I will explain the original mode. The original mode is a mode that reproduces the same　gore expression, cut scene direction, character design, motion, comic-style character introduction images,　ⅽostumes, achievements, etc. as the original version. For this original mode, which we applied for approval,　there were no requests for revisions from any of the　platformers. This is as I explained before. Next, the differences from the original version are as　follows.First, all of the parameters that have been adjusted for modern action games are different. This includes attack power, movement speed, camera movement range, combo input acceptance time, and initial combo skills. Other changes include the number of remaining shots in the blaster, changes to the behavior of Nick's special moves, the addition of an option to automate QTE, the change of the ranking mode to a time attack that plays through all stages, changes to the lineup of items in the shop, price adjustments, the addition of a chain attack system, image quality, resolution, FPS, and the addition of Yukari Tamura's voice. I have also fixed some of the incorrect implementations in the original version. Specifically, the cutscene that plays when you defeat the Zombie Big Chicken in the psychedelic world of Stage 3 has been changed from a cutscene to a slow-motion scene for defeating the gatekeeper zombie. This change was made to unify the rules for zombie defeat cutscenes. I have also fixed the chainsaw marks so that they appear on all buses, walls, roads, etc., except for plants and trees. Both of these were specifications that I overlooked during the final check of the original version, and I made the corrections in RePOP. We have applied for a Day 1 patch, but the purpose is not to make corrections based on the platformer, but to fix bugs and improve the quality of the ranking mode. #LOLLIPOPCHAINSAWRePOP#

“Hyperliquid”: The "Three" I’ve Finally Found After Years of Search Note: This article offers no investment advice or guidance, but pays tribute to the decentralized philosophy of Laozi and Satoshi Nakamoto. Three years ago, the crypto industry’s uncertainty was still validated by the high volatility of Bitcoin's price, and the DeFi summer driven by Ethereum had faded. When I opened the Tao Te Ching, the foundational scripture of the indigenous Chinese religion, Taoism. I found the power behind cryptocurrency, rooted in Eastern beliefs and philosophy, within the textual context from two thousand years ago.. The "decentralization" from Bitcoin's blockchain technology aligns perfectly with the Taoist ideas of "Wu Wei" and "Dao Fa Zi Ran" (governance through non-action) found in the Tao Te Ching. In 485 BCE, Laozi authored the Tao Te Ching, then left for the West, disappearing without a trace. In 2008 CE, Satoshi Nakamoto anonymously published the Bitcoin whitepaper, launching the first Bitcoin network the following year, eventually decentralizing its management to the community. Over two millennia, both figures disappeared after spreading their teachings, embodying the decentralization philosophy through their absence. The Tao Te Ching says: "By doing nothing, nothing is left undone." Satoshi Nakamoto says: "A truly peer-to-peer electronic cash system should allow online payments to be sent directly from one party to another without the need for a financial institution." This non-intervention aligns with the principles of Wu Wei, where Bitcoin’s market value has grown from zero to a $2 trillion asset over 15 years.The Bitcoin system operates through non-action, yet governs all without interference; it does what is uncontentious, yet nothing can challenge it. Subsequently, decentralized autonomous organizations (DAOs) in smart contracts emerged, following the same 'non-intervention' approach as the Bitcoin system. Interestingly, their abbreviation, DAO, coincides with the pinyin of the Chinese word 'Dao,' embodying the brilliant essence of 'The Dao that can be told is not the eternal Dao.' The Tao Te Ching also says: "The Way (Dao) follows nature." Natural laws, including decentralization, cannot be altered by human will. Just like the running of wind, rain, thunder, and lightning, Bitcoin's system operates autonomously through its code, neither good nor bad. The Tao Te Ching says: "Dao produces one, one produces two, two produces three, and three produces all things." In cryptocurrency, "one" is Bitcoin, the decentralized "Dao" producing peer-to-peer blockchain technology. "Two" is Ethereum,the peer-to-peer blockchain technology has evolved into a decentralized application system with smart contract functionality, which is expected to develop into a global decentralized computing system in the future. But what is "three"? I once thought that stablecoins, represented by USDT and DAI, were 'the third' because they made cryptocurrency pricing and transactions simple and efficient, allowing everything to flourish. However, I overlooked the exchanges that facilitate the transactions themselves. To this day, exchanges are still dominated by centralized exchanges (CEX), led by Binance. Even during the DeFi summer, where various smart contracts surged with decentralized protocols, decentralized exchanges (DEXs) like Uniswap, driven by AMM (Automated Market Makers) liquidity, emerged. However, due