🚨 THIS IS NOT NORMAL SpaceX, OpenAI, and Anthropic are preparing to go public. All at once. To absorb these giants, the market will need to find $200 BILLION in fresh liquidity. This is a brutal stress test for global investors. To buy into these IPOs, big funds need a lot of cash. They will ruthlessly cut overextended tech. First in line: NVIDIA, Microsoft, and Google. But these companies hold up the entire S&P 500. If they start to fall, the index will follow. We saw this before during the COVID bubble. Dozens of tech companies (Rivian, Coinbase, Robinhood) hit the market at crazy high prices. When the Fed hiked rates and liquidity dried up, these stocks crashed 80%. The AI and tech sectors are already running on fumes. Add a $200B drain, and you get a systemic flush. Get positioned before this massive rotation begins. As a reminder, I’ve called every major market turn for the last 10 years, including the $111K BTC peak in October. Turn on notifications. I will guide you through this liquidity squeeze, and when the real bottom finally forms, I’ll call it here publicly.

Two years ago, I wrote this post on the possible areas that I see for ethereum + AI intersections: This is a topic that many people are excited about, but where I always worry that we think about the two from completely separate philosophical perspectives. I am reminded of Toly's recent tweet that I should "work on AGI". I appreciate the compliment, for him to think that I am capable of contributing to such a lofty thing. However, I get this feeling that the frame of "work on AGI" itself contains an error: it is fundamentally undifferentiated, and has the connotation of "do the thing that, if you don't do it, someone else will do anyway two months later; the main difference is that you get to be the one at the top" (though this may not have been Toly's intention). It would be like describing Ethereum as "working in finance" or "working on computing". To me, Ethereum, and my own view of how our civilization should do AGI, are precisely about choosing a positive direction rather than embracing undifferentiated acceleration of the arrow, and also I think it's actually important to integrate the crypto and AI perspectives. I want an AI future where: * We foster human freedom and empowerment (ie. we avoid both humans being relegated to retirement by AIs, and permanently stripped of power by human power structures that become impossible to surpass or escape) * The world does not blow up (both "classic" superintelligent AI doom, and more chaotic scenarios from various forms of offense outpacing defense, cf. the four defense quadrants from the d/acc posts) In the long term, this may involve crazy things like humans uploading or merging with AI, for those who want to be able to keep up with highly intelligent entities that can think a million times faster on silicon substrate. In the shorter term, it involves much more "ordinary" ideas, but still ideas that require deep rethinking compared to previous computing paradigms. So now, my updated view, which definitely focuses on that shorter term, and where Ethereum plays an important role but is only one piece of a bigger puzzle: # Building tooling to make more trustless and/or private interaction with AIs possible. This includes: * Local LLM tooling * ZK-payment for API calls (so you can call remote models without linking your identity from call to call) * Ongoing work into cryptographic ways to improve AI privacy * Client-side verification of cryptographic proofs, TEE attestations, and any other forms of server-side assurance Basically, the kinds of things we might also build for non-LLM compute (see eg. my ethereum privacy roadmap from a year ago ), but for LLM calls as the compute we are protecting. # Ethereum as an economic layer for AI-related interactions This includes: * API calls * Bots hiring bots * Security deposits, potentially eventually more complicated contraptions like onchain dispute resolution * ERC-8004, AI reputation ideas The goal here is to enable AIs to interact economically, which makes viable more decentralized AI architectures (as opposed to non-economic coordination between AIs that are all designed and run by one organization "in-house"). Economies not for the sake of economies, but to enable more decentralized authority. # Make the cypherpunk "mountain man" vision a reality Basically, take the vision that cypherpunk radicals have always dreamed of (don't trust; verify everything), that has been nonviable in reality because humans are never actually going to verify all the code ourselves. Now, we can finally make that vision happen, with LLMs doing the hard parts. This includes: * Interacting with ethereum apps without needing third party UIs * Having a local model propose transactions for you on its own * Having a local model verify transactions created by dapp UIs * Local smart contract auditing, and assistance interpreting the meaning of FV proofs provided by others * Verifying trust models of applications and protocols # Make much better markets and governance a reality Prediction and decision markets, decentralized governance, quadratic voting, combinatorial auctions, universal barter economy, and all kinds of constructions are all beautiful in theory, but have been greatly hampered in reality by one big constraint: limits to human attention and decision-making power. LLMs remove that limitation, and massively scale human judgement. Hence, we can revisit all of those ideas. These are all things that Ethereum can help to make a reality. They are also ideas that are in the d/acc spirit: enabling decentralized cooperation, and improving defense. We can revisit the best ideas from 2014, and add on top many more new and better ones, and with AI (and ZK) we have a whole new set of tools to make them come to life. We can describe the above as a 2x2 chart. There's a lot to build!

