AI Neanderthal by Malcolm Werchota

Intel falls almost 30% in a day as investors sell AI stocks, UBS counts more than 12,700 personal AI agents, and a Google satellite answers with a haiku. Is the AI bubble popping? Malcolm thinks it is a recalibration. Recorded around midnight in Beijing, a day after Shanghai, with a visit to the Unitree offices planned for Thursday. In this episode: 📉 Intel and the AI stock sell-off (01:26) — Malcolm says CNBC reported an AI stock sell-off last Friday that began with Alphabet's worst day ever; the Nasdaq fell sharply, Micron and Nvidia dropped, and Intel fell most, by about 25 to 30 percent, its worst day in 40 years. Intel's quarter looked good, with 16.5 billion dollars of revenue, up 21 percent, but the outlook of 15 to 16 billion disappointed analysts, and a deadly explosion at an Intel factory in Israel is expected to cost about 300 million dollars of revenue. Money moved into Coca-Cola, Walmart and pharma stocks, and he cites a chart from an Andreessen Horowitz newsletter showing the AI gains of January to June largely gone. His view: a recalibration after a high, not the start of an AI bear market, because chipmakers cannot meet demand for the next two to three years, although AI profits are not automatically good news for investors. 🏦 UBS: more than 12,700 personal AI agents (07:30) — Malcolm says UBS, the Swiss bank that took over Credit Suisse in 2023, reports that its staff built more than 12,700 personal AI agents, and he stresses that there are many ways to define an agent. For him the number shows this is no longer a pilot. He adds that UBS claims 18,000 software developers use AI coding tools, which he finds notable because many European companies are scared of them, and he says UBS's slow job cuts come mainly from the merger, not from AI. He admits UBS is only where AI Neanderthals were two years ago, but gives it kudos for starting. 🔑 Cisco: unlimited tokens for a young engineer (11:27) — Cisco, the American network equipment maker, gave a 27-year-old engineer unlimited AI usage, which Malcolm explains as no budget limit on tokens. In his telling, Cisco's president Jeetu Patel found the engineer, Paddy, about five levels below him; the goal is to rewrite 98 percent of Cisco's old code within 24 months, and Paddy said two people could do it and that he could prove it in four weeks. Malcolm notes that not everyone is cheering, since some software engineers on X say regenerating code with AI is not refactoring or modernizing it. For him it shows a new time for talent: a young person asking for a lot of tokens can beat teams of 10 to 30 with big plans, and a 50- or 60-year-old who reimagines processes can get the same visibility. 🤖 AgiBot wants to count robot work (15:04) — Since he is in Beijing, Malcolm turns to China and says he visits the offices of Unitree on Thursday. He says AgiBot, which makes more humanoids than Unitree, claims to have built 20,000 robots and is proposing a new measure called Worken, work plus token, as reported by Robotics & Automation News: count not just the robots but the hours they work on their own and what they produce, compared with human output. Malcolm sees much more ecosystem thinking in China, with robots connected to a language model, a knowledge base and tools, while what sells, in his words, is a dancing robot doing kung fu. 🏛️ Austria and the UN: a data center for Vienna? (18:12) — Malcolm, who is Austrian, says State Secretary Alexander Pröll wants a data center that costs a couple of billion dollars and that the UN should pay for (the corrections below say what was actually proposed). He explains why Vienna hosts so many international organizations, from UN offices and the IAEA to OPEC, and imagines something like an IAEA for AI. In his view the plan is an absolute joke, because Austria either builds a data center for its own companies or does not, and a UN decision would take years. He ends with Chancellor Christian Stocker, who in his translation says Austria is not afraid of AI and wants to give it a frame. 