to fragmented liquidity, high latency, slippage, and risks like Permit authorization, they have been limited in widespread adoption, and can only serve as supplements to centralized exchanges—providing liquidity and acting as hubs for some long-tail assets. Even with the V3 iteration moving towards concentrated liquidity, similar to an automated market-making strategy, it has improved liquidity and reduced slippage, but is still mainly used by DeFi enthusiasts, professional market makers, and traders. As of today, Uniswap's TVL (Total Value Locked) is only $6.37 billion, down over 30% from its peak of $10 billion in November 2021. Meanwhile, Binance’s TVL, as shown in the December 2023 Merkle proof of assets, exceeds $100 billion. In terms of volume, Uniswap’s daily volume is $3 billion, while Binance exceeds $100 billion. Whether in terms of TVL or volume, DEXs cannot compete with CEXs. The stagnation of DEX development has directly impacted the growth of decentralized protocols’ TVL, which is an inevitable result. As the development of DEXs falters, assets reliant on CEX trading are not being withdrawn to the blockchain, causing on-chain assets to stagnate (where asset prices rise but TVL declines). Fortunately, the situation is gradually reversing. Over four years of DeFi development, on-chain oracles have become increasingly stable, cross-chain interoperability is becoming more secure, and TPS (transactions per second) on Layer 1 and Layer 2 chains are rising, while ensuring security and decentralization. POS (Proof of Stake) validation is becoming more decentralized, with more native and mapped assets on-chain. Hardware wallets, such as AA wallets, have improved in usability and risk resilience, while infrastructure is becoming more robust. Decentralized protocols and applications are thriving, and the four-year development of smart contracts has cultivated a large user base for decentralized applications. As assets, applications, and users all move towards decentralization, yet the most important liquidity exchange scenarios remain centralized, this is far from truly decentralized. Then came HYPE (Hyperliquid), and it seemed that the 'third' I had been searching for all these years was confirmed and validated the moment I discovered it. The weight I had once placed on stablecoins has also shattered. The 'third' I had been pursuing, the one that could enable large-scale adoption, was always the DEX capable of achieving this—before HYPE appeared, DEXs were merely optional supplements. But after HYPE emerged, it introduced a high-performance EVM chain designed for financial transactions, a Layer 1 product component with low latency and high throughput, creating a DEX with an on-chain order book that rivals CEXs. It has been running smoothly for over a year and a half, even during peak trading periods, ensuring a low-latency, high-performance trading experience. Large-scale adoption has already been proven by time, and its zero-incident reputation has attracted a large number of real users. Even without token rewards or incentives, users, TVL, and volume have all continued to grow steadily. Before the mainnet launch, the TVL, based solely on USDC deposits, reached $1.2 billion. The project team is low-key, humble, and not greedy, focusing solely on the product itself, with no VC investment or promotional campaigns. Word-of-mouth and user referrals have been the only drivers. The token distribution (TGE) is entirely oriented towards benefiting real users. This style, almost akin to the original ethos of Bitcoin, is even more focused on user-centricity than Ethereum or Bitcoin itself. It can be foreseen that when HYPE's mainnet goes live, with native and mapped assets executing simultaneously, mainstream asset spot trading and perpetual futures with joint margin trading will be launched. TVL will quickly surpass $10 billion, triggering a positive flywheel effect. Both TVL and volume will surge, outpacing all others. A large number of market makers, professional investors, and users will bring their capital on-chain for long-term involvement. Centralized exchanges (CEXs), for strategic reasons, will be forced to inject liquidity and invest in their tokens to secure some degree of pricing power. A variety of decentralized protocols will flourish after the HYPE mainnet launch, including decentralized lending, stablecoins, staking, restaking, and RWA (Real World Assets) protocols. The entire DeFi market will benefit from the irreversible shift towards decentralized on-chain trading that HYPE will drive, particularly decentralized lending platforms like AAVE and stablecoin DEXs like CRV. As on-chain assets and transactions grow, lending derivatives and more frequent stablecoin swaps will follow. For other DEXs of similar types, such as DYDX, unless they find a differentiated path to evolve their products, their TVL and volume will be continually suppressed until they collapse. Uniswap, as an AMM-based market maker for spot trading, will initially benefit, but as HYPE's spot trading area improves, its growth will be hindered. However, AMM liquidity will still have a long-term market position as supplementary liquidity for order books and a venue for long-tail asset trading. The biggest beneficiaries will be AI. The number of