OpenAIWhat racing reveals about working with AI — the OpenAI Podcast Ep. 22
CHAPTERS
- 0:00 – 0:43
AI meets motorsports: Joyce Ruffell + Chip Ganassi Racing, and RaceTek’s origin story
Andrew Mayne sets up the episode’s two perspectives on AI in racing: an OpenAI researcher embedded with a top IndyCar organization and a startup founder building race intelligence software. The framing emphasizes how AI can impact everything from strategy and performance to day-to-day operations.
- •Episode premise: AI improving logistics, decision-making, and driver performance
- •Joyce Ruffell’s collaboration with Chip Ganassi Racing (IndyCar/Indy NXT)
- •Chase Holden building RaceTek with ChatGPT/Codex
- •Motorsports as a uniquely data-intensive domain
- 0:43 – 2:43
Joyce’s role at OpenAI and why racing data is uniquely hard
Joyce explains her OpenAI research focus on data sources, quality, pipelines, and efficiency, and how that connects to motorsports. She highlights the challenge of high-bandwidth time-series telemetry and the need for real-time conclusions at the track.
- •OpenAI research focus: data pipelines, quality, and efficiency
- •IndyCar/Indy NXT collaboration context and team history
- •Telemetry is high-bandwidth time-series data
- •Racing demands fast retrieval to test more hypotheses and iterate quickly
- 2:43 – 3:32
What RaceTek builds: “racing intelligence systems” for better weekend decisions
Chase describes RaceTek Systems as a provider of bespoke racing intelligence tooling. The goal is a centralized system that captures the small but important details teams miss and supports smarter trackside decisions.
- •Bespoke software for teams: intelligence tools and workflows
- •A unified “one place to live” system for racing knowledge
- •Focus on improving crew decision-making during race weekends
- •Capturing small details that are easy to overlook
- 3:32 – 7:46
Chase’s path: from Talladega fandom to AI-enabled entrepreneurship
Chase recounts how early NASCAR experiences hooked him on racing culture and competition. He then explains a winding career path through finance, radio/podcasting, and finally a pivot when ChatGPT made it feasible to build the tools he’d imagined.
- •Early obsession sparked by Talladega and racing community traditions
- •Career shift from finance to radio/podcasting and NASCAR community building
- •ChatGPT as an inflection point: turning ideas into working products
- •Desire to give back to racing—especially helping smaller teams
- 7:46 – 10:02
Joyce’s racing origin and how the Chip Ganassi collaboration began
Joyce explains how her interest in cars and racing intensified during COVID through track experiences and community. A conference trip to Indianapolis opened doors to team shop visits—rare access that helped her connect AI work to racing needs, including Chip Ganassi’s interest in AI.
- •OpenAI car culture and Joyce’s internal “social car channel”
- •Racing interest grew through attending events and following multiple series
- •Indy conference shop visits as a catalyst for practical collaboration
- •Chip Ganassi already exploring AI; timing around late 2024
- 10:02 – 13:20
The real advantage: organization speed, and AI as education + preparation
The conversation shifts to what creates an edge: organizing knowledge quickly and turning it into action. Joyce argues one of the biggest deliverables has been education—demystifying models and teaching prompt/input discipline—paired with heavy pre-race preparation for race-day readiness.
- •Competitive edge is often “who can organize best, fastest”
- •Joyce: most impactful contribution is education about how models work
- •Prompting framed as input discipline to get the right outputs
- •Work is largely pre-race to prepare timing stand and strategy teams
- 13:20 – 18:08
Driver communication and turning messy notes into usable insight
Chase and Joyce discuss how AI can help bridge the communication gap between engineers and drivers by converting dense reports into digestible guidance. Joyce broadens this into a key AI value: integrating “soft” qualitative feedback (driver feel, engineer notes) with hard telemetry to make it searchable and comparable.