I am the Director of Summit Outcomes for the Presidential Advance Team. My job is to land in a foreign capital and leave with a word the President can say on the tarmac. We landed in Beijing 6 days after rolling back the tariffs we spent 4 years imposing. 145% to 30%. The average rate before the trade war was approximately 3%. In Geneva, we called this "creating the conditions for productive dialogue." The conditions were that we had already conceded. I want to be clear: Beijing was a success. We went in with 7 objectives. We left with 3 photo categories, a tentative agreement China has not confirmed, and a bag of burner phones we threw off Air Force One on the tarmac. Diplomacy. My team prepared the deliverables matrix in March. 241 line items organized by urgency, feasibility, and what we call "headline potential." The President reviewed it for 4 minutes. He circled "big deal" and "historic" and wrote "MORE" next to the Boeing section. That became the strategy. Boeing was the centerpiece. 500 aircraft was the White House number we briefed to reporters before departure. 300 was the floor. The Chinese offered 200. Their commerce ministry released the number before we could brief the press. Boeing stock dropped 4.73% that afternoon. Boeing referred questions about the order to the White House. The company receiving the aircraft could not confirm it was receiving aircraft. We called it "fantastic." In Washington, "fantastic" means the other side named the number and the market already priced in your failure. I should note: in 2017, the President announced $250 billion in deals during his first China trip. 300 aircraft. An $84 billion shale gas investment in West Virginia from China Energy Investment Corporation. I can tell you the exact amount of that investment that materialized. Zero. The shale facility was never built. The 2017 Boeing order was renegotiated twice and partially canceled during the trade war the President started 8 months later. There is a binder in my office labeled "2017 OUTCOMES: DO NOT REFERENCE." It is 3 inches thick. It has not been opened in 4 years. We do not reference it because the outcomes are the reference. The agricultural package was what we call a "scaffolding commitment." Billions in purchases over 3 years, structured so the announcement is front-loaded and the verification is someone else's administration. U.S. Trade Representative Greer said "double-digit billions." Beijing's Commerce Ministry issued a statement about "deepening cooperation in agricultural trade." Those are not the same sentence. By design. My deputy maintains a glossary of every term we have invented for agreements that are not agreements. It is 41 pages. He updates it after each summit. Last quarter he added "scaffolding commitment," "streamlined licensing framework," and "mutual recognition of shared concerns." He is in line for a promotion. NVIDIA was the quiet win. H200 chips approved for approximately 10 Chinese companies. We don't say "approved." We say "under a streamlined licensing framework." The chips ship. The export controls remain "in effect." The framework is the loophole wearing a lanyard. The controls exist because these chips in Chinese hands threaten American national security. The chips are shipping to Chinese hands. The controls remain in effect. Both of these are true. Fentanyl was discussed for 9 minutes. Both sides agreed it was a problem. Both sides agreed to continue discussing it. We added it to the deliverables matrix under "ongoing mutual engagement." The previous version of the matrix also listed it under "ongoing mutual engagement." That was in 2023. I copied the line item from the 2023 matrix into the 2026 version. Changed the date. The language was identical. But Taiwan. Taiwan was the deliverable we didn't put on the matrix. I watched the Taiwan exchange from the overflow room on a 12-second delay. I had the contingency statement drafted in 3 versions: "productive exchange," "frank discussion," and "both sides reaffirmed their respective positions." I used none of them. There was no contingency for silence. Chairman Xi released his remarks before the meeting was over. While the President was still seated across the table, Chinese state media published the transcript. "Clashes and even conflicts." His bluntest language on Taiwan in the history of the relationship, released to 1.4 billion people while we were still pouring tea. We called this "sequencing." The President was asked whether he would defend Taiwan if China attacked. He chose not to answer. We wrote that down as "a strong listen." The $14 billion arms sale. Already approved by Congress. The largest in the history of the Taiwan Relations Act. Taiwan's parliament spent months appropriating the $25 billion to proceed with this package and the $11 billion tranche approved last year. They finally secured the funding this month. The President told Fox News it was "a very good negotiating chip." He used the word "chip." Referring to the defense of 24 million people. Taiwan's Ministry of National Defense sent our office a letter requesting clarity on the delivery timeline. 