OpenAIWhat racing reveals about working with AI — the OpenAI Podcast Ep. 22
At a glance
WHAT IT’S REALLY ABOUT
Racing shows how AI wins through organization, speed, and teamwork
- OpenAI researcher Joyce Ruffell describes a collaboration with Chip Ganassi Racing focused on making high-bandwidth time-series racing data accessible, comparable, and actionable faster than traditional workflows allow.
- Chase Holden explains how RaceTek uses ChatGPT and Codex to build bespoke “racing intelligence” tools that centralize information and help smaller or mid-level teams close the gap with better-resourced competitors.
- Both guests emphasize that a major AI unlock in racing is bridging “hard” telemetry with “soft” human inputs like driver feel, engineer notes, and strategy thoughts, turning them into usable, retrievable data.
- A recurring theme is that education—demystifying how models work and how to prompt effectively—is often the most impactful early deliverable because it changes how teams think and work with AI.
- They argue AI won’t replace the human-centric nature of racing; instead, advantage shifts to humans with the best taste, instincts, creativity, and workflows once AI becomes broadly available.
IDEAS WORTH REMEMBERING
5 ideasIn racing, speed of insight is a competitive advantage.
Teams win marginal gains by shortening the loop from data → hypothesis → test; AI helps retrieve answers faster so practice and qualifying become higher-value experiments.
Education is an early, high-leverage AI deliverable.
Ruffell notes that demystifying models and teaching prompt/input quality helps engineers trust the tool and use it purposefully rather than treating it like magic.
The biggest AI value is unifying ‘hard’ and ‘soft’ data.
Telemetry is already analyzable with traditional tools, but AI shines at organizing and correlating qualitative notes (driver feedback, engineer journaling) with numeric signals for downstream decisions.
Excel is both the operating system and the bottleneck.
Motorsports workflows often live in complex, sprawling spreadsheets; moving data into more AI-native formats (or building tools that work with that reality) reduces friction and errors.
AI narrows resource gaps but doesn’t replace engineers.
Holden frames AI tools as “extra engineers” that catch what falls through the cracks; they won’t replicate a full engineering staff, but can elevate smaller teams’ decision quality.
WORDS WORTH SAVING
5 quotesThe man and the machine, how do you get those to work together as seamlessly as possible?
— Joyce Ruffell
I believe we are in what's called the data wars right now in racing, maybe on the IndyCar side and the NASCAR side.
— Chase Holden
One of the most impactful things that we've delivered to the team is, frankly, is education.
— Joyce Ruffell
When they start working with AI, it's not like, "Lead me. Hey, AI, lead, lead me in this direction." It's like, "No, I need AI to deliver me this," and they have a very clear idea of what it needs to be.
— Joyce Ruffell
That's when it all comes down to who has been doing it before that arrived the longest, who has the best instinct, who has the best taste, who can communicate what really needs to be done in that moment in the most effective way.
— Chase Holden
High quality AI-generated summary created from speaker-labeled transcript.