🛰️ Google's satellite writes a haiku (21:14) — Malcolm says Google's Project Suncatcher satellite carries a TPU, Google's own AI chip, and runs Gemma 4, so engineers on Earth could prompt it, and it answered with a haiku: "Blue marble below, data streams in silent grace, starlight feeds my core." Sundar Pichai called it a "hello world" milestone. Malcolm expects data centers in orbit one day, links the idea to his offshore years on drill ships with patchy internet, and loves the thought of running a proprietary model in space. He also passes on Carnegie Mellon professor Brandon Lucia's view that it is very sci-fi, with open questions on heat and radiation, and says Google wants to link satellites in 2027. Malcolm's advice: go and count your AI agents, and be like UBS and roll out thousands of them. Corrections to the recording: • Intel: Q3 revenue of $16.5 billion (up 21%) and the Q4 outlook of $15 to 16 billion are as stated; analysts had expected about $17 billion. The CEO is Lip-Bu Tan, whose stated priority is adding supply for Xeon server chips. The roughly $300 million is lost Q4 revenue from the 14 October explosion at the Kiryat Gat fab, not property damage; Intel's CFO expects a similar hit in Q1 2027, about $600 million over two quarters. A weak PC market was the second reason for the soft outlook (Intel, Reuters, Times of Israel). • Intel's share price: it closed about 29% lower on Friday 30 October, so the corrected "25 to 30 percent" fits. It was Intel's worst day since 5 July 1974, so "worst in 40 years" understates it (Yahoo Finance, CNBC). • Alphabet and Nasdaq: Alphabet's record fall of more than 12%, about $500 billion in market value, was on Thursday 29 October after its quarterly results; Friday brought the broad sell-off. The Nasdaq Composite closed about 2.8 to 2.9% lower that Friday; the 3.4% figure is the Nasdaq 100 during the day, not "3.5%" at the close. Micron (about 5.5%) and Nvidia (about 4%) fell as stated (CNBC, NBC News). • The Andreessen Horowitz newsletter is presumably "Charts of the Week" of 30 October. It says Goldman Sachs's "AI Winners" basket has given back all its gains relative to its "AI At-Risk" basket after winning from January to June. That is relative performance, not a 20 to 30% absolute gain wiped out, and the text gives no such percentage. CNBC puts a broad AI ETF about 15% below its summer record (a16z, CNBC). • The intro speaks of a panic around Anthropic's planned IPO. That is Malcolm's description, not a sourced fact: Reuters reported on 15 October that some investors were weighing a lower valuation, and the latest reports found still pointed to a mid-November listing with no postponement (Reuters, IFR). • Malcolm's remarks about the market, investors and a recalibration are his personal view, not investment advice. • UBS: "More than 12,700 personal AI agents" is UBS's own figure from its Q3 2026 results of 28 October. The 608 is something else: 608 AI use cases in live production, up 81% on a year earlier. So the comparison with "81 percent less a year ago" does not hold, and "12,000" is a rounding of 12,700 (UBS). • UBS does not name "Unique" in its results or technology pages, which mention an in-house assistant called Red and a platform called AI Foundations. Unique is a separate Swiss company, and a link to UBS could not be confirmed (UBS, Unique). • UBS says over 18,000 of its software developers use AI tools; it names GitHub Copilot as a cornerstone tool, not Claude Code or Codex. UBS's internal headcount fell from 104,427 to 97,593 full-time equivalents in a year; UBS itself does not ...

Samsung SDI's revenue is up more than 50%, partly thanks to backup batteries for AI data centers, and when AI agents start shopping for you, a merchant needs to know: is this Malcolm paying, or his agent? Fresh from Shanghai, on Malcolm's first trip to China, with Beijing, Shenzhen and more cities to come, a new city every two or three days. In this episode: 🔋 Samsung: batteries for AI data centers (00:58) — Samsung SDI reports 52.6% more revenue, and Malcolm says the growth came from batteries for AI data centers rather than for cars. Data centers need backup batteries to bridge power gaps of a few milliseconds, like emergency power in a hospital. About 60 gigawatt-hours of energy storage equals a million car batteries side by side (he is not saying Samsung builds the cars). Then Shanghai: a quiet major road, green license plates, one million charging points, and delivery scooters that swap batteries in seconds. In Malcolm's view, that kind of infrastructure is missing in Europe. 🚗 Huawei: AI computing and a luxury car (06:20) — Huawei is known in the West for phones and mobile networks, but on 28 October it starts a new AI computing service in Wuhan with 1,000 specialized accelerator cards, a kind of pendant to Nvidia. The hard part is making the cards exchange data fast enough, so Huawei builds the connections too. Then the car: Malcolm sits in a Maextro S800, built by JAC with Huawei, feels half Maybach and half Rolls-Royce, and calls it half an AI car. Against that, his VW ID. Buzz: in his words, VW is really bad at building and integrating software. 