In this episode, Lukas opens up about the real story behind W&B’s $1.7B acquisition, what it was like inside early OpenAI, why most founders move far too slowly, and the surprising truth that you’re almost always earlier than you think.
Lukas breaks down how W&B became the default toolset for the world’s best ML engineers, the strategy behind seeding researchers, why platforms often fail, and the singular advantage of being “in a rush”.
He also shares lessons from Travis Kalanick, what founders consistently get wrong, and how becoming a parent changed his leadership.If you care about AI, startups, product craftsmanship or developer tools, this episode is a masterclass.
Don’t forget to subscribe 🔔
Chapters:
00:00 Intro – The gold rush of AI
01:18 From CrowdFlower to OpenAI
02:54 The insight: data beats algorithms
04:47 Joining OpenAI as an unpaid intern
07:57 The origin story of Weights & Biases
09:03 Getting the first 100 users
11:00 “You think you're late, but you're probably early”
12:56 Fundraising: CrowdFlower vs W&B
16:45 Building for builders, not platforms
17:22 Why seeding researchers made W&B win
22:10 Advice most founders get wrong
37:15 Breaking through ‘paper walls’
44:56 What it was all for
⭐ Listener Suggestion – CoreWeave - https://www.coreweave.com/
Why subscribe?If you build, code, invest or care about the future of AI, startups or developer tools, this channel is for you.
🔔 Tap the bell so you never miss new episodes.
#AI #startups #wandb #founders #machinelearning #openai