Aakash Gupta (@aakashgupta)

A Chinese delivery rider in this video earns $4 a day. He tips $4 to a livestreamer. Douyin takes $2. Her agency takes $1. She nets $1 after eight hours on camera. She orders $1 of food through Meituan. Meituan takes 20%. The next courier earns 50 cents on the delivery. $4 of wages becomes 50 cents of wages in one rotation. Run it twice and the loop is dead. That's how the "self-sustaining cycle" actually works. Three rent-collectors at every node. Every rotation extracts more from the workers and pays the platforms more. Meituan and employ 11 million riders. Meituan's 7.45 million collectively pulled 80 billion yuan in 2023, roughly $1,500 each per year. Douyin and Kuaishou keep 50% of every virtual gift before the streamer sees a yuan. Recruitment agencies take another 20-25% off whatever's left. Meituan booked $40 billion in revenue last year. ByteDance booked $155 billion. The 11 million couriers and the 660 million livestream viewers are the kinetic energy. Every tip, every delivery, every minute watched converts directly into platform commission, ad inventory, or data. China's youth unemployment hit 21.3% in June 2023. The government stopped publishing the number the following month. The "world's strangest job cycle" is the official rebranding of what happens when 20% of a generation gets locked out of formal employment and three platforms figure out how to monetize the leftover labor.

Your brain has a circuit that doesn't know you live in a city. Its only job is to monitor whether birds are still singing. Right now, in this room, it is on. The circuit predates primates. Mammals have been using ambient soundscape continuity as a predator-detection system for roughly 200 million years. Birds stop singing when something larger moves through their territory. For most of mammalian history, a forest full of song meant no large predator was nearby, and the cessation of sound was the warning. Your nervous system never updated this software. The Max Planck Institute tested the inverse in 2022 with 295 participants. Six minutes of birdsong dropped anxiety with a medium effect size. Six minutes of traffic noise raised depression with the same. The effect worked on subjects who lived in dense urban environments and had no regular contact with nature. The brain still ran the check. Birdsong sits in the 1,000 to 8,000 Hz range. Your brainstem reads continuous patterns in that band as a signal that nothing dangerous is currently moving through the environment. EEG data shows birdsong at 45 to 50 decibels boosts alpha wave activity by 14.1% relative to silence. Alpha is the brainwave signature of relaxed alertness. Push the same birdsong above 60 decibels and the response flips. Stress markers rise 29%. The circuit only trusts the signal at the volume of quiet conversation, which is exactly the volume birds sing at from a typical distance. Three things happen simultaneously when the brain registers ambient safety. The amygdala downregulates. The parasympathetic nervous system takes over from the sympathetic. Heart rate variability rises, cortisol drops. The posterior cingulate cortex, which sits at the center of the rumination circuit, quiets down. King's College London tracked this through a smartphone study with over 1,200 participants and found the mood lift lasted hours after the sound stopped. People diagnosed with depression got the same response as healthy controls. Most of what gets labeled mental fatigue is hypervigilance running in the background. Birdsong tells the circuit it can stand down, and the brain reallocates the freed compute everywhere else. A quiet park feels different from a quiet office because the parks have sentinels.

BART spent $90 million on new fare gates. They're recovering about $10 million a year in fares. That's a 9-year payback on paper. The actual return hit in six months. Embarcadero station went from 112 hours of corrective maintenance in the six months before installation to 2 hours after. Daly City saved 109. Balboa Park saved 75. Across the system, 961 hours of cleanup work disappeared. Corrective maintenance is the term BART uses for graffiti, heavy soiling, vandalism, the damage that needs a crew not a janitor. At several stations it dropped to zero. Crime fell 41% year over year. Riders who reported seeing fare evasion on their trip dropped from 22% to 10%. Citations issued by BART police went from 2,200 in January to under 1,000 in July, because there was nothing to cite. The gates were a filtering project disguised as a revenue project. Old BART gates were waist-high orange fins designed in the 1970s. You could hop them in under a second. That made the station effectively a public space, and the rider mix reflected that. The new gates are 72 inches of polycarbonate with 3D sensors that detect tailgating. You either pay or you don't enter. Once you don't enter, you also don't smoke on the platform, sleep in the elevator, or harass other riders. BART tried hiring more police for years. Blitz operations at high-traffic stations. Increased patrols. Dedicated transit cops. None of it moved the numbers the way six feet of polycarbonate did. The $10 million in recovered fares is the smallest line in the return. Fare revenue used to cover 70% of BART operations. After the pandemic it collapsed to 22%. The gates won't fix that gap directly. They fix the precondition for fixing it: a system that office workers, families, and tourists are willing to use again. Ridership growth at stations with new gates outpaced ungated ones before the rollout finished. A $400 million annual deficit is heading to voters in November as a sales tax measure. Voters don't approve sales taxes for transit agencies they don't feel safe in. The $90 million on gates is buying BART the right to ask the public for more money. That's the real return on six feet of polycarbonate.

