Subscribe to stay ahead of technology trends. Never miss future editions.
Reid Hoffman, co-founder of LinkedIn, joins NEW ECONOMIES to explain why we’re still underutilizing AI, why agent management — not individual contribution — becomes the default mode of work within three years, and why the next decade of company-building rewards taste and judgment over raw skill: once AI can draw, code, and write as well as any specialist, the only thing left to own is the specific, unrepeatable bar for what excellence looks like.
We also cover why AI carries a positive perception across Asia, the Middle East, and Latin America while triggering data-center bans and slowdown politics in the US and Europe — and why 91% of non-Chinese AI market cap still sits within 30 miles of Silicon Valley. We get into the founding story behind Manas AI — Reid’s drug-discovery venture, plus whether Silicon Valley is still the best place to start a company.
We close on the lessons Reid has learned from Microsoft CEO Satya Nadella and the technology trends Reid is most excited about next.
Watch or listen now on YouTube, Apple Podcasts, and Spotify
Download the transcript
Timestamps
(0:00) Meet Reid Hoffman
(1:36) Why Most People Are Under Utilizing AI
(5:28) How Should We Be Using AI?
(7:17) Why Agent Managers Are Next
(8:47) Are We Taking AI Seriously Enough?
(11:58) How Is Blitzscaling Different Today?
(15:00) Is Silicon Valley Still The Best Region?
(16:27) Should Startups Partner with AI Models?
(21:37) Personalized AI & Taste
(24:28) Is Distribution The New Moat for Startups?
(26:19) Does AI Have a Negative PR Problem?
(28:48) What Would Reid Coach Governments?
(33:20) It’s Not Just Anthropic or OpenAI
(36:36) The IPO Gold Rush
(38:32) Sam vs Dario Characteristics
(42:30) Lessons from Satya Nadella
(47:30) Ollie Joins As Reid’s Chief of Staff
(49:39) Reid’s Trends
(55:27) Fire Round
Our notes from this conversation
The next career move isn’t a skill — it’s a headcount.
Individual contributors are giving way to agent managers, each overseeing anywhere from a handful to tens of thousands of agents. Reid’s timeline: broad-based adoption within three years, starting with coding and spreading into every “digital closing loop.”
Non-US, non-Europe just inherited the industrial revolution’s rematch.
AI carries a negative PR perception in the US and Europe (job-loss anxiety, data-center bans) and a positive one everywhere else — Asia, the Middle East, Latin America, Africa. Reid calls it the reverse revenge of the industrial revolution: the regions that missed the first wave may lead the cognitive one.
Distribution didn’t die — it became the AI amplifier.
Moats aren’t dead, but they’ve changed shape. Existing companies with distribution aren’t finished; they’re only finished if they refuse to plug AI into what they already own. Refuse, and “the dinosaur decay has started.”
Wrapper businesses are living on borrowed time.
If a competitor can point Claude Code or Codex at your product and rebuild it for the token cost, you were never defensible. The businesses that survive have a theory of the game: network effects, enterprise integration, or data loops the model can’t just absorb as a feature.
Hand skill is depreciating. Taste is the new scarce resource.
AI coordination doesn’t need better hand-eye coordination to draw well — so that skill’s value drops. What rises: judgment about what excellence actually looks like, in code, writing, and product. Reid expects taste and context-awareness to matter for decades, not quarters.
The market isn’t consolidating to three companies — it’s expanding to ten or fifteen.
Reid’s eight-year-old antitrust argument held: from five-to-seven heading toward ten-to-fifteen frontier and adjacent players, not three. That’s good for entrepreneurs — more potential acquirers, more room to build something valuable without having to out-build OpenAI or Anthropic.
Blitzscaling didn’t disappear — it moved from headcount to compute.
The old version was maximum organizational scale under uncertainty. The new version: build something in two weeks, throw it away, rebuild — at higher velocity and higher risk — while a much smaller human team manages a much larger agent workforce.
Trust gets built by value, not by messaging.
Reid’s answer to AI skepticism isn’t a better narrative — it’s a free medical assistant, a lease-reading legal agent, a tutor on every phone. “It isn’t by a story of words... how you get trust is by seeing value.”
Links
Follow Ollie on X - https://x.com/ollieforsyth
Follow Reid on X - https://x.com/reidhoffman
Subscribe to Reid’s Podcast, Masters of Scale - https://mastersofscale.com/
Discover Reid’s other projects: https://beacons.ai/reidhoffman
Reid’s Books - Blitzscaling & Superagency
