Daily TEA – The Legibility Trap: Why AI Might Eat Your Own Moat
AI jobs myths, quantum medicine, Telegram going full super app, and a $3.5M film contest to counter the doomers
Hello, dear TEA-mates! Here is what you need to know today.
1. 🧠 The Case for Strategic Illegibility
HubSpot co-founder Brian Halligan is warning founders: the race to make your company fully “legible” to AI (every action logged, every decision traceable) might be the fastest way to destroy your own competitive advantage. Citing a growing consensus among VCs including Sequoia and Redpoint, Halligan argues that when AI can read your entire org as data, it also makes that org instantly replicable. The framework, which he calls “Dorsey mode” after Jack Dorsey’s recent blueprint for AI-native companies, re-imagines competitive moats as deliberately shallow and accepts that structural advantages will be temporary. Speed of adaptation, not durable barriers, becomes the new defensibility. Halligan draws on circular org shapes, AI as primary decision-maker, and hiring for taste over credentials. The uncomfortable implication: the more legible you make your company for AI efficiency, the more you commoditize the very thing you thought was your edge. The real question is not whether to use AI, but how much opacity is worth keeping. (Read More)
🫖 TEA For Thought: “The next most important question is not ‘should we make everything legible.’ It is ‘how much legibility do you need for survival?’”
2. 💼 The AI Job Apocalypse Is a Complete Fantasy
A16z contributor David George makes a data-heavy case that AI-driven mass unemployment is a myth driven by a failure of imagination, not economics. He cites NBER Working Paper 34984, which found “AI adoption has not yet led to meaningful changes in total employment,” along with a Federal Reserve Atlanta survey showing more than 90% of firms estimate no headcount impact over three years. The Census Center found only 5% of AI-using firms reported any changes, split nearly evenly between increases and decreases. Goldman Sachs estimates AI augmentation effects outweigh substitution significantly, and management discussions favor augmentation over substitution 8 to 1. George points to historical precedents: agriculture fell from 33% of employment to 2% yet the workforce expanded, and Excel did not kill bookkeeping but grew the profession into 1.5 million financial analysts. App store submissions are up 60% year over year, new business formation is climbing, and software engineering job counts have increased since early 2025. George acknowledges AI will compress certain entry-level roles and some transition periods will be painful, but frames the apocalypse narrative as a lump-of-labor fallacy that ignores how unlimited human wants generate new demand. (Read More)
🫖 TEA For Thought: “It depends on what type of jobs people are doing. If all people do is a task, then the task itself can be automated. It still depends on the very nature of the job, whether the job itself is a task that can be automated or whether it has a purpose, which is mainly to solve different problems for other people. Fear is the best weapon to control people.”
3. 🎬 Google Bets $3.5M on Optimistic AI Films to Counter the Doomers
Google is teaming up with XPRIZE and Range Media Partners to launch a $3.5 million global short-film competition, asking filmmakers to imagine a technology-forward, optimistic future. Submissions of up to three minutes are accepted in traditional live-action, animation, or AI-generated formats, with an August 15, 2026 deadline. Google’s AI tool Flow is available to all participants. The grand prize winner gets full creative and production support to turn their three-minute entry into a feature film. The initiative sits under Google’s 100 ZEROS partnership framework and runs at futurevisionxprize.com. The timing is deliberate: as AI anxiety dominates public discourse, Google is funding a counter-narrative, amplifying voices who see abundance rather than threat. Emerging creators with limited production budgets can now compete by using the same AI tools the competition is meant to celebrate. (Read More)
🫖 TEA For Thought: “It is all about controlling the narrative. Whoever is louder wins. With all the doomers dominating, Google will now bring more positive voices about AI, using AI itself to do it.”
4. ⚛️ Quantum Computers Just Simulated a 12,000-Atom Protein
A 24-author research team led by Kenneth M. Merz Jr. has crossed a major barrier in quantum chemistry: simulating protein-ligand complexes at 11,608 and 12,635 atoms, a more than 40x increase in system size compared to prior work, with up to 210x improvement in accuracy. The team ran 9,200 quantum circuits over 100 hours on IBM 156-qubit processors alongside classical supercomputers Fugaku and Miyabi-G, collecting 1.3 billion measurement outcomes and achieving 72.5% parallel efficiency. Fragment energy calculations matched coupled-cluster accuracy levels (the gold standard for quantum chemistry). The paper, posted to arXiv as 2605.01138, establishes a scalable pathway for applying quantum computing to real biomolecular problems, including drug-protein binding simulations that previously required enormous approximations or were simply out of reach. (Read More)
🫖 TEA For Thought: “Imagine one day when quantum computers are as easy to access as AI is today. There will not be any disease.”
5. 💬 Toncoin Surges 36% as Telegram Takes the Wheel on TON Chain
Toncoin jumped 36% in 24 hours on May 4, hitting a four-month high of $1.80, after Telegram founder Pavel Durov announced that Telegram will replace the TON Foundation as the primary driver of The Open Network blockchain and become the network’s largest validator. Transaction fees dropped sixfold to approximately $0.0005 per transaction, with a near-zero fee model as the target. The announcement triggered a broader ecosystem rally: Notcoin gained nearly 26% and Dogs token rose over 100%. Despite the price spike, the network’s fundamentals remain far below 2024 peaks: TVL sits at $69 million versus $800 million at peak, and daily active wallets are just under 50,000 compared to 700,000 in August-September 2024. New developer tools, performance upgrades, and a redesigned website are expected within two to three weeks. (Read More)
🫖 TEA For Thought: “Imagine when the bots on Telegram are real AI agents with our own configurations plus an agentic payment system. Super app in the making.”
🛠️ Tools of the Day
warpdotdev/warp — Warp is a modern, AI-native terminal rebuilt from the ground up with built-in agent workflows, collaborative sharing, and intelligent command suggestions baked directly into the shell experience. 55,781 stars.
TEAHEE Moment
Stay sharp, stay informed. See you tomorrow.
If you enjoyed this TEA, follow along on social for more:
Twitter/X