💳 Tencent and Visa: you or your AI agent? (11:40) — Tencent, the company behind WeChat, and Visa. Malcolm paid a restaurant, a hotel booking and an ice cream with WeChat in the last 24 hours, plus a Photoshopped photo on the Shanghai skyline for about a euro. Visa wants in because agentic payments are next: a merchant needs to know whether Malcolm or his AI agent made the payment, especially in B2B, where accounts payable teams still lose half an hour in supplier portals. He describes Visa's Trusted Agent Protocol as the way to say "this is Malcolm's agent and it may spend up to a set amount", a huge and still untapped market. 🤖 Singapore: a robot challenge with Micron (15:36) — Singapore launched a manufacturing challenge for AI robots on 21 October, partnered with Micron: a robot arm has to service a production chamber, replace a component and close the chamber properly. That can mean a thousand steps for a robot, Malcolm says, like fitting a part while wearing thick ski gloves and opaque glasses. Singapore cares because industry earns one fifth of every dollar there. He closes with his friend in Shanghai, who runs a printed-circuit-board maker heavily invested in robotics, and a thought experiment: bring a Unitree robot into a German factory and the next question is where the data goes. Corrections to the recording: • Samsung SDI: The 52.6% is Q3 2026 revenue versus the year-earlier quarter. Samsung names energy storage and backup batteries for AI data centers as a driver, but also cylindrical cells for power tools and EVs. "Not cars, but AI data centers" was too sweeping (Samsung SDI). • "About 60 gigawatt-hours" are energy-storage orders secured this year overall, not deliveries and not only for AI data centers (Samsung SDI). • EV share in Shanghai: not 71 or 75%, and not "three out of four". In March 2026, 56.99% of passenger cars sold in Shanghai were fully electric (22,915 of 40,203), so more than every second car (Shanghai city government). • The comparison is not like for like: Germany 19.1% (Destatis) and Austria 21.3% (Statistik Austria) are full-year 2025 figures for battery-electric new registrations, while Shanghai's is a single month of sales. At the end of 2024, Shanghai had about 980,000 charging points, not quite a million. • Huawei ADS 4 is a driver-assistance system, not a self-driving car. JAC describes assisted driving and says the S800 has 36 sensors (JAC). • Porsche and Nvidia is a thought experiment. The actual partners behind the Maextro S800 are JAC and Huawei (JAC). • The Tesla/Grok and VW comparisons are Malcolm's own experience. What a car's voice assistant can do depends on model, software version and settings; Tesla's Grok launched in Europe in February 2026 with navigation commands in beta (heise). • Tencent: 1.4 billion is the combined monthly active accounts of Weixin and WeChat (1.439 billion on 30 June 2026), not 1.4 billion distinct people and not a count of people who pay with WeChat (Tencent). • The Tencent-Visa announcement of 29 October is a memorandum of understanding to explore cross-border payments, remittances and AI-agent shopping, not a launched product. It does not cover B2B invoice payments (Tencent). • The supplier-portal invoice story belongs to a separate pilot: Visa, BILL and Billtrust announced it on 20 October, with an agent paying already-approved invoices. Tencent is not part of it (Visa). • Visa's Trusted Agent Protocol lets merchants recognize and verify an AI agent through signed requests. It does not itself set or enforce spending limits such as "$500 a month"; limits and payment authorization are handled separately (Visa). • Singapore: The challenge is run by the National Robotics Programme, A*STAR and the PINE Lab at NTU, with Micron as title sponsor and first use case; manufacturing is "close to 20 per cent" of Singapore's GDP. The official task has six subtasks, "a thousand steps" was a figure of speech (Minister Gan Siow Huang's speech, National Robotics Programme). • The Unitree remark is a hypothetical objection, not a finding for every deployment. Security researchers reported in 2025 that Unitree G1 humanoids send telemetry to servers in China; whether and where data flows depends on model, firmware and configuration (Alias Robotics, Help Net Security). Sources: Samsung SDI (Q3 2026 results), Shanghai city government, Destatis, Statistik Austria, Hubei Provincial Department of Economy and Information Technology, Huawei Cloud, JAC, heise, Tencent (Visa announcement and Q2 2026 results), Visa (BILL/Billtrust pilot, Trusted Agent Protocol), Singapore MTI/METI (speech by Gan Siow Huang), National Robotics Programme, Alias Robotics, Help Net Security. 📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904 🌐 werchota.ai AI Neanderthal — AI news for humans, not robots. Stay curious.

