Synthesis

I must say… private transactions and uncensorable AI are more timely than ever. Private, Proof-based Money powered by decentralized compute networks. Goodness.

Anthropic is releasing an open source library, FlashProve, for provable inference. It was measured on Llama 4 Maverick, a 400B-parameter mixture of experts model. Code: github.com/anthropics/FlashP… Blog post: anthropic.com/research/prova…

Not sure the point you’re making. Let’s see how this plays out !

I feel sorry for people who follow @tulipking and think his theories on digital assets are valuable or deep. Real respect for someone out there giving it their best, but anyone who has been in this space for a while all read this as a very surface level interpretation of the reaction against generalist chains. Like an analyst read one blog post on “moneyness” and never took the theory any further. If he truly wants to develop this theory further, he needs to build a thesis around how proof based computation and proof of work offer programmability for historically would have been account based POS assets with poor economics. I haven’t seen this anywhere in his recent posts (if it’s there, would love to read it). He’s still conflating the core issue of the last cycle: pre-mined POS chains at artificially high valuations made the underlying cryptography, no matter the breakthrough, almost impossible to defend as having “moneyness”. He’s going to have a very hard time when the next wave of protocols appear that are better than Pearl and offer programmability….however his HARDEST times will come when Pearl, or someone in the Pearl ecosystem, realizes they need smart contracts and launch a fragmented, uncoordinated L2 initiative. I give it 12-18 months. See Kaspa. I wonder how his perception of Pearl will change when this happens.

"It is not necessary that one must fail and the other must succeed", It takes courage to be a hater More refined: you need to decide what these two coins really are bc if they are competing, pick a side The analysis of Pearl is pretty simple. It wants to be AI Bitcoin. Make a simple swap of SHA256 for MatMul and you're done. Obviously there's a bunch of downstream implications: Does useful work solve the security budget? Can you bootstrap a useful network? Is this more or less secure than useless work? etc... But the design space for Pearl is straightforward. We all at least know what Pearl is trying to do, whether or not we agree it will do that is a little irrelevant to the point I'm making Nockchain is much harder to wrap your head around. I'm not necessarily saying that's a good or a bad thing, complicated things can be valuable, but notice how much word count the article dedicates to explaining Nockchain vs Pearl (again not a bad thing or a dig at the article, it legitimately takes more words to explain Nockchain given the complexity) So you get left in a few kinda awkward spots with respect to Nockchain: 1. Is Nockchain trying to be an AI money like Pearl? They've clearly identified AI compute as the next trend to ride and are spending a bunch of time developing and narrating about it (again, not a bad thing) For many reasons I obviously don't think Nockchain is money: It's too complex (Bitcoin and Zcash may be technically complicated but architecturally and memetically they are simple), The NockVM introduces programmability which is -EV for moneyness, & it's not pure (I can see MatMul Bitcoin being AI Money, it's harder to see zkp verifiable compute being AI Money) > I believe there's two rational points of disagreement with me. 1a, Maybe you think Nock is something other than money. Fair enough, in that environment I agree Nock & Pearl are not competitive, you can own both so long as you have your own thesis for underwriting Nock as a money making protocol rather than a money protocol. 1b, Maybe you think Nock is money but disagree with my definition of money. Again, fair enough but obviously I agree with myself instead of you 2. If Nockchain is not AI money like Pearl, then what is it? To me it seems most similar to Ethereum but designed around the paradigm of verifiable compute. That really is cool and interesting, but it comes with all the problems I believe Ethereum has: Blockspace is a commodity it's hard to see proving networks not being commoditized if they become super relevant, The network can charge based on throughput but not value (it's more like the post office than an exchange), and Interoperability makes vamp attacks that kill Lindy too viable ⭐️ So why does it matter? Why am I hating? If Nock is money, then we're playing a zero sum game. Look a the zec/xmr chart, the market is learning how to price money & you don't want to be hold the wrong one If Nock is something else then I'm not really hating, I'm just being kinda inflammatory. I genuinely think what they're building is conceptually cool and an experiment worth building. I'm just not a buyer for now

More evidence of the Tulip King thesis. These smart contract L1s are not money. They cannot dilute holders ad infinium and simultaneously collect monetary premium Expect all the serious L1s to follow Solana's example like NEAR is in a general push towards sustainable economics. Basically they're becoming more like businesses (good because that's what they are) Apple goes up because while they issue ~$60M/yr in share compensation to employees, they buy back ~$400M/yr a year. Compound this reliably for many decades (include a good dividend although we don't really do that in crypto, it's fine) and viola the stock price goes to the moon Your downstream research should be on the following: Do staking rates go up or down as yield falls? What's the impact on price in either situation? Lowering inflation is only one side of the equation, fees generated also needs to go up. Which, if any, of these networks can get to break even and eventually profits? What profits are required to sustain the valuations these coins need to hit for your bags to make money? Is it possible to get there? It's obviously up for debate but my opinion is that: 1. staking rates fall with yields. My suspicion is most PoS yield is just capturing some inefficiency between those who stake & those who don't. As this arb closes, so should the incentive to stake. More supply will effectively become liquid and have a generally negative impact on price 2. To make things doubly bad, I also doubt the smart contract networks can ever become super profitable. There's too many flaws in the business model: * Fees do not scale with txn value. I pay the same fee to send you $10 as $100k. If I'm sending $100k perhaps I'm willing to pay more to send the txn in a private way or faster, but i'm not willing to pay 10,000x more. At BEST, fees scale extremely sub-linearly with value > I'll re-iterate my post office example. Just because I mail you a trillion dollars, does not make the post office worth a trillion dollars. A smart contract network can facilitate a huge amount of value and commerce without actually capturing a relevant amount of that value itself * Blockspace is a commodity. Just look at gas prices & total fees. The median gas price on Ethereum is like $0.04 and transaction fees on solana are down ~80% YoY DESPITE the total count of transactions going up ! Pick your favorite chain, it doesn't seem to be able to extract fees either way Btw all of this is GOOOD for other coins. The money the comes out of sc l1s (bad money) can go into the select l1s that are good money. The money can reallocate into onchain finance protocols and send them higher than any of you can imagine. More mature & wise capital allocation is good, just don't hold bad bags !

Are you a privacy project, programmable gold, or compute backed money? Because you claim all three (usually depending on whichever narrative is hot at any given time). He who chases two rabbits catches neither.

changed my mind on $nock might be a very big opportunity here decentralized AI PoW with ZK proofs EXTREMELY on narrative

I read everything about $PRL so you don't have to: It's Bitcoin but AI

No coin has ever eclipsed #bitcoin The narrative is always "it does more stuff" The reality is that it means it's just a software company without an actual market Zcash is trying to ride bitcoins narrative Ethereunet already tried this and is far more decentralized than Zcash

Zero knowledge / Zero trust will sell for a significant premium this decade. Plan accordingly.

Another really weak argument for $PRL. You do realize that $NOCK has the flexibly to launch a way wider array of compute puzzles and future AI systems that pearl right? Its surface area for future partners is exponentially larger. Not only does it have a better path - its programmability and zkvm will make it *way* easier to integrate with external systems.

I completely agree about those differences between pearl and nock, that’s why pearl will be AI Bitcoin. A simple focus on low overhead dual use inference. FP8 enables more frontier open source models, and game theory says more of the inference run by the network will come from subsiding production workloads over time ^ that’s all that needs to be and should be done if you’re trying to redesign Bitcoin around an AI algorithm (matmul) Nock actually goes wrong by doing too much. It’s more like AI Ethereum. Which is not a good thing