3 pages. It referenced specific weapons systems by name: F-16V Block 70 fighters, HIMARS launchers, Harpoon coastal defense missiles. The letter was addressed to me. I filed it under "pending." On Air Force One, a reporter asked about the 1982 Six Assurances, the framework in which the United States committed not to consult with Beijing before selling arms to Taiwan. The President said: "What am I going to do, say I don't want to talk to you about it because I have an agreement wrote in 1982? No, we discussed arms sales." 44 years of bipartisan Taiwan policy, dismissed in 2 sentences at 38,000 feet. We are calling this "a modernized approach to alliance management." Our readout mentioned trade, agriculture, energy, and regional stability. It did not mention Taiwan. I wrote it. Their readout opened with Taiwan. I have staffed 7 summits across 2 administrations. This is the first where I could not draft a single deliverable as a success without a qualifier. In my office there is a laminated card that lists every synonym for "undecided" that polls above 40% approval. "Active review" is 3rd. "Determination" is 7th. Both tested well with independents in the Midwest. He also said: "Taiwan would be very smart to cool it a little bit. China would be very smart to cool it a little bit." He was eating a cheeseburger. He said this while eating a cheeseburger. Secretary Rubio told NBC that Taiwan arms sales "did not feature prominently." This is accurate in the same way that the iceberg did not feature prominently in the Titanic's itinerary. Representative McCaul, Republican of Texas, former chairman of the House Foreign Affairs Committee, said the United States must "arm Taiwan so they can defend themselves." He said Xi was "very aggressive" regarding Taiwan during the summit and that "most of what Xi talked about was Taiwan." Representative Meeks, Democrat of New York, ranking member of the same committee, said Xi has "leverage over the president" but not "over the United States Congress and the American people." He noted that Congress already approved the package. "The president is the one that's holding it up." Representative Fitzpatrick, Republican of Pennsylvania, compared Taiwan to Ukraine. He called both "fortresses of democracy on the front lines." Speaker Johnson said Taiwan needs to "stay independent and secure." The bipartisan consensus was that something had gone wrong. The bipartisan action was press quotes. No vote. No resolution. No hearing scheduled. 4 members of Congress from both parties said the right words to reporters and then went to lunch. That's how the system processes alarm. I monitor 14 accounts we classify as "aligned messaging amplifiers." Within 4 hours of the Taiwan exchange, 9 went silent. 2 pivoted to fentanyl. 1 posted 3 words: "Not like this." It received 280,000 impressions in 90 minutes. He deleted it and posted about the border instead. The President patted Chairman Xi on the back 7 times during the Zhongnanhai garden walk. We counted. He called him "my friend" in 4 languages, 2 of which he does not speak. He asked if other world leaders had been invited to the compound. They had. Putin was there last year. The President asked if his tour was longer. 15 CEOs flew with us to Beijing. Their combined net worth approaches $1 trillion. Cook. Musk. Jensen Huang. Larry Fink from BlackRock. Jane Fraser from Citigroup. David Solomon from Goldman Sachs. Stephen Schwarzman from Blackstone. Kelly Ortberg from Boeing. The CEO of Visa. The CEO of Mastercard. The CEO of Qualcomm. Illumina. Micron. Cargill. GE Aerospace. Musk and Huang rode on Air Force One. The others flew commercial. Tesla's Shanghai factory produces approximately half of the company's vehicles worldwide. Musk's presence on Air Force One was noted by my counterintelligence liaison. No further action was taken. We organized the state banquet seating chart by net worth. I am told this was the President's suggestion. They came for market access. Xi told them China would "open further to American business." That was the deliverable. Those 5 words. No specifics. No timeline. No sectors named. 