Leopold is just Getting Started
3 Min Read | No AI used
I began my career in 2006 on the graduate program of the Bank of Ireland. It was the heyday for Irish property and for Irish banks. They were expanding into international markets, the loan books were bloated yet growing and at that time it was still unheard of to question the boom that had been occurring in property and banks. Our first week was an initiation week and during lunch one day, one of my new colleagues was preaching the merits of Bank of Ireland stock to the table. This thing was a rocketship and we all needed to get on it he evangelised. He was so convinced of the prospects of Bank of Ireland he was going to borrow, pretty heavily, to purchase the stock. Many on the table agreed with him. I was too sceptical and too broke to give it any more thought after lunch so I didn’t think much more about it. We all know how that story ended and my grad program colleague and several others may have had a serious lesson from it.
This is similar but different to the Situational Awareness situation.
Leopold Aschenbrenner, if your not aware is a wunderkind who graduated Columbia at 19, joined OpenAI, was let go (according to reports) before penning a 167 page essay entitled “Situational Awareness” outlining his diehard belief that the next decade of advances in AI would herald a new age of advanced technology and productivity for humanity. His essay went viral and attracted huge investment from Silicon Valley and Wall Street. Without any investing experience, Leopold started a 2 and 20 hedge fund, launched with $225m, leveraged at 4x - 5x and finished out the year with assets of $20bn, exposure of $45bn and returns of c.439%. That was just before markets moved against him in July, “crashing” his returns to 80% for the year, and causing a liquidity crisis forcing him to sell his public equity book to cover margin. His remaining investments are privately held. He issued a new culpa letter and vowed to do better.
Many in the industry cheered his comeuppance and issued I told you so’s citing the obvious foolishness of using so much leverage. They've pointed at Aschenbrenner as naive and out of his depth and the Silicon Valley Investors to be even more so. But that's just not right. The fact is, Aschenbrenner has done more in his first two years than most do in a career.
At 24 he convinced people to invest billions in him and his vision and for the most part he grew that investment at an outstanding pace. He got caught out by margin calls but he made the right decision for his investors to sell his public equity book to Citadel and he survived. He did everything a fund manager should have done.
You might say, how could he be so blind to the risks. But that was the strategy he was selling. Those investing in him knew what he was doing and undoubtedly accounted for such an outcome. They transferred the risk of conviction and leverage away from their book and onto his. This is why it is so different to my Bank of Ireland story above. Situational Awareness investors could reap the upside gains but only lose what they put in. An individual or fund who levers up can lose more than they have but their investors are not exposed. It’s most likely they allocated an appropriate amount from their own exposures to such a fund, 1% - 2% and noted it’s a highly volatile allocation. Investors that did that are not going to be concerned with what happened, they’ll have lost a chunk of their investment for now but it will still form part of their investment strategy. They will look at this and will think that the manager of their AI exposure demonstrated professionalism in the face of a crisis and has gained incredible experience.
There is another lesson here for us. The whole episode was a great case study in the power of conviction.
Aswath Damodaran just wrote an excellent paper on Situational Awareness and I recommend you read it. In it, he dedicates a lot of the paper to discussing conviction in investing. It's a deeply fascinating topic. Investors must have high conviction but they can never really have complete conviction because the future always carries uncertainty. An investor must know where conviction ends and doubt begins and the potential value and cost of each respectively.
Complete conviction is irrational and naive for the most part but conversely, most great fortunes are made through complete conviction. Famously, Bill Gates would have made more if he had left his shares in Microsoft MSFT 0.00%↑ instead of diversifying into a range of investments. There is also a powerful additional benefit that comes from complete conviction - the marketing edge. People who say they are all in on something when others agree with them, can do very well, whether the thesis ultimately holds or not. The last 7 or so years have had many great examples of this. History is littered with operators who preach a complete fate in a certain idea being rewarded by the masses. It is human nature. When we see someone go all in on something, and that person appears intelligent, is heavily validated by previous achievements or by high profile supporters then we reasonably assume they're on to something and it makes sense to back them.
And in Silicon Valley, complete conviction and a failed first-run are the cornerstones of going on to greater things. Many of the great investors stumbled early and most of the great tech entrepreneurs did - in silicon valley it's expected.
If Leopold could get $20bn AUM on $45bn notional at 24, with zero experience and generate incredible returns with it (before losing a lot of it through hubris), imagine what he'll do in the next 5, 10, or 15 years having learned all he’s learned.