Some links 03/2025
Ben Evan’s “State of AI” presentation is a must read
Despite lackluster performance, David Einhorn’s Q4 letter is worth reading
Annual letter time: Partners Fund 2024 letter, Rubicon Stockpicker 2024 letter, Compound Interest 2024 letter
An interesting new Subtack called “Building Things that last” has just launched with a few great posts
For me, the story of the week: Chinese LLM Deepseek seems to be (almost) on par with the OpenAI &Co at a much lower cost. Maybe it’s just Chinese propaganda, maybe it’s not.
A very detailed Nvidia short case including an (expert) assessment of Deepseek
The “Behind the Balance Sheet” Substack and Podcast are both worth the time
value and opportunity 
Any decent FundManager/ManCo would take as benchmark an index (or composite) comparable to the market they are invested in, which is typ. defined in the SAA/TAA docs. Those that do otherwise are cheaters. Manipulators. You can find any excuse, there is ultimately NONE !
When I read Partners Fund (and a few others btw), they just do that: they invest systematically out of the DAX market, yet they keep it as a reference for benchmarking. It’s a shame that they do that and nobody calls it out.
One of the reasons that they use the Dax is that you don’t have to pay for it. If you would use MSCI etc., a fund easly needs to pay a large amount of money to the index providers which in turn will be paid by the investors.
All the investors in the fund understand what these guys are doing. Perfromance fee is charged like the old Buffet partnerships on top of an annual, fixed CAGR. They are not cheating/manipulating anyone. Sorry for you if you think so.
Mathias here – the advisor of Partners Fund. This isn’t an excuse, just an explanation:
You’re right – the DAX isn’t the perfect index. But to be honest, I don’t know which one would be better for most of my Partners. Since we started with it, I believe we should stick with it. Over time, the fund has invested across a broad spectrum – Megacaps in the US, oil and gas, Nanocaps in the UK and concentrated positions – shifting its universe along the way. And this will continue over the next decades.
I’ve written about this several times (2017/2023), and I genuinely believe my Partners should judge my performance against any reasonable alternative they would use for their own investing. But consistency matters—it’s not about comparing us to the S&P 500 today, the MDAX tomorrow, and Biotech the day after.
Most fund investors are German natives, and the DAX is a familiar benchmark – it’s on TV, in newspapers, and part of everyday conversations. It’s also fairly diversified and, unlike the S&P500, a performance index.
As MMI points out, using MSCI World as a benchmark – especially in public materials – comes with significant costs. These costs would also be charged against the fund, and in my view, that is completely unnecessary.
Best
Mathias
The problem is we are nearly two years after the Google “We Have No Moat, And Neither Does OpenAI” and more of a month after Ilya’s presentation openly stating that scaling won’t work anymore (and before that strong hints about this being the case). How much longer does this go on? Will switching to a robotics hype work?