So the whole tech world is learning about OCP2.0 today via pmarca on Rogan. What happened exactly? For those curious, let's dig into OCP as it pertained to crypto. Quick thread
In 2023, I noticed post FTX crypto firms were getting debanked - just like a16z, many of our portcos. I wrote this piece in Piratewires, coining OCP2.0 and summarizing what I was seeing
Six weeks later, crypto banks Silvergate and Signature had collapsed under mysterious circumstances. Many saw this as just desserts for banks serving crypto. I felt differently, that they were "suicided" as part of OCP2.0
In this second post, I also discovered the exact mechanism of OCP2.0, which had not been reported anywhere else, because it was considered "confidential supervisory information" – the FDIC verbally messaged a secret 15% cap on crypto-related deposits to banks.
Over a year later, and choke point 2.0 was still in place. I took another look at Silvergate, post bankruptcy filing, and was increasingly convinced they were murdered by government edict.
I was in Congress last week and raised the topic of OCP2.0 with several members. The good news is that Trump's FDIC and OCC appointees will end it day 1, Trump himself has vowed to end OCP2.0 as he said in a speech. But that's not enough...
First of all, we need a full and complete investigation into how a purportedly sleepy and nonpartisan bank regulatory body like the FDIC was deputized into a political death squad. We need to learn the role of the Fed, the OCC, the DoJ, the NEC, the names of those responsible
Second, we need legislation to end the politicization of the bank regulatory apparatus, because two democrat admins in a row haven't been able to keep their hands out of the cookie jar (Trump ended OCP1.0 in his first term). This will happen again and again...
We need a law on the books that stipulates that bank regulators cannot quash a specific legal industry just because it is politically disfavored. Choke Point style behavior has now targeted over a dozen industries, with crypto just one victim among many.
More resources for those curious. In 2021 I sat down with @ismurray to cover the history of OCP1.0. He wrote a great paper on OCP 1.0: cei.org/wp-content/upl…
@ismurray In April 2023 after OCP2.0 became clear the law firm Cooper and Kirk wrote a paper on how OCP2.0 is unconstitutional: cooperkirk.com/wp-content/upl…
@ismurray More recently, Coinbase FOIAed the FDIC and received documents summarizing a systematic FDIC campaign to get banks to de-bank crypto startups - hard evidence of OCP2.0
Last week, Rep French Hill asked FDIC Chair Gruenberg under oath whether the FDIC was in the habit of sending verbal guidance to banks to have them derisk crypto. He says no. We know that he's not telling the truth. Rep Hill has vowed to get to the bottom of it if he gets the chair of the House FinServ Cmte
@ismurray Trump has specifically vowed to end OCP2.0 – used the term directly in his speech in Nashville. I believe he will end it on day 1, but we need to go further and push for full investigations and legislation ensuring it can't happen again
walling off banks from crypto makes the system _more_ fragile, not less. the SAB121 rule prohibiting banks from custodying crypto means that big custodians like BNY mellon can't enter the market. they would be manifestly more competent than the crypto firms that sprung up instead.
maintaining the anti-crypto posture meant that only boutiques were willing to serve crypto and ended up with high deposit concentration.
in an ordinary regime, banks would be free to serve crypto firms at their pleasure, meaning less deposit concentration.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
agreed. it can be fixed sans soft fork for anyone that's live on the network, but old/abandoned coins in vulnerable addresses (p2pk, reused p2pkh) are at risk. no easy way to fix. some quantum timelines have a break by 2030.
and yes bitcoin is more vulnerable than other systems that rely on cryptography, because they can be upgraded to post quantum encryption whereas there's millions of coins in old addresses (assumed lost/inactive owners) which are fundamentally at risk.
basically you have to make your peace with the fact that there's a 2-6m BTC (aka $200b-600b) prize waiting to be claimed by the first entity to achieve quantum supremacy. the other alternative is getting everyone to agree to pre-emptively burn/steal the coins which is unpalatable
unfortunately the yale academics are at it again and are trying to gaslight everyone into thinking that debanking crypto didn't happen. it did, everyone dealt with it, everyone knows it, and the evidence is excruciatingly clear. full treatment soon
1. Steven points out that Coinbase has banking. Yes. They actually have had struggles with banking and are only really around today because a16z had to intervene to ensure they had bank access prior to 2020 when they landed the JPM relationship. One of the reasons Coinbase emerged as the winner in the US is precisely because they had better bank relationships than others. It proves our point that banking is a weird thing that determines winners and losers in crypto, and it shouldn't be that way.