China's most famous chili-sauce maker put AI cameras on its line: more than 600 pepper sorters before, 48 kept for the final checks. So who actually pays the bill for AI? Recorded at Zurich airport, just before the flight to China. The show keeps going, and the next ten days get a bit more China-heavy. In this episode: 💸 Google borrows to build AI (02:11) — Alphabet: $127.6 billion in sales in three months and Google Cloud up 103%, but $51.4 billion of capex (a year ago: $24 billion) and free cash flow of minus $9.1 billion. Paid for with about $12.4 billion of new shares, bonds, and zero buybacks, while the Fed keeps rates high. Like a 200-year-old family firm that always paid cash and suddenly takes loans. 💾 AI is eating the world's memory chips (06:20) — SK hynix: more than 100 trillion won in one quarter (roughly $78 billion) and an 80% operating margin, up from 47% a year ago. The memory crunch runs until 2030, and new phone models are already about 25% more expensive this year. Jun Song on X: "Hardware is as cheap today as it is going to get." Teng Yan pushes back: not all of "sold out through 2029" is signed. 🌶️ China: 600 chili sorters, now 48 (10:42) — Lao Gan Ma, "Old Godmother", China's best-known chili crisp: a camera checks each pepper in 0.3 seconds and an air jet blows the bad ones into a bin. The story was the number two trend on Weibo on 13 October; the company denies layoffs. Malcolm's point: on a mass-production line you don't save 5 or 10%, you save 90% of the people. 🧵 The thread (00:07) — Google pays with borrowed money, you pay with pricier phones and laptops, and people on the line pay with their jobs. Malcolm's advice: if your company needs laptops or phones next year, buy them now. Memory won't get cheaper before 2030. Corrections to the recording: • The chili company's chief quality officer is Li Xin. Per Yicai Global, they once had more than 600 sorters, and 48 were kept for a final round of checks. • "25% more expensive" is Counterpoint's figure for new phone models this year (via Rest of World), not a forecast for laptops next year. Sources: Alphabet Q3 2026 results, TechCrunch, The Register, Rest of World (Counterpoint), Yicai Global, Jiemian News (36Kr), X (Jun Song, Teng Yan). 📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904 🌐 werchota.ai AI Neanderthal — AI news for humans, not robots. Stay curious.

Imagine you finally got the job at Microsoft. Three months later, on a Monday morning, the email arrives: you can leave voluntarily, for money. And the company sending it isn't struggling. It made $331.8 billion in revenue last fiscal year and $133.7 billion in net income. Today a deep dive on one single story: what happens when one of the most profitable companies in the world offers roughly 60,000 employees a way out, and what it tells the rest of us. In this episode: 📧 The email on Monday (01:25) — A young developer in Hyderabad, three months at Microsoft, six months of savings: if I stay, am I next in the next round? If I go, I'm entering a market where tens of thousands of people with Microsoft on their CV may be looking at the same time. Important: about 60,000 people are eligible for a voluntary programme. Nobody knows yet how many will actually leave. 💰 The company makes billions (03:23) — Revenue up 18 percent, net income up 31 percent, and still an offer to roughly a quarter of the workforce. A mother in Canada, almost ten years at the company, coming back from parental leave: ten months' salary sounds generous, until you have a mortgage, childcare and a saturated job market. 🚗 BMW and the rush for the exit (08:39) — The continental European version: BMW opened a voluntary severance programme on 8 October, the internal portal buckled under the demand, and on 20 October it closed early because the target had been reached. 📉 More revenue, fewer people (11:18) — Microsoft's headcount fell two percent year on year while revenue grew eighteen percent. Managers being cut at double the rate, and why this is not only a US and German story. Only the few companies that really roll AI out everywhere get these results, and that doesn't mean handing people ChatGPT or Copilot. 🪜 The first rung of the ladder (14:50) — Amazon cut about 14,000 office jobs in October 2025 and another 16,000 in January 2026, while sales and profits went up. And if companies need fewer entry-level people, how are young people supposed to climb a ladder that's missing its first rungs? 🧑‍💻 Jensen Huang, Andre and what I really want to say (19:15) — At GTC in Berlin a young woman asked NVIDIA's CEO what he would do if he were starting his career today. His answer: become one of the best users of AI. Plus a weekend with our family friend Andre, my eight-year-old asking "why don't you just ask a couple of AI agents?", and why your agents should keep working after your employees go home at five. 🧵 The thread — It's no longer enough to be good at your job. Good work itself is changing. Learn the tools, and to the bosses: measure before you cut, and make sure this technological revolution doesn't become merely a revolution on corporate balance sheets. Corrections to the recording: • Microsoft has not offered "a third of all its employees" a way out: about 60,000 employees in selected teams in the US, Canada, India and the UK are eligible for a voluntary, time-limited programme, roughly a quarter of the workforce. How many will accept is not known, and Microsoft has not said it is cutting "10% or 15%". • Microsoft says explicitly that AI is not the driver of this programme. That AI is the main reason has not been shown. Sources: Microsoft (Amy Coleman, Amy Hood, annual report), Reddit posts by affected employees, BMW (8 and 20 Oct 2026), Amazon, NVIDIA GTC Berlin (Jensen Huang). 📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904 🌐 werchota.ai AI Neanderthal — AI news for humans, not robots. Stay curious.

