We spoke a lot during 2022 about the idea that while the downdraft was similar in magnitude to prior bear markets, that it was unique because blockchain was not facing an existential crisis. Most of the price action was driven by leverage and the headlines caused by bad actors. Thats why it is not surprising that we see the same projects coming back. They were not down because they were not good projects, they were down in sympathy with the overall market. Solana is a particularly good example.

THREE FLAVORS OF BLOCKCHAIN If you abstract out to just the three main flavors of blockchain: Bitcoin, Ethereum, and all other projects combined you can see the cycles. The first thing to note is once Ethereum became established in 2017-18, it has had a fairly steady share. The only major pullback was during the 2020-21 bull market when competing layer 1 hyper-scalable blockchains like Solana and Avalanche gained meaningful market share. Its Bitcoin and the other things that gyrate. One of the earliest non-bitcoin tokens and Panteras first venture investment Ripple and their XRP token exploded to a remarkable 27% of the entire market on May 17, 2017. The graph below depicts the three main things in blockchain over the recent extremes in Bitcoin market share since 2016. When Bitcoin and Ethereum were at their historical low market share in January 2018 the non-Bitcoin+Ethereum share was 55% of the market.

AREAS OF INTEREST IN 2024 The pace in which innovation occurs in the blockchain tech ecosystem produces new niches and verticals cycle-over-cycle. The research we do to expand our theses helps us stay ahead of the curve, but also helps us maximize our coverage when it comes to sourcing and investing in deals. Over the next couple letters, well share our theses on areas we are actively looking at. AIxWEB3 By Chia Jeng Yang, Principal, and Caroline Cahilly, Intern

AI: Melding Human & Computer Intelligence The outputs generated by AI models, like large language models (LLMs), should be a result of optimal interaction between human and computer brains, data, and incentivization systems. The ability to speak in natural languages is what makes LLMs exciting, as humans and AI can use the same language to expound on complex processes. This is a big step toward a future of coordinated systems that bring humans in the loop. To make this collaboration even better, we still need to develop strong human-computer frameworks, mechanics, and tooling, which would encourage AI systems to think more effectively, produce more useful answers, and achieve optimal results.