From Fired CEO to Billion-Dollar Exit: How Lukas Biewald Turned Failure into the Future of AI
The Eric Ries Show
Why Large Language Models Work
Lukas explains how simple neural architectures, massive data, and compute let models encode surprising amounts of human knowledge.
Lukas Biewald has been shaping the future of AI for nearly two decades. In this episode of The Eric Ries Show, we explore how his early passion for the board game Go, his hard-won lessons as a founder, and his focus on alignment and culture helped him build Weights & Biases into a trusted toolmaker in machine learning.
In our conversation, we discuss:
• Lessons Lukas drew from Go about pattern recognition, persistence, and the power of computation
• The hard-won lessons Lukas took from CrowdFlower about timing, focus, and staying in the wrong role too long
• How Lukas created his own unpaid internship at OpenAI to sharpen and expand his technical skills
• The pain point that inspired Lukas and his cofounders to build Weights & Biases
• Practical ways Lukas kept his board aligned
• Why caring deeply about your product and users beats clever hacks and pitch decks every time
• How W&B’s ‘Year in Review’ showed that delight can be a moat in enterprise software
• Why the CoreWeave deal feels fundamentally different for Lukas than the sale of CrowdFlower
• And much more!
—
Where to find Lukas Biewald:
• LinkedIn: https://www.linkedin.com/in/lbiewald/
• X: https://x.com/l2k
• Website: https://lukasbiewald.com/
—
Where to find Eric:
• Newsletter:https://ericries.carrd.co/
• Podcast:https://ericriesshow.com/
—
In This Episode We Cover:
(00:00) Intro
(02:05) How Go shaped Lukas’s view of AI
(08:32) How large language models actually work, and why they still feel mysterious
(12:58) Why AI’s abilities force us to rethink what intelligence means
(14:55) Key lessons from Lukas’s time at CrowdFlower
(23:59) What Lukas gained from teaching AI and interning at OpenAI
(28:28) Weights & Biases initial product
(30:06) How Lukas and his cofounders turned a pain point into a company
(32:18) What Lukas learned about the reality of raising money
(34:30) The non-negotiables in startup building
(37:00) Why genuine care outperforms every other growth strategy
(39:11) How caring about AI research and delighting users created W&B’s moat
(43:47) How a contrarian culture helped W&B preserve its ethos as it scaled
(47:00) The challenges of being a middle manager and how to identify the right bold ideas
(55:07) How Lukas knew W&B had achieved real traction
(56:36) How Lukas built alignment and transparency with his board
(59:05) Why the acquisition of W&B feels less stressful than Lukas’s previous company
(1:03:09) Why Lukas believes skepticism about AI is premature
—
You can find episode references at https://www.ericriesshow.com/
—
Production and marketing by Pen Name.
Eric may be an investor in the companies discussed.
The AI-powered Podcast Player