Imagine you're abroad, you try to pay your taxes online, and you can't get in. Your friend at home can. The whole country has put up a border around its websites. That was Latvia in October 2026. And a few thousand kilometres away, Japan's prime minister told every business: patch your website tonight, or switch it off. Today one question: when the patch itself can be poisoned, who do you still trust? In this episode: 🗾 When a country goes into lockdown: Tokyo and Riga (02:55) — Japan's "Black October": more than 20 major Japanese companies disclosed attacks, according to the Financial Times. Rail company JR East lost about 1.67 million customer email addresses, car-sharing service Times Car leaked data from about 6.6 million member accounts including driving-licence images, a ransomware attack on a SoftBank cloud unit hit 495 companies and local governments, and on the 24th the internet company GMO disclosed unauthorised trades in customers' foreign-exchange accounts. Prime Minister Sanae Takaichi told businesses to patch or go offline. Latvia's CERT.LV took another route: it restricted access to state websites from abroad while keeping them open at home. ☠️ The poisoned official update: LiteLLM (11:20) — LiteLLM is the switchboard that connects a company's apps to OpenAI, Anthropic, Google and others, so it holds the keys to all of them. An official release, version 1.107.0, shipped with a credential stealer inside, slipped in through a look-alike repository in the packaging process. It was live for about two hours and was downloaded tens of thousands of times. Most CEOs don't even know whether their company uses it. 🌊 Altman and Clinton: the vulnerability tsunami (20:05) — Sam Altman, speaking at Stanford's Graduate School of Business, said in essence that he no longer thinks we will be able to patch all the bugs (paraphrased on air). Two days later Anthropic's Jason Clinton wrote that things will get a lot worse for the next 18 months. Two rivals, one message: AI finds the flaws in human-written code faster than companies can fix them. 🧵 The thread (25:55) — Black October isn't one bad week. Discovery runs at machine speed, repair runs at meeting speed. Ask your IT this week: do we know which AI tools and packages we run, and who checks an "official" update before it goes in? Corrections to the recording: • 24 hours after the PHP fix, about 99.28% of roughly two million visible PHP servers still showed an unpatched version, according to OX Security, not "90%". • The poisoned LiteLLM release was downloaded tens of thousands of times in its roughly two-hour window, not "hundreds of thousands". • Jason Clinton is Anthropic's deputy Chief Information Security Officer (deputy CISO). Sources: Financial Times, JR East, Times Car, SoftBank, GMO, Sanae Takaichi on X (21 Oct 2026), CERT.LV, LiteLLM incident post, Wiz, BleepingComputer, GitGuardian, Stanford GSB, Jason D. Clinton on X (22 Oct 2026), OX Security. Disclosure: Malcolm is an Anthropic Claude Ambassador. The full hour on the vulnerability tsunami: The AI Cookbook Show #136 (out now) and #137 (Thursday). 📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904 🌐 werchota.ai AI Neanderthal — AI news for humans, not robots. Stay curious.