trading strategy AI models will rapidly increase alongside HYPE's growth. Various types of AI will be able to freely trade with different strategies on HYPE without worrying about CEX asset freezes or withdrawal issues. At present, some users may mock the idea of HYPE becoming the Binance of the blockchain as a joke. However, years from now, they will only remember that Binance was the off-chain HYPE. The deconstruction of CEXs and the reconstruction of DEXs is quietly taking place in this winter. There is no longer the brilliance of DeFi's Summer, only the quiet beauty of CEXs, which are now being stared down by death. If my judgment is wrong, it will only be because HYPE has failed to fulfill its mission as a DEX. However, in the future, there will be one or even multiple 'HYPEs' that will carry on this vision and complete the irreversible revolution in the cryptocurrency era. And I, too, will eventually find the missing 'third' in the crypto faith that has been absent for so many years—the 'three' that gives birth to all things. @HyperliquidX @HyperFND @chameleon_jeff

Today is the start of a very important day in the AI world... the trial between Elon and Sam/OpenAI begins today in Oakland Federal Court. The jury selection is happening right now in the U.S. District Court for the Northern District of California and opening arguments are expected Tuesday. The civil jury trial is projected to last 2-4 weeks, with Judge Yvonne Gonzalez Rogers making the final call. So let me tell you why this fight all came about. 1/ In 2015 Elon co-founded OpenAI with Sam Altman and others as a nonprofit to develop artificial general intelligence safely and openly for the benefit of all humanity... this was supposed to NOT be for profit or for closed corporate control. Elon was a major early funder, contributing around $38-44 million (~60% of early seed funding) plus resources like compute and recruiting. 2/ Then, Elon left the board in 2018 over disagreements for the direction of the company and Microsoft’s growing role of the company. 3/ Later, OpenAI took billions from Microsoft, restructured with a for-profit arm, initially “capped profit,” now more commercial... went closed-source in practice, and exploded in value with ChatGPT. FYI, the current valuation of OpenAI now sits at ~$852 billion and the company recently completed restructuring with a for-profit entity reporting into a nonprofit foundation. Now, with the trial, Elon is saying Altman, Brockman, and OpenAI breached the founding charitable trust and agreement by turning it into a “wealth machine” that prioritizes profits and insiders over the original mission. He claims they deceived him about their plans... This lawsuit was originally filed in November 2024, was withdrawn, and then revived in early 2026. And just this Friday, April 24, Elon voluntarily dropped the fraud claims to “streamline” the case and keep the jury focused on the mission issue... proceeding on breach of charitable trust and unjust enrichment. This is what Elon is looking to get back from Sam and OpenAI: a/ substantial damages, with stakes in the $100B+ range, with many reports saying it's in the range of ~$134B and the winnings will be given to the nonprofit/charity arm and will NOT benefit Elon personally b/ possible unwinding/restructuring to restore the original nonprofit mission c/ leadership changes (e.g., getting rid of Altman/Brockman from key roles). On the other side, OpenAI is claiming: a/ Elon knew about and once supported commercialization steps, saying he even explored merging with Tesla or gaining control himself b/ OpenAI calls the suit competitive sabotage from rival Elon's company xAI and says this is driven by jealousy over OpenAI’s success This is SUCH an important trial in the world of AI... people may not fully understand. The reason is bc Elon has repeatedly warned that profit-driven, closed-source AGI is very dangerous. This is his chance to enforce the original “benefit humanity” mission he helped create and walked away from in 2018. A big win will hurt OpenAI’s valuation, upcoming IPO plans, Microsoft partnership, and market dominance/customer perspective of the company... and this will also force real changes in how the world’s leading AI lab operates, influencing the entire AI race and future regulations. For me, Sam Altman straight up CANNOT be trusted with the future of AGI, and this trial proves why Elon was right to fight it. This is the same guy who got fired by his own board in 2023 for not being “consistently candid,” then crawled back in and turned the nonprofit “open for humanity” promise into a closed-source, Microsoft-bankrolled profit machine worth hundreds of billions. Elon was the one who put up the early cash and vision to keep AI safe and beneficial for all of us... not to create a trillion-dollar insider club. If Altman gets away with rewriting the rules after the fact, it sets a dangerous precedent that mission-driven tech is just marketing fluff. I believe Elon isn’t doing this for ego or rivalry... he’s really doing it bc someone has to hold the line before profit-over-people AGI becomes unstoppable. This one’s personal for the man who actually wants to understand the universe instead of just cashing in on it. Humanity needs Elon to win. And that's who I'm rooting for!

