Episode summary
Databricks CEO: AGI Is Already Here, Why Companies Are Behind, and the Superintelligence Debate
Summary
Ali Ghodsi says Databricks was born out of frustration that the industry was not ready for the company’s early ideas about combining data processing and AI. He traces the company’s roots to UC Berkeley’s AMPLab, early recommendation systems like Netflix, and a set of contrarian bets on cloud, open source, and AI that he says proved decisive. Databricks’ product story is framed as an enterprise infrastructure play: help organizations use their own data, govern it, secure it, and make AI useful inside messy real-world businesses rather than just as a chatbot or coding assistant.
A central argument is that enterprises are far behind the frontier labs because they face four practical barriers: getting enough organizational context, securing access to data, controlling AI costs, and preserving ownership of internal IP. Ghodsi says many companies are still using AI in rudimentary ways, while Databricks is building an “ontology” and agentic workflow layer to analyze org charts, meetings, surveys, and customer feedback. He presents this as “management moneyball,” where AI can surface blockers, measure culture, and speed decisions across large organizations.
On the superintelligence debate, Ghodsi argues that AGI is already here by an older definition, but that recursive self-improvement is not. He says the real issue is applicability in enterprises, not a runaway godlike system. He also addresses Databricks’ IPO timing, saying the company will eventually go public but prefers to stay private during an AI transition that he thinks is too volatile and distracting.
Key takeaways
- Ali Ghodsi says Databricks was built from early Berkeley work on machine learning and recommender systems, after the industry initially showed “zero interest in AI.”
- He argues enterprises are mostly using AI as “a glorified Google search” or for coding, not as a full operating layer for the business.
- Ghodsi says Databricks has to solve four enterprise blockers at once: context, security, cost, and IP ownership.
- He says AGI is already here by the definition his team used years ago, but superintelligence would require recursive self-improvement and a much tighter training loop.
- Ghodsi says Databricks will eventually go public, but staying private now avoids the stock-driven distraction of an AI market that is still highly volatile.
Notable quotes
“we have all the intelligence we need to completely transform the planet”
“They're just using AI in very rudimentary ways. They're just using it as a glorified Google search”
“It's not possible to have a thousand agents doing stuff realistically”
“Just because 98% of the code is being written by AI, we don't have RSI”
“my goal is to make database as successful as possible in the next decade or two”
Topics
- agi readiness
- enterprise ai gap
- agentic ai
- databricks ipo
- cyber risk
Companies and tickers mentioned
- Databricks
Watch the episode
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from All-In Podcast
-
Freedom vs. Safety: Axon CEO Rick Smith on the Future of Policing -
Midterm Blood Bath, $4T AI Sell Off, Data Center Freeze, What’s Next for Trump -
All-In Podcast REACTS to Democrat Sweep! -
Anthropic S-1 Shocker, Greatest Business Ever? Google’s $500B Wipeout, Midterm Predictions -
WHOOP CEO, Will Ahmed: 50B Hours of Health Data, a 4 Million Member Data Moat, & the FDA Fight -
AI Cyber Panic, Ken Griffin’s Bubble Warning, The Fall of France?
All-In Podcast, this episode and its artwork belong to All-In Podcast, LLC. Octagon is not affiliated with the show; this page is an independent summary.