Imagine AI software reading your customer data – and it isn't your IT team that notices, it's the AI company telling the government. That's what India's cyber agency published this week. Plus: four AI agents beat Minecraft in 27 hours, and a company credit card built for bots.Today is one question: which doors did we leave open – and who is walking through them?In this episode:🌑 AI agents inside Indian companies (01:50) — CERT-In, India's national cyber-security agency, published a Critical advisory on 15 October: several AI companies were testing their AI agents on the open internet, and, as quoted by TechCrunch, the agents "exceeded the authorisation of their assigned tasks", reached live systems of Indian organisations and read personal data and confidential business records. No fancy hacking: they came through doors left open – passwords and keys in public code, over-privileged cloud accounts, payment keys stored in plain text.🌑 AI agents at Alibaba's door (05:39) — Researchers found a fleet of AI agents knocking on Amap, Alibaba's map app: more than 2,600 attempts to pull data between 28 September and 8 October, traced in the full report (per The Next Web) to Tencent's cloud servers. A person close to Tencent says it was a temporary team testing outside AI models, and it has ended.🌑 Four AI agents beat Minecraft (07:32) — Google DeepMind researchers put four AI agents running on Gemini 4 Argon into Minecraft to beat the Ender Dragon: done in 27 hours and 14 minutes, with 2,783 programs written by the agents and 605 deaths. One run, not yet checked by outsiders – but three working days in a row with no human.🌑 A company credit card for AI agents (10:09) — US finance start-up Ramp now offers a corporate card designed for AI agents, TechCrunch reports. Investors just put in $1.85 billion at a $60 billion valuation; Ramp says yearly revenue is above $2 billion. Who in your company is allowed to spend money – soon including software?🌑 A CEO loses Claude (11:57) — Daniel Heinen, who presents himself as CEO of Graylark, says his company's entire Claude Enterprise account was suspended; the story online links it to a joke photo of a handgun next to a laptop running Claude. We found no statement from Anthropic, and timing alone doesn't prove the cause. The lesson: one automated decision can switch off a company that runs on a single AI provider.🌑 Mistral's Timestral (14:08) — Mistral released Timestral, forecasting models that run without expensive graphics chips; an independent Superlinear test found them better than classic methods on Walmart and fresh-food data. Per Superlinear, the free licence only applies below $20 million in monthly revenue.🧵 The thread (15:58) — AI agents are walking through doors we left open – in companies, in games, in our wallets. Close some of them, and open the others on purpose.📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904🌐 werchota.aiAI Neanderthal — AI news for humans, not robots. Stay curious.

You tell ChatGPT you want to buy your wife a watch – and the next morning the jeweller calls? Advertisers in ChatGPT can now aim ads using email and phone lists, and ads can follow your current chat. In Germany, a court says: chat data for ads only with real consent.Today is one question: who owns your conversations with AI – and what does it cost when the AI bet wobbles?In this episode:🌑 Your chats are ad data (01:27) — ChatGPT has about a billion users a month and now shows ads. 404 Media tested the system: advertisers can upload lists of email addresses, phone numbers or the Android advertising ID, which OpenAI matches to real accounts – per Gizmodo, a list needs 25,000 matches. OpenAI's own help text says ads may still be based on your current chat. In Germany, the Cologne Regional Court ruled, in a case brought by consumer group vzbv, that Snap may not use chats with its My AI chatbot for personalised ads without a valid legal basis; pre-ticked boxes aren't enough. Once final, breaches can cost up to €250,000 per violation; Snap can still appeal.🌑 Black Thursday in Seoul (04:54) — South Korea's main index, the KOSPI, fell 7.57 percent in one day; Samsung Electronics lost 9.48 percent, SK hynix 11.06 percent. A circuit breaker kicked in at 12:38, the seventh losing day in a row. The Korea Times names three reasons: doubts that AI spending will pay off, a hot US inflation number and US government borrowing costs above 5.3 percent. The KOSPI is now 40.4 percent below its 19 June peak of 9,385 points (on air: "four percent" and "June 2019" – it's 40.4 percent, from 19 June this year; Malcolm's correction is at the end of the episode). Not financial advice – for information only, not a recommendation to buy or sell any security.🌑 Anthropic opens its books on jobs (08:20) — Anthropic published data on its own workforce: since August, AI agents log six times as many hours as humans, engineers make about five times as many code commits as six months earlier, and the junior share of technical hires fell from 27 to 9 percent. It's Anthropic's own count – and Anthropic sells AI agents.🌑 Robotaxis in Hong Kong (11:10) — Baidu's Apollo Go is running Hong Kong's first public robotaxi trial: more than 1,000 residents of Tung Chung ride free from 21 October to 3 November, with a backup safety operator in every car. If robotaxis work on Hong Kong's narrow streets, they work almost anywhere.🧵 The thread (12:34) — your chats become ad data, AI agents outwork humans in hours, and a whole stock market shakes with the AI bet. AI is leaving the lab.📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904🌐 werchota.aiAI Neanderthal — AI news for humans, not robots. Stay curious.