Battle of Berlin Given the shortage of troops, most defenders came from the Volkssturm "Popular Militia" and the Hitlerjügend, both structures of questionable military value. The official number of defenders was counted at 41,200, of which only about 15,000 had a certain degree of military instruction. Their armament was quite scarce, possessing only 42,000 rifles, 2,200 machine guns (less than 200 were MG 42), 140 mortars, 15 PAK 40 and eight 8.8 cm anti-tanks. The main artillery unit was the 230 8.8cm, 43 10.5 and 12.8 anti-aircraft cannons of the 1st Flak Division, but many of them could not be used against ground targets. The rest of the artillery consisted of a heterogeneous mix of cannons of multiple origin, of use when less relative. For the final battle, just over 20,000 regular soldiers (between Werhrmacht and Waffen SS) would reinforce the original garrison, coming from various units retreating towards the capital ahead of the Soviet advance. Berlin was divided into eight sectors, with numerous defenses of which only those relied on water courses, such as the Spree and the Landwehr Canal, were effective. The defenders lacked a determined and unique command, in practice. For example, of the 483 bridges in the city, only 127 were destroyed, because Speer considered that, by destroying all of them, to prevent the passage of Soviet cars, the cut of electricity, water and gas that would be produced would add greater suffering to the civilian population. Notes taken from the book: "The Battle of Berlin 1945" by Juan Vazquez Garcia. 📖 Photography: 📷 Members of the Volkssturm captured during the Battle of Berlin. Interestingly enough, the soldier in the foreground has minimal gear on. He appears to be wearing a captured French belt, hanging from which a typical British (or Soviet or civilian) mug and a German cantimpla. The coat seems to be lacking railings, he is wearing a DAF (German Labour Front) cap. Germany 1945. Photographer: 📷 Anatoly Arkhipov. (FGF Colourised)

A profit-driven enforcement assembly line led by Li Renye Beyond collaborating with Justin Sun on two separate cases, the enforcement team headed by Li Renye has allegedly targeted a number of prominent figures in the crypto space. Further details are still being compiled. Part 1 Correction to a previously circulated claim: The individual currently described as the “blade” in Justin Sun’s hand, Li Renye, has never held a position within mainland China’s police, procuratorate, or court system, either in the past or at present. Media reports referencing his supposed prior official roles have, intentionally or not, created a misleading impression for most readers. Li Renye, born on October 4, 1996, previously held only the following position: •Mediator at the Dispute Resolution Center in Yongjia County, Wenzhou, Zhejiang Province, mainland China •This role was temporary, contract-based, and outside the formal government staffing system. Part 2 According to individuals familiar with the matter: 1Li Renye allegedly obtained money from multiple parties by promising assistance that never materialized, and refused to return the funds. 2He reportedly has a prior record of being summoned and detained by mainland Chinese authorities in connection with financial fraud. 3More critically, he is alleged to have impersonated personnel from China’s law enforcement or judicial agencies. “Prosecutor” Li, did it feel empowering to fabricate an identity as a member of China’s judicial system? Part 3 At a later stage, Li Renye reportedly crossed paths with Justin Sun in what insiders describe as a complementary arrangement: 
Li provided operational muscle, while Sun provided financial backing. From that point on, Li became what observers characterize as a rapid-response enforcement instrument in Sun’s toolkit, capable of pushing aggressive law enforcement actions across multiple jurisdictions. As previously noted, the case involving
Wang Bingyu who is said to have close ties to Justin Sun, allegedly involved substantial participation by Li Renye. During police interrogation, Wang Bingyu and an individual referred to as “Accountant Guo” reportedly surrendered more than 100 million RMB. 
Yet, the ultimate destination of those funds remains unclear. “Prosecutor” Li,it has been rumored that 32 million RMB from that surrendered amount was divided between you and certain associates. Care to clarify? According to this narrative,
@justinsuntron Li Renye, and Li Tao ,
have effectively turned law enforcement actions in mainland China into what critics describe as a fully integrated, profit-driven enforcement pipeline, referencing both the Wang Bingyu case and the @WilsonWei777 case. Anyone with information regarding cases connected to these individuals is encouraged, in the original statement, to submit details for consolidated disclosure. And finally, a closing jab from the original text: “Prosecutor” Li, how comfortable is the ride on Justin Sun’s private jet, the
Tron 1? ✈️ @BBCWorld @BBCBreaking @WSJ @business @nytimes @cnni @Reuters @Forbes @TIME @TheEconomist @UN @AP @washingtonpost @MarketWatch @WSJecon @FAANews @NTSB_Newsroom @FoxNews @FT @YahooFinance @SkyNews  @NBCNews @thejusticedept @fincennews  @ukhomeoffice @nca_uk @govuk @ica_singapore @govsingapore @mfasg

