Big Technology Podcast AGI or Bust, OpenAI’s $1 Trillion Gamble, Apple’s Next CEO?
311 snips
Oct 10, 2025 
Ranjan Roy, a tech business writer at Margins, returns with insights on AI and investment dynamics. They discuss the critical need for achieving AGI to justify hefty investments. Ranjan argues that the current focus on model improvement may delay this goal. They also explore OpenAI's ambitious $1 trillion compute deals and the potential economic risks associated with rising debt in AI funding. Additionally, they ponder who might succeed Tim Cook as Apple's CEO and critique OpenAI’s Sora regarding copyright issues.
AI Snips
Chapters
Transcript
Episode notes
CapEx Bets Depend On AGI Reality
- Massive new AI capex only makes economic sense if AGI is achievable at scale.
- Ranjan Roy and Alex argue current evidence doesn't support that assumption.
Research Progress Vs. Infrastructure Buildout
- Model progress is tapering and experts are walking back AGI timelines.
- That divergence makes heavy infrastructure spending look disconnected from research reality.
Three Paths To AI Progress
- There are three paths: scale-heavy compute, software algorithmic efficiency, and product/orchestration.
- Ranjan emphasizes compute-efficient algorithms and product workflows as viable alternatives to brute force scaling.