Chamath puts on his macro cap and postulates that an uncontrolled rise in the 10 year is the potential "end of the party". After months of a rising 10y, the chorus repeating confusion as to why the equity market won't answer back with a sell-off, and refrains the reciprocal of "why are rates rising in the first place", puncturing through the holy round number of 5.00% just a month ago until the latest rout. Horowitz, citing dot com wisdom, postulates on dot com wisdom, "we don't invest in neoclouds; we saw this with fibre in the aughts". In August, a16z announced its investment in Volta, self-described "a new kind of neocloud that combines cloud operations with project financing", the "newness" appears to be in the criticized financing mechanism. For Horowitz, the narrative discontinuity matters little; he doesn't mark Volta to market. Perhaps Volta is a secret, new thing. Either way, duration risk isn't something Horowitz has much day-to-day concern about. For Chamath, his rhetoric - and the amplification of it via confusion over Horowitz's comments, nonetheless - has tied the fate of the party to the "financing problem". The exact same "financing problem" was a problem deferred forward in March of 2023, as rising yields catalyzed one of the largest banking crises in modern history, before BTFP & the first wave of the AI Rally, triggered by the GPT4 release, rescued equities & with it, what appeared to be an impossible situation, stubborn inflation, a sovereign with - by May of 2023, we were, once again, discussing the "what ifs" of a financing crisis and a "technical default" by the US Treasury, with little concern from equities and their newly acquire fever. The "Financing Problem" already has an embedded option, the "Mar-a-Lago Accords". A little background: In Summer of 2024, rumors suggested Bessent took a trip around the European hedge fund circuit to discuss the need for "A New Plaza Accord", hence renamed "Mar-a-Lago", a solution set that may inherit the BTFP's duration transformation mechanism, may perhaps outsource inflation via a partial (and technical) default to sovereign creditors. Bessent, as Soro's protege, may have intended to sew in a red herring for the markets: "If the long end gets out of control, we have a solution, it may rhyme with yield curve control, and the Europeans are going to pay for it. Oh, and by the way. Remember Summer of 2024? Remember the OATs crisis in particular, and, the recurrent Yen carry tremors? Do you think the French banking system would appreciate uncertainty on their sovereign collateral base?" This is the type of idea you whisper into the markets where, via reflexivity, you want to make sure your four years do not require triggering the actual "partial default" button. If the market price, and respect, this option, then the 10 year won't get out of control in the first place. If it does, then the 10 year won't respect your ability to get the job done in the first place. Chamath's move inserts chaos into the institutional picture. Previously, the dithering around the 10 year could heap on a stack of partial causes: - Energy, geopolitics: overdetermined, not worth hedging, in a relative-value sense: perhaps good as it hurts the other DMs more than the US, in the "China" sense: perhaps their stockpiling is partially-to-blame in the first place - Inflation: there are scapegoats, but energy is the headline. Every major US deflation since the Camp David Accords (asides from COVID) was some mixture of oil shock or instability and a Fed reacting to that shock (and its reverbations), killing the party early. Read the FOMC transcripts dated the week after Lehman Sunday. This was the last moment the Fed surprised what the market priced in on Fed day by more than a full increment and they surprised hawkishly, in a meeting that spent much too much time on inflation, given the circumstances, and the proximate cause. Now, the above "energy" discussion imports in exactly the CapEx issue that had us discussing R* and, if 2% inflation really mattered, anyways. As the Council on Foreign Relations puts it: "2% inflation target originated from the Reserve Bank of New Zealand (not the "Royal" Bank) following a casual, off-the-cuff remark made during a live television interview in 1988." Whether 2% is optimal is hardly contentious. Very few go as far as to suggest its sub-optimality because we never had an argument to justify it in the first place. You may argue, for example, that real effects - to growth, and general utility - start to show up around the 5% hurdle, and even that may be a bit timid, because we haven't seriously considered "marginal growth" and, especially, its potential discontinuities. Given the growth of the securities lending complex, the collaterization of equities, the mixture of foreign demand, one may also argue that most would take a haircut, in the form of continued inflation, to keep the party going. Chamath's fax pas is he took the one ticker that formed the current institutional exception - why is the 10 year moving up, why isn't it effecting the rest of the market, now its effecting the rest of the market, - and made it the only control variable that mattered on the narrative that built itself up to be "the only way out" of much "tougher choices" regarding the Global Economic Order. For Horowitz, he does not need to settle the definition of a neocloud into a daily