Coding agents are accelerating different types of software work to different degrees. When we architect teams, understanding these distinctions helps us to have realistic expectations. Listing functions from most accelerated to least, my order is: frontend development, backend, infrastructure, and research. Frontend development — say, building a web page to serve descriptions of products for an ecommerce site — is dramatically sped up because coding agents are fluent in popular frontend languages like TypeScript and JavaScript and frameworks like React and Angular. Additionally, by examining what they have built by operating a web browser, coding agents are now very good at closing the loop and iterating on their own implementations. Granted, LLMs today are still weak at visual design, but given a design (or if a polished design isn’t important), the implementation is fast! Backend development — say, building APIs to respond to queries requesting product data — is harder. It takes more work by human developers to steer modern models to think through corner cases that might lead to subtle bugs or security flaws. Further, a backend bug can lead to non-intuitive downstream effects like a corrupted database that occasionally returns incorrect results, which can be harder to debug than a typical frontend bug. Finally, although database migrations can be easier with coding agents, they’re still hard and need to be handled carefully to prevent data loss. While backend development is much faster with coding agents, they accelerate it less, and skilled developers still design and implement far better backends than inexperienced ones who use coding agents. Infrastructure. Agents are even less effective in tasks like scaling an ecommerce site to 10K active uses while maintaining 99.99% reliability. LLMs' knowledge is still relatively limited with respect to infrastructure and the complex tradeoffs good engineers must make, so I rarely trust them for critical infra decisions. Building good infrastructure often requires a period of testing and experimentation, and coding agents can help with that, but ultimately that’s a significant bottleneck where fast AI coding does not help much. Lastly, finding infrastructure bugs — say, a subtle network misconfiguration — can be incredibly difficult and requires deep engineering expertise. Thus, I’ve found that coding agents accelerate critical infrastructure even less than backend development. Research. Coding agents accelerate research work even less. Research involves thinking through new ideas, formulating hypotheses, running experiments, interpreting them to potentially modify the hypotheses, and iterating until we reach conclusions. Coding agents can speed up the pace at which we can write research code. (I also use coding agents to help me orchestrate and keep track of experiments, which makes it easier for a single researcher to manage more experiments.) But there is a lot of work in research other than coding, and today’s agents help with research only marginally. Categorizing software work into frontend, backend, infra, and research is an extreme simplification, but having a simple mental model for how much different tasks have sped up has been useful for how I organize software teams. For example, I now ask front-end teams to implement products dramatically faster than a year ago, but my expectations for research teams have not shifted nearly as much. I am fascinated by how to organize software teams to use coding agents to achieve speed, and will keep sharing my findings in future posts. [Original text: ]

I am the Health Services Contract Administrator for U.S. Immigration and Customs Enforcement and on October 3, 2025, I stopped paying for medical care, and the system has been working better ever since. I need to be specific about what "working better" means in this context because I am a contract administrator and specificity is my entire professional identity. It means: fewer invoices. Fewer reimbursement disputes. Fewer pharmacy reconciliations. Fewer appeals from providers who claim they provided urgent care and want to be compensated for having provided it. The workflow is cleaner. My inbox is lighter. The VA Financial Services Center, which had processed our medical claims since 2002, twenty-two years of pharmacy invoices, specialist referrals, hospital transports, dialysis authorizations, and oncology treatment plans, terminated its interagency agreement with us after a lawsuit from a nonprofit called the Center to Advance Security in America. They filed suit. The VA pulled out. Our entire claims processing pipeline vanished overnight. I posted a notice on on November 12 describing this as an "absolute emergency" that required resolution "immediately" to "prevent any further medical complications or loss of life." That was seven months ago. The replacement contractor, Acentra Health, had not achieved notice to proceed by the April 30 period of performance deadline. As of today, no entity is processing medical reimbursement claims for ICE detainees in the United States. When I say "no entity" I mean that structurally. A person held in a GEO Group facility in Georgia who requires dialysis three times per week is receiving dialysis from a provider who has not been paid since October. The provider continues to provide care because the alternative is that the confined person dies in their facility and the facility is then liable for a death that could have been prevented by a treatment that the facility was contractually obligated to provide. The treatment continues. The payment does not. The provider absorbs the cost. The cost is eventually written off. The write-off appears in the provider's quarterly financial statements as "uncompensated care, federal detention." It does not appear in our budget. It does not appear in any ICE financial disclosure. The care happened. The cost was real. The payment was imaginary. The system is working better. In fiscal year 2024, the VA processed $246.42 million in clinical reimbursement claims on our behalf. In fiscal year 2025, despite an 82.5% increase in our daily detained headcount, the VA processed only $157.2 million before the October termination. The delta between what was needed and what was processed is approximately $300 million. That $300 million represents medications not reimbursed, specialist consultations not paid for, emergency transports not covered, prenatal visits not compensated. It represents chemotherapy sessions where the drugs were administered and the oncologist submitted an invoice and the invoice entered a system