- •AI can translate complex engineering info into driver-friendly summaries
- •Not every driver is an engineer; readability and speed matter
- •“Soft” data (notes, feelings) is real data but hard to organize traditionally
- •AI helps correlate qualitative notes with ECU/telemetry for downstream use
- 18:08 – 21:43
Leveling the field: AI for smaller teams, time savings across the whole organization
Chase explains how AI tools can approximate some benefits of having more engineers—helping smaller and mid-tier teams close gaps with top operations. Joyce adds that even strong teams are time-constrained; AI that reduces repetitive work can unlock higher-value questions, and it can help beyond performance—like logistics and parts tracking.
- •AI tools as “gap fillers” when teams can’t afford more engineers
- •Time is the universal currency in racing operations
- •Freeing individuals to pursue deprioritized questions that may yield an edge
- •Applications beyond track calls: logistics, parts tracking, error reduction
- 21:43 – 26:39
Spreadsheets, complexity, and the “data wars” in modern racing
They dig into how deeply racing workflows live in Excel—massive, formula-heavy, interconnected sheets that encode team knowledge. The chapter closes with the idea that racing is in “data wars,” where winning depends on capturing every signal—telemetry, notes, strategy thoughts—and breaking down team communication barriers.
- •Racing’s dependence on sprawling Excel-based workflows
- •Data conversion into more AI-native formats as an enabling investment
- •AI as a way to simplify and share understanding across roles
- •“Data wars”: every telemetry point, handwritten note, and idea matters
- 26:39 – 30:08
Getting started with AI and Codex: curiosity, better inputs, better workflows
Chase offers practical advice for newcomers: ignore stereotypes, explore tools directly, and learn how to communicate intent clearly. He also highlights the productivity jump from coding with chat to using Codex/IDE-style agents, while acknowledging practical constraints like rate limits.
- •Adopt a curious mindset; move past negative noise and stereotypes
- •Good results come from strong intent and well-crafted inputs
- •Learn via communities (Reddit/Discord) and sustained exploration
- •Codex/IDE agents speed iteration, with real-world constraints (rate limits)
- 30:08 – 31:36
How RaceTek landed its first customer: interest alignment + wearing multiple hats
Chase explains that the first deal came from committing fully—moving to North Carolina and networking in the racing ecosystem. A key ingredient was a partner/customer already interested in AI, combined with Chase’s media and community-building background to create multiple ways to add value.
- •“All-in” move to North Carolina to be close to the NASCAR ecosystem
- •First customer: SS Green Light BRK Racing with Garrett Smithley
- •Shared curiosity about AI lowered adoption friction
- •Media/podcasting experience helped create broader value beyond software
- 31:36 – 36:14
Racing, research, and human-centric competition (not robot replacement)
Joyce connects racing’s appeal—measurable outcomes and marginal gains—to the feel of ML research where progress is often a single number moving. The group argues that AI in motorsports is fundamentally human-centric and goal-driven, and while autonomous racing exists, it won’t replace the desire to watch humans compete.
- •Racing as a distillation of effort into time/position; parallels to ML metrics
- •Marginal gains mindset in elite competition
- •Teams demand precise, high-quality outputs—pushing model performance
- •Autonomous racing may coexist, but human competition remains central
- 36:14 – 39:26
When everyone has AI: advantage shifts to taste, prompting skill, and creativity
They explore a future where AI tools are widely accessible and the baseline capability rises. Chase predicts differentiation will come from experience, instinct, and the ability to operationalize workflows quickly; Joyce adds that cheaper experimentation makes creativity and idea generation a bigger competitive lever.
- •As AI becomes ubiquitous, differentiation returns to human judgment
- •Winning hinges on workflow design, communication, and fast execution
- •Smaller teams catching up could reshape competitive tiers
- •AI lowers experimentation costs, amplifying creativity and rapid prototyping
- 39:26 – 41:23
Beyond the racetrack: AI for operations, scheduling, and business building
Chase describes using AI across the company for project management, organization, and business development—not just race analytics. He uses agents to pressure-test ideas quickly, manage outreach and sponsorship efforts, and keep complex schedules on track, ending with a light wrap-up question about favorite racing movies.
- •AI as an always-on assistant for planning and prioritization
- •Coding agents for rapid prototyping and pressure-testing post-race ideas
- •Support for outreach, sponsorship, content opportunities, and scheduling
- •Wrap-up: favorite racing movies and closing thanks