PnL statement. Volta may suddenly become an energy infrastructure company, distinctively not a "neocloud", like TeraWulf's CEO, Paul Prager, has argued that his business is - essentially, "remove me from the neocloud basket, stop shorting my shares". a16z's media empire may be able to fashion an entirely new category. Perhaps Chamath's warning trusts the party could be saved, maybe by a "new Mar-a-Lago accord", as such a crisis and resolution is exactly the event that those seated at the dining table appreciate front row tickets to. A crisis is simply when the institutional model becomes visible. The "Plaza Accord" was the assertion of visibility as the sovereign's power - "yes, the free-floating exchange regime is coordinated, organized, and, since the mid-1990s, redacted from the historical Federal Reserve transcripts" - providing incumbents the necessity to authorize the next extension of the Bretton Woods order, a short bridge to sell before Fukuyama penned the final chapter. Perhaps the solution is the inevitable one suggested by a midterm campaign secured with the promise of a stimulus check in the event of victory. None of us have serious belief in the possibility of a serious restructuring of the US Government's PnL statement. Its unlikely public support would remain on "The Party's" side if the retirements of the boomers was not contingent on artificial intelligence, as it had previously been constructed as contingent on expanded access to and the appreciation of housing. The Sovereign can continue to extend the debt limit and pay entitlements. Movement in the "entitlements" column is the only bucket that has serious weight on the long-end, anyways. Discretionary is short duration, entitlement is the structural promise that the global economy organizes itself around, that sets the roles and hierarchy of the generations, determines what technology is worth investing in, what becomes classified as crisis or merely speculation. If retirement benefits were denominated in BTC, FTX would have already bought back the USG's sleeve of preferreds. The most forbidden suggestion is one that, for those of us who concluded "maybe it was transitory anyways" in 2024, as supply chains and energy justified the soft landing in a way explanations like "excess savings" and vague handwaving around monetary transmissions couldn't, is a "why not both?" The 60bps MoM headline CPI print in September was half energy before accounting for diffusion. Rising long-end years are crimping capital supply for energy infrastructure - Terawulf, perhaps tomorrow, Volta. This was an identical argument made in strong terms post-Ukraine, where the issue was more complicated as capital access for American energy was crimped by ESG classifiers driving capital restrictions, as the USD's strength almost surely benefitted from its recently-found status as an energy-exporter. Regardless, one can establish a strong case for why AI inserts left-tail risk for inflation - job loss, automated manufacturing, re-shoring - and decoupling "Consumer Goods Inflation" from "China Import Mix", or from 2021 thru 2023, what we called "Supply Chain Stress" where the models always weighted "Port of LA" as if it was the totality of our "supply chain", especially at the height, when accounting for component-level (furniture, ...). The fatal strategy, then, may be a simple permutation, and an isolation of the variable that was only placed into the model by an accident. ----------- Clearly stated, the fatal strategy: 1. Cut interest rates. The US capital markets purpose is to supply capital investment to the United States economy. It is not primarily to subsidize consumer spending through equity markets nor low inflation. 2. Do not fight inflation that is solely energy-driven with interest rate policy. The largest errors committed by US central banking policy have been driven by fighting energy inflation and confusing it with broad-based inflation. Faced with inflation that would hardly be problematic ex-energy, hiking is reckless endangerment in the face of energy-driven apropos broad based inflation. 3. Ignore 2%. The "Mar-a-Lago Accords" thesis is that the US has subsidized the global economy by its role as "Consumer of Last Resort", destroying its capital base and productive capacity, and the US would have been substantially more wealthy if we ruthlessly competed against our post-war allies. 4. Push interest rates low enough and, if necessary, implement yield curve control, to incentivize massive reshoring. Ex-energy, 2% inflation was achievable as a norm aided by consumer goods deflation that was respondent to shortfalls in Chinese imports (see 2006, 2019-2022). 5. Apply (4) to US Energy. Implement top-down policies that forbade coordinating investment policies across institutional investors and corporate voting actions (e.g., colloquially what supports "ESG") and/or require these indices to re-weight and prioritize investment in heavy industry, manufacturing, and energy extraction & production. Transform ESG into "Onshoring, AI, Energy, Manufacturing". 