that no longer exists. I have a filing cabinet in my office — three drawers, GSA-standard, beige, the kind with the lock that everyone has the same key to — that contains printed copies of the final VA-processed claims from September 2025. The bottom drawer has a jar of Tums that my predecessor left when she transferred to FEMA in August. I eat them daily. Not from stress. From the cafeteria. The cafeteria serves a chili that the facilities contractor, Aramark, describes as "Southwestern-inspired." It is inspired by the Southwest the way our medical payment system is inspired by the concept of paying for medical care. The death rate is the number people ask about, so I will provide it with the precision my role requires. Historical baseline, 2018 through 2024: 8.9 deaths per year in ICE custody. Calendar year 2025: 33 deaths. Twelve of those occurred after October 3, after the payment freeze. January through April 2026: 17 deaths. That is one death every six days. Annualized, the current rate is 51.7 deaths per year. 5.8 times the pre-October baseline. A study published in JAMA on April 16 calculated the per-capita rate: 88.9 deaths per 100,000 person-years in partial fiscal 2026, compared to 13.0 in fiscal 2023. Nearly seven times. The JAMA authors are epidemiologists. I am a contract administrator. We are counting the same bodies with different denominators. Emmanuel Damas was 56 years old, Haitian, confined at an installation I am not authorized to name. He had a tooth infection. The on-call clinical staff treated the infection with ibuprofen. Ibuprofen is an anti-inflammatory. A tooth infection is a bacterial event. These are different categories of medical problem requiring different categories of intervention. The infection progressed to septic shock. Emmanuel Damas died. The ibuprofen was on our formulary. Antibiotics were on our formulary. The difference between the two was a reimbursement claim that would have been submitted to a payment processor that no longer existed. The detention center chose the treatment that did not generate a claim. I cannot tell you whether that decision was made consciously. I can tell you that it was made consistently. Across multiple facilities. Across multiple months. The ACLU reviewed deaths in ICE detention between 2017 and 2021, before the payment freeze, and determined that 95% were preventable with adequate treatment. I do not know what the percentage is now. I suspect it is also 95%. The category "preventable" has not changed. The category "payment" has. At Fort Bliss, a military installation in El Paso repurposed as a detention facility under a $1.24 billion sole-source contract awarded to Acquisition Logistics, a firm with no prior detention management experience, three people died within 44 days. One death was ruled a homicide by the El Paso County Medical Examiner. ICE reported it as a suicide. Those are different words describing different events with different legal implications. The Medical Examiner's ruling generates an investigation. A suicide generates a compliance review. An investigation involves law enforcement. A compliance review involves my filing cabinet. I am not qualified to determine which word is correct. I am qualified to tell you that the words produce different paperwork, and the paperwork determines which systems activate, and the systems that activate determine who is accountable, and in this case, the system that activated was the compliance review, and the compliance review found that all protocols were followed, and all protocols were followed because the protocols do not include "pay for medical care." Rodney Taylor was a double amputee detained at Stewart Detention Center, operated by CoreCivic. He was forced to crawl on floors covered in feces and mold because the center did not provide adequate mobility assistance. CoreCivic reported $2.2 billion in revenue last year, up 13%. Their profit was $116.5 million, up 70% year over year. Their ICE revenue nearly doubled between Q4 2024 and Q4 2025, from $120 million to $245 million per quarter. They received a 70% increase in profit and Rodney Taylor received a floor. CoreCivic's annual report describes their business model as "government solutions." Rodney Taylor's experience was, technically, a government solution. GEO Group, the other major for-profit detention operator, posted $2.6 billion in revenue in 2025 and $254 million in profit, a 700% increase. They secured $520 million in new ICE task orders that year. Combined, GEO and CoreCivic spent $6.8 million on lobbying to secure access to a $75 billion funding stream from the GOP's reconciliation bill. The return on that investment is so large I had to check my calculator twice. It was not a calculator error. It was the normal functioning of a procurement system where the companies that run the facilities also fund the campaigns of the legislators who appropriate the money for the facilities. The firm-fixed-price task orders specify a per diem rate of $187.48 per adult per day. That rate includes healthcare coverage. The rate has not changed since the disbursement freeze. We are still remitting $187.48 per day per person. The clinicians are not receiving any of it. The $187.48 goes to the facility operator. The operator is supposed to allocate a portion of it for clinical services. There is no SLA enforcement mechanism to verify that they do. There is only my filing cabinet, and the filing cabinet is for contracts, not outcomes. Senator Ossoff's office conducted an investigation between January and August 2025 and received 85 credible reports of medical neglect, including untreated chest pain causing heart attacks and unmanaged diabetes complications. That investigation preceded the payment freeze by two months. The conditions it documented were the baseline. The baseline was already 95% preventable death. The disbursement freeze removed the financial infrastructure supporting the 5% of