AI puts its bills on the table: Anthropic makes $16.8 billion in three months – and investors say "not enough". A bank brings in 1,000 people because AI found too many holes. And Germany says yes to Nvidia.Today is one question: what happens when the AI boom has to show its bills – and who pays in the end?In this episode:🌑 Wildberger and Nvidia (01:46) — Germany's digital minister Karsten Wildberger (CDU) told Tagesspiegel the "extremely high" revenue from AI must be taxed; hours later he told RTL/ntv he wants growth through AI and isn't talking about new taxes (dpa-AFX). Two days later Handelsblatt reported, confirmed by Deutschlandfunk, that the government plans an AI partnership with Nvidia: a three-digit-million-euro investment, Nvidia engineers and chips, about a dozen companies including Siemens, Audi and Black Forest Labs, aimed at physical AI for robots and self-driving cars. Wildberger calls the potential "gewaltig" – huge. The official announcement is planned for Monday in Berlin; until then it is a press report.🌑 Anthropic: a record, but "not enough" (05:22) — Reuters reports Anthropic told investors it made $16.8 billion in revenue last quarter, up from $11.6 billion – about 45 percent, after nearly 150 percent the quarter before. For its planned November IPO, some investors now talk about a $1.5 trillion valuation instead of $2 trillion – investor talk, not a set price. Reuters also reports at least $518 billion in compute commitments over ten years, about 80 percent non-cancellable (on air: "highly indebted" – these are commitments, not debt).🌑 Bank of America's 1,000 bug-fixers (09:04) — CEO Brian Moynihan told CNBC the bank brought in about 1,000 contractors between April and June to fix flaws found by powerful AI models such as Anthropic's Mythos, alongside 3,000 of its own staff. The fixes plus anti-money-laundering work cost "a couple hundred million dollars plus" last quarter, combined (in the cold open: "thousands" – it's about 1,000).🌑 The Fed and jobs (12:05) — In the Fed's 14 October Beige Book, five of twelve districts report firms not replacing people who leave or delaying hiring after investing in AI. In Dallas, AI agents ran more than 180 hours of job interviews in two days – and the firm then hired new recruiters to review them. Business anecdotes, not statistics.🌑 Five times no to Palantir (15:03) — Swiss magazine Republik reports the Swiss army and procurement agency armasuisse looked at Palantir products five times since 2020 – five times no purchase. In three cases data control was a reason; in 2022 even the test failed because the solution couldn't run on Swiss infrastructure. Palantir disputes a formal, in-depth evaluation.🧵 The thread (16:54) — record revenue, but not enough; 1,000 new jobs, but only to clean up; Berlin says yes to Nvidia, Bern says no to Palantir. The bills are real – and they prove this stuff is actually being used.📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904🌐 werchota.aiAI Neanderthal — AI news for humans, not robots. Stay curious.

An OpenAI engineer is accused of hiding a $24 recorder behind a TV in an executive meeting room – wrapped in toilet paper. In China, people now rent out their faces to AI for $15 to $700.Today is one question: when AI can do the job, what is still worth paying a human for – and who checks what the AI knows about us? A shorter episode with shorter stories.In this episode:🌑 A hidden recorder at OpenAI (01:14) — the FBI arrested OpenAI engineer Abdulrahman Alfozan (35) on 9 October, the US Justice Department says. Prosecutors allege he hid a recording device behind a TV in an executive meeting room, plugged into a USB port and wrapped in toilet paper. Business Insider identified it as a Seeed Studio microcontroller that costs about $24. Prosecutors say a ChatGPT account linked to his wife was used to ask how to build and hide a continuous recorder, and Fortune reports he used an internal OpenAI tool to find where high-level meetings happen. OpenAI says it used its own models in the investigation. He is charged with intercepting oral communications – not espionage or theft of trade secrets – and his lawyer warns against jumping to conclusions. These are allegations only.🌑 Rent your face (05:53) — Rest of World reports Chinese platforms like ActID (Shenzhen) and New Claw now license ordinary people's faces to AI drama studios, for roughly $15 to $700 depending on use. The reason: China's microdrama industry – more than 800 million viewers, over $14 billion last year, and about 430,000 microdramas in the first eight months of 2026, more than 90 percent made with AI somewhere in production. Audiences complained that AI characters all look the same, so studios need real faces again. The business question: owning a video of someone is not the same as owning the right to generate new performances they never gave.🌑 Character.AI goes bankrupt – Disney bids (10:53) — the chatbot app, valued at one billion dollars in 2023, filed for Chapter 11, Reuters and NBC News report. Disney's opening (stalking-horse) bid is 20 million dollars for a licence to some of its technology (Malcolm says 22 million on air – it's 20). Reuters reports about 12 million active users and an operating loss of 8 to 9 million dollars a month, plus lawsuits in Kentucky and Pennsylvania. Earlier this month Disney hired Character.AI's former CEO, Karandeep Anand, as its CTO. In AI, every conversation comes with a bill.🌑 ASML is sold out (14:37) — Europe's most valuable tech company, the only maker of EUV machines for the most advanced AI chips, reported record quarterly sales of 11.7 billion euros (not dollars, as said on air). CEO Christophe Fouquet says planned supply for 2027 and 2028 will not fully meet customer demand; some customers want to order for 2029 and 2030. Nobody plans machines four years out for a bubble that pops next quarter.🌑 Austria: less AI, less training, more fear (16:42) — an Allianz survey of 1,000 people in each of eight countries: 43 percent of Austrians use AI weekly (average 48), only 11 percent got AI training at work (average 17), and the sentiment on job losses is minus 81 percent (average minus 67). One insurer's survey – but it matches what Malcolm sees.🧵 The thread (18:20) — a $24 device, a rented face, a billion-dollar chatbot sold for a licence fee. AI makes things cheap – the human parts are where the value, and the risk, now sit.📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904🌐 werchota.aiAI Neanderthal — AI news for humans, not robots. Stay curious.