But that's not really what we're complaining about. Coinbase is the largest and most credible crypto firm in the US, they are publicly traded. Obviously they're able to get banking. It's everyone else that has suffered. Especially startups, who aren't able to easily make the case that they are a valuable relationship to a bank that's already leery of banking crypto / facing pressure from regulators. Coinbase being adequately banked is largely irrelevant.
2. Steven says the FDIC showed favoritism to crypto banks SVB and SBNY by making depositors whole. This is silly. First of all, there was a massive regional banking crisis going on, and the FDIC had to step in to stop the run. SVB in particular was systemic to US tech and startups. They had to be saved or most of the dynamism in our tech sector would have evaporated overnight. Saving them was the right decision. They also only had one (1) crypto client, Circle.
SBNY are a different story. They had more crypto clients. But they were not shown favoritism. Arguably, they were executed for serving crypto. That's what the board member Barney Frank alleges. They were sent into receivership on a sunday night without being given a chance to raise capital and save themselves. Also, when they were sold to Flagstar, all their crypto related deposits and IP (Signet) were stripped out. The resolution of SBNY shows no preference towards crypto, only animosity. The refusal to transfer the crypto business was one of the most obvious signs in early 2023 that something was amiss.
Deepseek has completely upended the conventional wisdom around AI
- China will only do closed source / proprietary
- Silicon Valley is the global nexus of AI development, has a huge headstart
- OpenAI has an unbeatable moat
- you need to spend tens, maybe hundreds of billions for SOTA model development
- value will accrete to the models (fat model hypothesis)
- scaling hypothesis means model performance is a ~linear function of training inputs costs (compute, data, GPUs)
All these narratives, shaken if not completely undermined overnight. Stay frosty
All of that said, I don’t worry too too much about equity value in NVIDIA/AI datacenter companies (although I have massive bag bias here, am invested in several of the largest AI datacenters cos). Why? When a commodity gets cheaper, and it looks like even disregarding DS’ claims around training costs, factually speaking inference costs are down by a factor of 30 compared with OAI SOTA model, the use of that commodity increases. This is jevons’ paradox. So inference overnight became vastly more abundant. DS’ innovations will be rapidly incorporated by other model cos. So AI can be embedded cheaply everywhere. This probably shifts the ratio of training to inference in AI Capex in favor of the latter but I don’t believe it undermines equity value in the firms that produce the inputs for inference - GPUs, datacenters, etc. Just accelerates the transition from pre AI world to fully embedded world.
All of that said, the investor premise that the model cos - OAI, Anthropic etc - are where equity value will accrete has a massive hole in it now. I always felt and have said that I thought model cos would be capital incinerators due to high quality open source models and a race to the bottom, and I think that is more true now. But overall I don’t worry about the rest of the stack, whether it’s the producers (NVDA, datacenters) or the firms (mag7) that are actually bundling up compute and selling it to the end user in the form of better consumer experiences.
TLDR- for most of you no need to panic. Although I think it will take the market some time to digest and the ride will be bumpy in the near term.
Receipts on me thinking fat model hypothesis is bad / model cos as capital incinerators
Thanks to @iampaulgrewal and the rest of the team at Coinbase, the FDIC has now been compelled to further un-redact their "pause letters" to banks during 2022-23. We know know what the FDIC was asking banks not to do (thread)
For an introduction on what the pause letters were all about and the context, see this thread:
I've summarized the new batch of pause letters here. When I add 'suspected' that's me guessing what bank or product the redactions pertain to. Credit to @john_iller for making the USDF connection.
Longer thread on the FDIC’s 2022-23 “pause letters” to banks regarding crypto-asset activities in 2022 and 2023. Strap in
So we all know that banks have been stymied from doing crypto stuff for years: regulators installed 15% deposit caps on fiat deposits from crypto firms, cutting off bank access for the whole industry; the SEC used SAB121 to stop banks custodying crypto; and the Fed effectively prohibited banks from doing stablecoin stuff. But maybe the worst and most brazen was the FDIC unilaterally barring banks from doing any crypto stuff at all…
So recently, Coinbase made FOIA requests to the FDIC regarding their guidance to banks on crypto matters. Last month, the FDIC published summaries of these letters, without saying what they said exactly, or which banks they were addressed to. assets.ctfassets.net/c5bd0wqjc7v0/7…
@pmarca Pmarca claims debanking is targeting crypto founders, fintech founders, and conservatives. He says the CFPI and other financial regulators are responsible. Lee Fang and others say the CFPI is anti-debanking. Who is right?