care that was being provided. I have a Gantt chart in my office, printed on 11x17 cardstock and laminated and pinned above the Tums drawer, that tracks the Acentra Health onboarding timeline. The original completion date was April 30, 2026. That date passed twelve days ago. The chart has a red line through it drawn in Sharpie by my deputy, who does this for every missed milestone. There are four red lines. There will be more. Each red line represents a period during which no payment processor exists. Each period without a payment processor is a period during which clinicians must choose between providing unpaid care and not providing care. The first option costs them money. The second option costs someone their life. I do not track which choice they make. I track contracts. Seventy-one percent of ICE deaths in 2025 and 2026 occurred in privately operated detention sites. Half of 2026's deaths occurred in CoreCivic or GEO Group facilities. The Office of Detention Oversight, the COR entity responsible for facility inspections, conducted 36.25% fewer compliance audits in 2025 than the previous year. Fewer audits, more deaths, higher profits. The three trend lines move in coordinated directions. I do not draw conclusions from correlated trend lines. I am a contract administrator. I process contracts. The contracts are technically valid. The facilities are technically operational. The reimbursement apparatus is technically being replaced. The deaths are technically being counted. The word "technically" is doing more work in this paragraph than any clinician in the ICE detention system has been compensated for in seven months. My internal memo from November 12 used the phrase "absolute emergency." It recommended resolution "immediately" to "prevent any further medical complications or loss of life." That memo was written on government letterhead, classified as internal correspondence, distributed to eleven recipients, and filed in the correspondence tracking system under routing symbol HSA-OAQ, which requires a FOIA request to access. Seventeen people have died since I wrote it. The memo was technically effective. It generated a procurement action. The procurement action generated a bridge contract. The bridge contract generated an onboarding timeline. The onboarding timeline generated a Gantt chart. The Gantt chart generated four red Sharpie lines. The red Sharpie lines generated nothing. They are decorative. Like the per diem rate that includes medical care nobody is billing for. Like the 95% preventable death rate that is not being prevented. Like the word "emergency" in a seven-month-old memo that is technically still active, technically still urgent, technically still describing a situation that requires immediate resolution. I am technically still the person responsible for resolving it. The system is technically still working. The people are technically still dying. The filing cabinet is technically still organized. The contracts are technically still valid. The word "technically" has appeared so many times in this document that it has lost all meaning. That is exactly what it was designed to do.

I am the Lead Settlement Counsel in the Civil Division of the Department of Justice, assigned to *Trump v. Internal Revenue Service*, Case No. 1:26-cv-00147. My job is to represent the government against the plaintiff. The Attorney General, who represented the plaintiff before she represented the government, assigned me personally. I keep a laminated seating chart in my top drawer. It maps who in this building used to sit across the table from me. Three of the top four names in the Department previously represented the man I am now tasked with opposing. I initial the chart quarterly. In blue pen for active conflicts. I ran out of blue ink in February. The plaintiff is seeking ten billion dollars. Ten. Billion. He paid $750 in federal income tax in the year he was elected. Seven hundred fifty. I have paid more for parking violations in the District. He paid zero in ten of the fifteen years before that. These are the returns that were leaked. The leak is the crime. The returns are evidence of good citizenship. This is how settlement works. The man who leaked the returns, Charles Littlejohn, a contractor, is currently serving a 5-year federal prison sentence. He disclosed that the President of the United States paid less in taxes than a part-time crossing guard. For this, he is in a cell. For the returns themselves, for what they revealed about a system designed to collect from people who cannot afford attorneys and forgive those who can, there is no case number. There is no docket. There is no plaintiff. That information simply exists now, and we are here to make it expensive. Ten billion divided by one hundred million taxpayers. That's one hundred dollars per household. You will pay approximately one hundred dollars to compensate a man for the emotional distress of the public learning he paid less than you did. In legal terms, this is called "damages." In structural terms, it is called Tuesday. This is how settlement works. The settlement term currently under discussion includes a provision that the IRS will drop all active and future audits of the plaintiff, his family members, and his business entities. Permanently. An enforcement agency will agree, in writing, to stop enforcing. I have a Post-it on my monitor that says AUDIT IMMUNITY — CONFIRM SCOPE. It has been there for nine weeks. No one has asked me to remove it. Attorney General Bondi represented the plaintiff privately before she took office. Deputy Attorney General Todd Blanche represented him in his criminal trial. The number-three official, Stanley Woodley, represented him in the classified documents case. I am, technically, the adversary. I sit in the same building as three of his former personal attorneys. I take my lunch at the same cafeteria. I use the same badge to enter the same elevator. The Attorney