donut⚡️

On Pearl Having been back in NYC the last week, a lot of GPs at various funds have been asking me about my view on Pearl. I'm not spending time on crypto at all, but wanted to share my take. This is not a shill (PoW is insanely fickle) and NFA but wanted to share my take nontheless. For my overarching views on PoW, folks can see here(cyber.fund/content/the-spect…), here (bitcoinmagazine.com/bigread/…), and here (forum.celestia.org/t/assessi…). While I’m not really into crypto these days, I am a big fan of what the Pearl team is doing. Full disclosure: I am an investor and holder of Pearl in a multitude of ways and have known the founding team for years, having co-invested with one of their founders several times, and met @weinsteinomri in the process of them building out the initial idea. Backdrop My sense for quite a while is most investors don’t get proof of work and hence they don’t get Pearl. They make comparisons to proof of stake chains, which are categorically different: proof of stake is easier to launch, easier to maintain, easier to design a marketplace for (e.g., priority cancels, etc), and yes, easier to capture (source: Richard Cantillon). In a phrase, PoS is much closer to a shared financial ledger than to a real cryptocurrency. Yes, PoS networks have a type of currency, but not a very good one. If you know anything about me, though, you know I’m not like most investors – I love proof of work (see: my exchange with @hasufl here: x.com/waltjsmith/status/2089…)! That is not me bragging: I only love it and sort of get it because my brother runs a media company (@Blockspace, which also runs OPNEXT, a Bitcoin tech conference) — he’s mocked me and every VC he’s met for years for comparing PoS to PoW. More importantly, through him and through his media company, I’ve heard and read Miners transition to AI compute (e.g., HPC) for years. More on that later. Given Trump 2.0’s (will be generous here) willingness to ‘embrace’ crypto, we’ve seen a lot of proof of stake chains becoming increasingly (less generous here) entrenched with the federal government. Don’t agree with me? You don’t see the ‘Bitcoin Foundation’ forming a lobby group to fly to DC to make sure their casino stays online and gets preferential market structure in the Clarity Act, do you? Albeit it’s not perfect, proof of work incentives geographic decentralization and overall is a lot more resistant to capture and hence, credibly neutral. The upshot of that and of Trump embracing and/or capturing PoS networks is that any real attempt at proof of work is pretty interesting. That includes ZCash, which has a storied history as a more cypherpunk Bitcoin and now that includes Pearl. Pearl Proof of work is insanely hard to get right and also to maintain (see: all the PoW networks that launched but never got a critical mass of hashrate online). One thing that is clear to me is the deceleration in the growth of Bitcoin’s hashrate is in part driven by price but also by AI. Why is this happening? Because GPUs make more money than ASICs (Bitcoin’s difficulty adjustments squeezes margins and folks love dancing cat videos a lot). Don’t believe me? You should believe the Bitcoin miners – one of the big ones told my brother (told you we’d get back to this) in 2023 at a miner meetup in Jackson Hole that 1. they would be worried about Bitcoin’s security budget within a few halvings and 2. they wanted to pivot entirely to HPC. Into this vacuum, the Pearl team launched and are iterating on their network. I don’t totally get the math (90% chance anyone who says they do is lying IMO) but essentially they take proof of AI inference, grind it, and create a proof of work hash out of it! And va la – you have a new proof of work network which promises to amortize costly LLM inference (with seemingly zero overhead: x.com/WeinsteinOmri/status/2…) and secure a new PoW chain all at once. Pretty cool right? While it’s not perfect yet (proof of work is hard, remember), I invested in Pearl and remain excited about it because of: 1. the erosion of crypto’s principles in Trump 2.0 and within the context of PoS in general. 2. the slow burn on Bitcoin’s security budget 3. the proliferation of AI. This sort of makes Pearl the culmination of Ribbit’s token letter (docsend.com/view/hcv2rx2cq7h…) or even Black Rock’s thesis on AI and digital assets (docsend.com/view/hcv2rx2cq7h…). In another person’s words turning intelligence (x.com/prlnet/status/20531808…) into a new type of currency is “the unholy marriage of crypto and AI.” While the network needs to grow and ensure its PoW ecosystem maintains stability above all else, the team, early work with other AI players (e.g., Together AI: x.com/waltjsmith/status/2055…), etc make me incredibly excited about it. Much like @zooko's Zcash, it's a real shot on goal to do something amazing in an industry whose innovative days are mostly behind it.