In the summer of 2024, former OpenAI researcher Leopold Aschenbrenner’s Situational Awareness essay became required reading across Silicon Valley. Its central thesis – that AGI is plausible by 2027, that trillions will funnel into the GPU and datacenters as a result, and that the underlying security stakes are vast – helped shape boardroom conversations, startup strategies, and even the multibillion-dollar investment fund bearing the same name.
Then, just days ago, came the reversal.
After building a leveraged book of roughly $45 billion (~4x its underlying equity) on an aggressive thesis, Aschenbrenner’s Situational Awareness fund was forced into a dramatic unwind following a sharp selloff in AI infrastructure stocks. The fund reportedly liquidated much of its public equity portfolio to Citadel after leveraged bets on AI names turned against it.
Today, we consider what happened under the surface with an op-ed from investor Alec Litowitz, who believes the lesson from this fallout has nothing to do with leverage but about sameness in thought. He argues the greatest systemic AI risk isn’t superintelligence at all – it’s artificial consensus – and further explores this idea in his forthcoming book, The Adaptability Quotient.
As Litowitz reminds us, plurality in thought is a great hedge against sameness, and independent judgment often means not building alone.
Which brings us to this week’s second announcement: our launch of Built Together: The Future Nexus Founder–Investor Awards, recognizing the best partnerships in venture.
We’re looking for both the early believers and those who stuck around for the long run. Those with shared conviction. Those who showed up for each other, day after day, even in moments of hardship. Those who brought their networks, their complementary strengths, and even their honest disagreement to the table – all in the name of category creation.
Because the strongest partnerships do not eliminate differences in judgment. They put those differences to work.
-The Editors
