When your AI has to make decisions in the real world, the data you don’t have can hurt people.
DiffuseDrive CEO Balint Pasztor joins IT Visionaries to unpack the $124B data scarcity problem holding back autonomous systems — and how synthetic data (done right) can compress years of data collection into hours.
We dig into: why edge cases matter (and how to safely create them), closing the sim-to-real gap with diffusion models, preventing model drift, and building a data engine for physical AI across defense, robotics, and industrial automation. If you care about self-driving cars, drones, QA on the factory floor—or just shipping AI that survives the messiness of reality—this one’s for you.
Key Moments:
00:00 Introduction to Autonomous Driving Challenges
00:26 Meet Balint Pasztor and Diffuse Drive
01:14 The Importance of Synthetic Data
06:39 The Role of Synthetic Data in AI Training
18:07 Understanding Diffusion Models
23:28 Challenges in Real-World Data Collection
26:46 Three Steps to Improve AI Performance
32:13 Overcoming Non-Obvious Data Challenges
36:00 Balancing Data Quantity and Quality
40:28 Future of Autonomous Systems and Physical AI
--
This episode of IT Visionaries is brought to you by Meter - the company building better networks. Businesses today are frustrated with outdated providers, rigid pricing, and fragmented tools. Meter changes that with a single integrated solution that covers everything wired, wireless, and even cellular networking. They design the hardware, write the firmware, build the software, and manage it all so your team doesn't have to.
That means you get fast, secure, and scalable connectivity without the complexity of juggling multiple providers. Thanks to meter for sponsoring. Go to meter.com/itv to book a demo.
---
Mission.org is a media studio producing content alongside world-class clients. Learn more at
mission.org
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
Informações
- Podcast
- Canal
- FrequênciaQuinzenal
- Publicado18 de setembro de 2025 às 13:30 UTC
- Duração45min
- Episódio500
- ClassificaçãoLivre