15 chief executives flew to Beijing and received a sentence. Chairman Xi has delivered this sentence at every summit I have staffed. It has not once been followed by a named sector, a timeline, or a specific commitment. It is received as news each time. 43 lobby badges in a Ziploc bag. That's what my team collected from the CEOs after the garden tour. Standard protocol. The badges were embossed with the Great Hall of the People seal. Several executives asked if they could keep them. We said no. One asked twice. 15 executives with combined access to American financial, defense, and technology infrastructure had spent 3 hours inside the Great Hall of the People. We secured the lobby badges. The S&P 500 futures dropped 1% on the morning after the summit. The KOSPI fell 6.12%. China's CSI 300 fell 1.12%. UBS told clients that "much increasingly scarce jet fuel has been burned to produce nothing of real substance." Fortune's headline was "Wall Street sees nothing of real substance." The markets liked the anticipation. The markets did not like the deliverables matrix. Iran was the item we listed as "mutual recognition of shared concerns." The President told reporters they "feel very similar." Xi sat in silence. China's Foreign Ministry did not comment on any commitment regarding the Strait of Hormuz. The President then told reporters the United States "doesn't need the Strait of Hormuz open at all." Oil hit $109 per barrel. Deutsche Bank flagged it as a market-killing statement within the hour. The President described Iran as "a little bit crazy." This was during a toast. Over Peking duck. Rare earths. I prepared a 40-page brief on critical mineral dependency. Supply chain maps for 14 minerals. $1.2 trillion in dependent U.S. industries. Roughly 4% of GDP. The President circled the GDP figure and wrote "big." In the meeting, he asked Chairman Xi if rare earths were "the things in magnets." They are. They are also in every F-35, every Patriot missile battery, and every MRI machine in the country. The discussion lasted 11 minutes. 3 of them were about magnets. No agreement on export licenses. China exposed our dependency last year and has not let us forget it. The Supreme Court struck down our tariffs separately, which was helpful context for the discussions. Fentanyl received 9 minutes. Magnets received 3. We are calling the rare earth outcome "a foundation for continued engagement." There is a poster in the Advance Team office that says "A foundation is not a building." It has been there since my first summit. No one has removed it. On the flight home, my team collected every item the Chinese government had distributed. The credentials. The pins. The keepsakes. The rose seeds Chairman Xi offered for the White House Rose Garden. Standard counterintelligence protocol. All of it went into a bag and off the plane before wheels-up. We threw away the roses. We kept the talking points. The Boeing order grew on the flight home. 500 before departure. 200 in Beijing. 750 somewhere over the Pacific. Boeing had not confirmed 200. The President told reporters on Air Force One it was "a pretty historic couple days." I wrote the line that preceded it: "Tonal reset with significant forward momentum." He used "fantastic" instead. In previous administrations, a tonal reset preceded the deliverables. In this administration, the tonal reset is the deliverable. He has used "fantastic" for every summit since 2017. I have not checked whether the word still polls well. I am told it does. Beijing has not confirmed any of the agreements announced by U.S. officials. This is consistent with the 2017 visit, where $250 billion in deals were announced and an estimated $10 billion materialized. It is consistent with the October summit, where pledges were also made and also not fulfilled. We have a term for this in the Advance Team. We call it "precedent." I have already labeled the binder for 2026. We go back in September. Same matrix. New line items. The verification will be someone else's administration. The President has already asked for the word "monumental." I am told it polls well.

Crazy — another hack just happened! According to @dcfgod, @EchoProtocol_ on Monad was exploited. The hacker: minted 1,000 $eBTC ($76.64M) on Monad; deposited 45 $eBTC ($3.45M) into Curvance; borrowed 11.3 $WBTC ($867K) from Curvance; bridged the 11.3 $WBTC to Ethereum and swapped it for 385 $ETH ($821K); then deposited the 385 $ETH into Tornado Cash to launder the funds. The hacker still holds 955 $eBTC ($73.2M).

Crazy AI shopping stats from @fromedome “shoppers who arrive on merchant sites from AI search have higher purchase intent, convert at nearly 50% higher rates, and spend more — average order values from AI-referred sessions are 14% higher.”