An interview with Skype co-founder Jaan Tallinn on investing $100M+ in 100+ AI startups and why he thinks he failed to steer AI development toward human safety
Jaan Tallinn used the fortune he made selling Skype in 2009 to invest in AI companies like Anthropic and DeepMind … Tweets: @carnage4life , @semafor , @semaforben , @semafor , and @reedalbergotti Tweets: Dare Obasanjo / @carnage4life : The problem with AI safety discourse is people rarely list concrete concerns. This article talks about threats to humanity from AI but actually lists none. At least when people said social media was bad for democracy there were concrete examples. https://www.semafor.com/... @semafor : 🟡 NEW: The co-founder of Skype used his fortune to invest in AI — not because he was excited about the technology — but because he believed it was a threat and hoped he could steer it in a safe direction. Now, he's admitting defeat. https://www.semafor.com/... Ben Smith / @semaforben : Another glimpse into the pre-history of AI from @ReedAlbergotti https://www.semafor.com/... @semafor : “My philosophy has been that I want to displace money that doesn't care,” he told @ReedAlbergotti. “Plan A failed. There is a dissonance between privately being concerned and then publicly trying to avoid any steps that would address the issue.” https://www.semafor.com/... Reed Albergotti / @reedalbergotti : Spoke with Jaan Tallinn, the founder of @Skype for this article with some really interesting new information on the behind-the-scenes battles over AI safety. https://www.semafor.com/...
Context & Ripple Effects
Jaan Tallinn's bet was that capital buys steering: after selling Skype in 2009 he put more than $100M across 100+ AI startups, including Anthropic and DeepMind, explicitly to pull development toward human safety. The Semafor interview is him conceding the premise didn't hold — the money went in, the safety influence did not materialize.
The admission lands mid-fight: Forbes has since framed the same cohort of internet-era billionaires as battling over whether AI's future is concentrated safety or unfettered advancement, while Tallinn himself has begun doubting the effective-altruism philosophy underpinning his approach — against Vinod Khosla's camp arguing AI will free humanity from the need to work.
First-order effects
- Tallinn's concession weakens the core pitch behind 'safety-motivated' checks into Anthropic, DeepMind, and peers: investors who claimed proximity would buy alignment influence now have their most prominent example admitting it bought neither.
- Dare Obasanjo's critique that the interview names no concrete threats sharpens the immediate discourse problem — Tallinn's failure story gives skeptics a template for demanding specifics from safety advocates.
Second-order effects
- Other billionaire backers staking their legitimacy on steering AI — the group Forbes profiles as fighting over concentrated safety versus advancement — face pressure to show measurable influence or concede they are simply early financial returns players.
- Labs like Anthropic and DeepMind learn they can absorb mission-driven capital without ceding control, which shifts bargaining power further toward founders when raising future rounds.
Third-order effects
- If investment-as-steering keeps failing, AI-safety effort migrates from capital channels toward regulation and public-pressure campaigns — the Max Tegmark/Future of Life Institute lane rather than the cap-table lane.
- The pattern feeds a broader reassessment of effective altruism's role in AI governance, with one of its most prominent funders already questioning running companies on the philosophy at all.
The trend: Internet-era fortunes are being deployed to shape AI's direction, and the Tallinn episode is evidence that investor influence over frontier labs is proving far weaker than the safety-motivated-capital thesis assumed.