:52) Driving zero-shot in 500+ cities (04:22) Growing up on a farm in New Zealand (05:11) Why the company is called Wayve (06:10) Building the first end-to-end prototype (07:28) First lane following with 10 interventions (09:09) Scaling beyond on-policy learning (10:37) Why London was a great training ground (11:15) The autonomous vehicle market then and now (12:46) Mass market vehicles versus expensive retrofits (14:15) Algorithm improvements versus logistics (16:15) Plugging perception into an LLM (17:07) Adding language to the driving model (18:04) Learning social cues like flashing lights (18:42) Personalization and driving styles (19:23) Waymo's impact on San Francisco (20:30) Human-like driving behavior (20:49) Why Wayve doesn't build its own cars (22:53) Stack ranking business opportunities (24:00) Target geographies for launch (25:08) The AI 500 road show results (26:44) Is this superhuman performance? (27:56) Scenarios that still challenge the model (29:20) The trolley
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Related AI Lessons

\The Vibe Coding Reckoning: What Leaders Need to Know Before It’s Too Late
Learn about the Vibe Coding Reckoning and its implications for leaders in the AI industry, including the importance of clarity and governance in enterprise AI
Medium · AI
Microsoft Just Triggered China’s Rise as an AI Hegemon
Microsoft's cancellation of subscription-based AI services may lead to China's rise as an AI hegemon, shifting global geopolitics
Medium · AI
South Africa used AI to write its AI policy. The citations were fake.
South Africa's AI policy was generated using AI, but with fake citations, highlighting the need for responsible AI use in governance
The Next Web AI
OpenAI called the growth report clickbait. The market disagreed by tens of billions of dollars.
OpenAI's growth report sparked market disagreement, despite the company calling it clickbait, resulting in a tens of billions of dollars discrepancy
The Next Web AI
Up next
The Metrics That Actually Matter
AssemblyAI
Watch →