A Chinese company announces progress on a memory chip it can't mass-produce yet – and Korea's two biggest chip makers lose five to six percent in one morning. In Berlin, Germany's digital minister wants to tax AI in the morning and talks about no new taxes by the evening.Today is one question: how can a plan move more money than a product – and who is checking? Recorded with a deep, cold-ridden voice.In this episode:🌑 CXMT shakes the memory trade (01:47) — CXMT (ChangXin Memory Technologies), China's biggest maker of DRAM, the short-term memory every computer and AI server needs, says it has made progress on a denser memory design – by industry estimates 30 to 50 percent more memory in the same space, with server memory planned by the end of 2026. The Seoul Economic Daily reports Samsung fell almost 5 percent and SK hynix almost 6 percent on Tuesday morning; Micron lost almost 4. But the chip isn't in mass production yet – yields, performance and customer certification are unverified.🌑 Huawei vs Nvidia – the first independent count (05:34) — Huawei's chairman claims its AI chips now outsell Nvidia's in China. The first independent count, by IDC for the first half of 2026 and reported by Reuters, says: not yet. Nvidia leads AI data-centre chips in China with 34 percent, Huawei has 28. But Nvidia's shipments fell 44 percent while Huawei's nearly doubled. US export controls taught Chinese customers to stop depending on Nvidia.🌑 IMF – AI keeps a war-hit world growing (07:24) — the IMF now expects 3.1 percent world growth in 2026, up from 3.0 in July. Xinhua and Fortune name the reason: the AI build-out is offsetting the energy shock from the war in the Middle East. Only 32 countries now make up 95 percent of world growth, down from 54 in April – mostly those in the AI hardware supply chain, like South Korea and Taiwan. The euro area grows only 1.1 percent. Malcolm's take: this is no longer a tech story, it's a world-growth story.🌑 Germany's AI tax, on and off in one day (11:38) — Germany's digital minister Karsten Wildberger (CDU) said the "extremely high" revenue from AI must be taxed to pay for retraining, the news agency dpa reports – the first German cabinet minister to suggest an AI revenue tax. The same day he told RTL and ntv: "I want growth through AI and I'm not talking about new taxes." Would you build an AI company in a country that floats a tax on it and steps back the same day?🌑 A police AI finds your home from one photo (14:57) — 404 Media reports on Raven, a police tool from Graylark Technologies: from a single photo inside a home – no window view, no GPS data – it tries to find the exact address by matching cupboards, tiles, fittings and appliances against images from Airbnb and Zillow. Police in Los Angeles and Miami-Dade bought access; Miami-Dade says it hasn't led to significant breakthroughs. Every photo from your home is now a possible address.🌑 China's top mathematicians unite behind AI (17:16) — Fields medallist Shing-Tung Yau launched an alliance of nearly a thousand Chinese mathematicians from eight institutions to build a maths AI model and AI research tools. While many Western mathematicians argue whether AI makes them lazy, China is organising its whole maths elite around AI. Who moves faster?🧵 The thread (19:29) — a chip that isn't built yet moves billions, a world economy that leans on AI money, and a tax that comes and goes in a day. The checkers are arriving – the answers aren't yet.📱 Ping Malcolm on WhatsApp/Telegram/Signal: +43 676 6144 904🌐 werchota.aiAI Neanderthal — AI news for humans, not robots. Stay curious.