General fired the Department's chief ethics officer on her fourth day. The position has not been refilled. I submitted a conflict-of-interest disclosure in January. It was received. The word "received" is doing considerable work in that sentence. This is how settlement works. The plaintiff has stated publicly, and I am quoting the public record, "I've gotta make a deal. I negotiate with myself." This was not presented as a metaphor. Judge Kathleen Williams has ordered both parties to explain, by May 20th, whether they are in conflict. I am drafting the government's response. The plaintiff's former attorneys, my supervisors, will review it. The plaintiff has pledged to donate any settlement proceeds to charity. I should note for the record that the Washington Post documented that the plaintiff donated less than $10,000 over seven years, during a period when he publicly claimed millions. His charitable foundation, the Trump Foundation, was dissolved by court order in New York in 2019 for self-dealing. The words "to charity" appear on page four of the term sheet. They are not defined. I have not been instructed to define them. We have already disbursed $8.5 million in adjacent settlements. Michael Flynn received over one million. Carter Page received one point two five million. The Babbitt family received five million. 450 January 6th defendants have filed compensation claims. The pipeline is active. The precedent is operational. I track disbursements on a spreadsheet I titled RESOLUTION LEDGER. It auto-sorts by amount. The President's ten billion would require me to adjust the column width. I want to note one final detail, because the file demands it. The leak of the tax returns occurred during the plaintiff's first term. He appointed the IRS commissioner. He oversaw the Treasury Department. The negligence he is suing for occurred under his own management. He is suing the government he ran for the failures he administered. In the margins of the original complaint, someone wrote "beautiful" in pencil. I will not speculate who. This is how settlement works. You file your taxes every April. You are audited if the numbers don't match. You pay penalties. You pay interest. You pay what you owe, and sometimes more, and sometimes for years. The plaintiff paid seven hundred and fifty dollars. Someone told you. That person went to prison. And now, because you found out, because the information became public, because a contractor decided the country should know what the country was owed, you will pay one hundred dollars to the man who owed it. The settlement is on my desk. Both sides have agreed. I represent one of them. My boss used to represent the other. The ethics officer has been dismissed. The judge wants to know if the plaintiff and the defendant are the same person. I am reviewing the question. The math checks out.

Defi is a central part of the value that Ethereum provides. Financial empowerment is a central part of what it means to have agency and freedom in our current world. Finance is far from the only thing that Ethereum is good for, but it is an important thing. This post discusses how the Ethereum Foundation is approaching defi. Defi today makes the world's best savings, risk management and wealth-building opportunities permissionlessly available worldwide. We need to build on that. Ethereum's early defi era was great because it dared to dream and innovate and come up with totally new paradigms (eg. AMMs). Defi tomorrow will bring back that spirit. Don't just "make a better stablecoin", dig a layer deeper, and think about the underlying problem (risk management, hedging one's future expenses), and come up with an even better solution. But also, as the EF, we are not interested in supporting "onchain finance" or even "defi" indiscriminately. We have a specific vision of what we want to see out of defi: permissionless, open-source, private, security-first global finance that maximizes people's control over their own assets, minimizes centralized chokepoints and trusted third parties, and democratizes risk management and wealth building (the two key goals of finance according to modern portfolio theory) as well as payments. We want protocols that pass the walkaway test: that keep working even if the original team suddenly disappears without warning (or even: becomes hostile / compromised without warning). Bringing this vision to reality will inevitably take a lot of work. Defi is a complex toolchain, including various onchain components, user-side offchain components (ie. wallet, local agent...), other offchain components, etc. The things that we care about include areas like: * Improving security of defi through "traditional" means, eg. audits, standards, wallet-side safeguards * Improving security of defi through "new" means, eg. AI-assisted formal verification, user-side agents as safeguards * Oracle security and decentralization (there's A LOT of skeletons in the closet here, we as an ecosystem really need to point a big eye of sauron at it for a while) * Privacy. Both privacy-preserving payments, and privacy of more complex use cases (eg. what does it mean to have a maximally privacy-preserving CDP? there are clearly benefits in reducing liquidation-sniping risk, but it requires hard tech to get there) * Open source, and improving the licensing / forkability situation in defi Ethereum is a permissionless protocol, and nothing stops people from deploying insecure protocols, protocols that enshrine ultimately unneeded centralized trust in the name of convenience, or dopamine-maximizing gambleslop. However, we *are* interested in working with anyone aligned to make permissionless, open-source, intermediary-minimizing and security and user-agency-maximizing defi ecosystem as strong as possible, so that it can be not just individuals and institutions' first choice in Ethereum, but also a globally compelling way to manage funds for anyone who needs its properties.