Practical AI The mythos of Mythos and Allbirds takes flight to the neocloud
91 snips
Apr 23, 2026 A frontier model sparks talk about cybersecurity fears, hype cycles, and the growing push for AI governance. Then things get weird with a sneaker company trying to become a neocloud player. The conversation also hits tokenmaxxing, pricey coding habits, AI chat logs turning into legal evidence, and the market for private AI messaging.
AI Snips
Chapters
Transcript
Episode notes
Neocloud Means GPU First AI Infrastructure
- Daniel Whitenack describes neocloud as AI-native cloud infrastructure built for training and inference rather than general-purpose apps, storage, and databases.
- He contrasts hyperscalers with GPU-first providers like CoreWeave, Together AI, and Lambda Labs that optimize around scarce GPUs and multi-node AI workloads.
Allbirds Has Cash But Not An AI Moat
- Daniel Whitenack argues Allbirds brings mainly cash and a public-company shell, not obvious AI infrastructure expertise, making the pivot look fragile.
- He notes roughly $50 million is tiny beside the billions flowing into AI data centers, so the move may be more foothold than competitive threat.
Embedded AI May Outshine Neocloud Over Time
- Chris Benson expects both centralized AI data centers and embedded physical AI to grow, but thinks far-edge deployments still have huge untapped room.
- He suspects neocloud gets crowded and less profitable over time, while specialized AI inside devices, cars, phones, and factories may offer stronger differentiation.