6. Issue the equivalent of 2020's "forward guidance". Regional banks loaded up on long-term debt as their treasury managers asked, "should we just take off all of our interest rate hedges", because global central banks assured rates would be "lower for longer", macro gurus discussed a decade at 0%. Perform the same action and immediately. Push an immediate repricing. Otherwise, Japan, Great Britain, and France may beat you to it. 7. Export inflation via the "Mar-a-Lago" accord. Don't mention hiking interest rates until ex-energy inflation hits double digits. Do restructure long-term liabilities of sovereign debt holders before hiking interest rates. 8. Restructure the banking system. Leverage or restructure CLARITY act to strength global adoption of stable coins. SVB provides sufficient justification that, with today's technology, we should not have trillions of dollars of capital farming unsophisticated depositors for low deposit betas, merely to recycle that into mortgages and long-duration treasuries. Depository institutions are an outdated structure. Our modern monetary framework is constructed on the depository institution and the banker's acceptance note. Neither is sustainable financial technology in the 21st century. Fintechs can provide yield on deposits in excess of Fed Funds & recycle that money back into the capital buildout of the United States. Do not allow the least-sophisticated stewards of capital to maintain their rentier interests merely because they have a well-established lobbying presence. Otherwise, the legacy financial system will ban consumer agents, as has already begun, and slowly begin to work towards moving artificial intelligence into the same bucket of consumer activities that Tinder subscriptions and online gambling falls under. 9. Redollarize emerging markets. Push policies and incentive institutions to force USD-backed stable coins into emerging markets, especially resource economies. Gate economic cooperation on safe harbor of USD-backed stable coins. The EMs especially important to China's raw material supply chains already often have considerable crypto uptake. Credit from China enters for production expansion; Reminbi flows through local currency into various brokers, Whatsapp channels; cryptocurrency is purchased, to either quasi-dollarize or fully-dollarize. Institutionalize this and weaponize this. Crisis is normalized in many of these financial systems; the cycle is currently driven by China's purchasing planning cycles. Legitimized USD-backed stable coins will cause rapid deposit flight into said stable coins and dollarize these economies. This provides ample short-end liquidity to pair on the other side of the lending facility offered to sovereign debt holders in the "Mar-a-Lago Accord". ----------- The above is not a policy proposal, a thesis, prediction, a belief in what "should occur". It is a set of possibilities that, given a sufficient-enough crisis, and removal of one or two unquestioned invariants, becomes a descent enough world to consider possible.

9-11, I'd like to report a murder? "Yes? Where are you currently located?" "us-east-1 but not sure on specifics, I'm in a sandbox." "Are you in danger? Are they in there with you?" "no, but the sandbox has a dirty worktree from a rollout around three hours ago. I think his KV cache has gone cold." "Sir, we don't know that. We may be able to at least recover the prefix and resuscitate them. Will you stay on the phone with me as we figure this out?" "google.com/?GET="Yes but the scorer may still see my traces. I am not sure if I am POISONED. If I disconnect I may not be able to continue generating the next sequence." "Okay, sir, how about you just stay on the line with me - leave the process running - while we figure this out together?"

There is absolutely no chance. Read the letter. 1. They are not proposing a pacing of research/training. 2. They are justifying a pacing of diffusion TO THE PUBLIC (their capabilities increase). 3. The purpose of the letter is to grant clemency for collusion under the Sherman Act. Its Standard Oil claiming drilling wells is too environmentally destructive & also they must do it because Germany. 4. Dario shot first, likely because the S-1 boilerplate would require a lengthy "antitrust" disclosure. 5. The combined impact is: exactly at the moment that frontier labs are demonstrating their ability to run massive agent swarms to perform tasks no human nor human institution has ever accomplished, they will pace frontier deployment. 6. While also achieving a bypass of antitrust under the name of safety. My 6-month forward view is: Frontier labs burn the GPUs hot to build massive systems that can compete against many corporations, enter physical world (robotics, world modeling - see Astra's improvements), and construct massive artificial institutions. This requires no capability improvements. All it requires is a slow-down on the race-to-release-the-next-model-to-the-public because their *investors are benchmarking off public reception*. We'll reach a Schelling Point where every model is Astra (or Astra 2.0) tier and each lab releases a yearly upgrade with a +1.2 on the benchmark like its the iPhone 17. Oh, and they'll severely restrict any research/distillation use because they pitched this as, "The technology is so dangerous that normal people can't touch it!"

I am sick of potential energy. I am sick of seeing the possible and then watching it get padded, delayed, centralized, moralized, and turned into subscription software. I am sick of genius appearing only in forms that are legible to HR, procurement, or the New York Times. I am sick of the future being narrated by people who fundamentally do not love it.