Three products killed in seven days. That's the number that tells you where OpenAI actually is right now. Monday: Instant Checkout scrapped. Only 12 Shopify merchants ever went live. Walmart said conversion rates inside ChatGPT were 3x lower than on their own site. Six months of "agentic commerce" produced almost nothing. Tuesday: Sora shut down entirely. Disney's $1 billion investment collapsed before any money changed hands. Their teams were working on Sora projects Monday evening and got blindsided by the announcement 30 minutes later. Downloads had already fallen 32% month over month by December. Thursday: Erotic chatbot shelved indefinitely. Internally called "Citron mode." They couldn't train models that previously avoided explicit content to reliably exclude illegal behavior. A senior employee quit over it. Their age-verification system has a 10%+ error rate. Now look at the financials behind these decisions. OpenAI hit $25 billion in annualized revenue in February. They're projecting $14 billion in losses for 2026 and $17 billion in cash burn. The IPO is targeting Q4 2026 at an $840 billion valuation. They need to file an S-1 in months. Every one of these killed products was a liability on that S-1. E-commerce checkout with no tax compliance infrastructure. A video app burning compute with falling downloads. An adult chatbot while the FTC is investigating AI harm to minors and Meta just got hit with $375 million in a child exploitation case. This is what pre-IPO cleanup looks like at $840 billion. You kill everything that creates a headline risk, consolidate into a "superapp" that combines ChatGPT, Codex, and Atlas, and pray the coding market is big enough to justify 65x revenue. The Pentagon contract was the tell. OpenAI rushed a $200 million defense deal the same day Anthropic got blacklisted, admitted it was sloppy, then spent a week rewriting the terms. That's a company optimizing for one thing: making the investor deck look inevitable before the roadshow starts.

Cursor is raising at a $50 billion valuation on the claim that its “in-house models generate more code than almost any other LLMs in the world.” Less than 24 hours after launching Composer 2, a developer found the model ID in the API response: kimi-k2p5-rl-0317-s515-fast. That’s Moonshot AI’s Kimi K2.5 with reinforcement learning appended. A developer named Fynn was testing Cursor’s OpenAI-compatible base URL when the identifier leaked through the response headers. Moonshot’s head of pretraining, Yulun Du, confirmed on X that the tokenizer is identical to Kimi’s and questioned Cursor’s license compliance. Two other Moonshot employees posted confirmations. All three posts have since been deleted. This is the second time. When Cursor launched Composer 1 in October 2025, users across multiple countries reported the model spontaneously switching its inner monologue to Chinese mid-session. Kenneth Auchenberg, a partner at Alley Corp, posted a screenshot calling it a smoking gun. KR-Asia and 36Kr confirmed both Cursor and Windsurf were running fine-tuned Chinese open-weight models underneath. Cursor never disclosed what Composer 1 was built on. They shipped Composer 1.5 in February and moved on. The pattern: take a Chinese open-weight model, run RL on coding tasks, ship it as a proprietary breakthrough, publish a cost-performance chart comparing yourself against Opus 4.6 and GPT-5.4 without disclosing that your base model was free, then raise another round. That chart from the Composer 2 announcement deserves its own paragraph. Cursor plotted Composer 2 against frontier models on a price-vs-quality axis to argue they’d hit a superior tradeoff. What the chart doesn’t show is that Anthropic and OpenAI trained their models from scratch. Cursor took an open-weight model that Moonshot spent hundreds of millions developing, ran RL on top, and presented the output as evidence of in-house research. That’s margin arbitrage on someone else’s R&D dressed up as a benchmark slide. The license makes this more than an attribution oversight. Kimi K2.5 ships under a Modified MIT License with one clause designed for exactly this scenario: if your product exceeds $20 million in monthly revenue, you must prominently display “Kimi K2.5” on the user interface. Cursor’s ARR crossed $2 billion in February. That’s roughly $167 million per month, 8x the threshold. The clause covers derivative works explicitly. Cursor is valued at $29.3 billion and raising at $50 billion. Moonshot’s last reported valuation was $4.3 billion. The company worth 12x more took the smaller company’s model and shipped it as proprietary technology to justify a valuation built on the frontier lab narrative. Three Composer releases in five months. Composer 1 caught speaking Chinese. Composer 2 caught with a Kimi model ID in the API. A P0 incident this year. And a benchmark chart that compares an RL fine-tune against models requiring billions in training compute without disclosing the base was free. The question for investors in the $50 billion round: what exactly are you buying? A VS Code fork with strong distribution, or a frontier research lab? The model ID in the API answers that. If Moonshot doesn’t enforce this license against a company generating $2 billion annually from a derivative of their model, the attribution clause becomes decoration for every future open-weight release. Every AI lab watching this is running the same math: why open-source your model if companies with better distribution can strip attribution, call it proprietary, and raise at 12x your valuation? kimi-k2p5-rl-0317-s515-fast is the most expensive model ID leak in the history of AI licensing.