The new MacBook Pro starts at $1,999
Your M1 MacBook just got a lot harder to defend.
- Apple launched a 14-inch MacBook Pro with its new M6 chip. Battery life is rated at 24 hours, and Wi-Fi 7 and Bluetooth 6 are new. It starts at $1,999 with 16GB of memory.
- Apple says the M6 runs AI workloads up to 25% faster than the M5 and handles LLM prompt processing up to 13 times faster than the M1 model. Every number comes from Apple.
- Preorders open October 28 and the laptop ships November 6. A new M6 iMac follows the same schedule, starting at $1,499.
Is your M1 really still fast enough?
Time your heaviest local model or AI workflow on your current Mac now, so you have a real baseline when independent M6 reviews land.
A $2T IPO whose filing flags doom risks
The Claude maker is asking Wall Street for a $2 trillion price tag.
- Anthropic filed to list on Nasdaq as ANT on October 26. The reported value runs $1.5 trillion to $2 trillion, and the top end beats SpaceX’s $1.77 trillion record.
- Once it trades, index funds and retirement accounts will hold it by default. Plenty of savers will own AI risk without ever picking it.
- Price and share count land in the coming weeks. A debut could come as soon as November 16, and OpenAI is expected to follow next year.
1,300%. That’s Anthropic’s year-over-year revenue growth in the first half of 2026.
If you build on Claude, skim the S-1’s risk factors and line up a backup model for your core workflow.
Genie 4 holds a scene for two minutes
Google’s new AI builds a world, then lets you rewrite it while you’re standing inside.
- Genie 4 is out in research preview from Google DeepMind. Type a prompt mid-simulation and the weather flips or a new character walks in. Matching audio plays in real time, and a scene now holds together for up to two minutes.
- Gemini acts as the director here, planning steps while Genie simulates them. In a robotics test, that setup ran task steps live, a possible way to test robots before they touch the real world.
- Ultra subscribers get it in Project Genie in the coming weeks, along with early access partners in the Gemini API and Google AI Studio. Objects still drift the longer a simulation runs, and legible in-world text isn’t solved.
Two minutes isn’t long, but it doubled in about a year.
Ask for early access through the Gemini API or Google AI Studio, then test whether one storyboard scene survives its full two minutes.
Anthropic Files for Potentially the Biggest IPO Ever
Anthropic has officially filed to go public, setting the stage for one of the most closely watched IPOs in the AI industry. The company plans to list under the ticker ANT, with its latest filing revealing rapid revenue growth, ongoing losses and a governance structure designed to protect its AI safety mission.
- Anthropic reported $16.3 billion in revenue for the first half of 2026, up from $1.2 billion in the same period last year.
- Its operating loss narrowed to $693 million, compared with $2.1 billion a year earlier. However, its reported net loss reached $80.7 billion, largely due to noncash accounting adjustments tied to convertible debt. Anthropic warned it may not achieve or maintain profitability.
- The company’s enterprise business continues to grow. By the end of September, more than 11,900 customers were spending at least $100,000 annually, including 1,900 spending $1 million or more and around 250 spending at least $10 million.
- Anthropic’s seven co-founders will collectively control 50.1% of shareholder voting power through a new Founder LLC, giving them significant control over the company after it goes public.
- The Long-Term Benefit Trust, established to protect Anthropic’s public benefit mission, will have the authority to appoint most of the company’s board.
The governance structure has raised concerns because voting power will not directly reflect economic ownership, leaving public shareholders with limited practical influence over the company’s decisions. Anthropic has not yet disclosed how much it plans to raise or its proposed share price.
OpenAI Announces Three Scientific Discoveries
OpenAI has announced new AI-generated results in mathematics, computer science and physics, while introducing a process that brings external academics into the review of its findings.
- In mathematics, OpenAI says its model produced a proof of the local smoothing conjecture for the Euclidean wave equation across all spatial dimensions of two or more. The result also addresses related problems involving the Kakeya conjecture and Stein’s restriction estimates.
- In computer science, OpenAI claims to have found a deterministic algorithm that challenges the Exponential Time Hypothesis, a long-standing assumption that certain computational problems cannot be solved in subexponential time.
- Computer scientists Russell Impagliazzo, Ramamohan Paturi and Ryan Williams will help examine the proof.
- The physics information has not been released in full, but is a “mathematically rigorous existence proof of certain interacting quantum field theories containing gauge fields and fermions in four dimensions, including Quantum Chromodynamics and a unified completion of the Standard Model of particle physics.”
- A group led by physicist David Gross will guide the review of this work.
The company says it will not seek any academic credit for its models’ discoveries.
How to Have ChatGPT Take Your Meeting Notes and Turn Them into Action
ChatGPT’s Meetings plugin can capture meeting audio, create a personalized summary, and suggest follow-up tasks, such as drafting emails or updating project plans.
- Download and open the ChatGPT desktop app on your Mac.
- Open Plugins in the sidebar, search for Meetings, and install it.
- Open Meetings from the sidebar and complete the setup. Allow access to your microphone and system audio when prompted.
- Click Take notes to start. You can add your own notes to the meeting page to capture personal thoughts, questions, or extra context.
- Once finished, Meetings will process the audio and use the transcript, your notes, and relevant authorized connected-app context to prepare a summary and suggested action items.
- Review the notes for accuracy, then edit or start suggested actions. You can also ask ChatGPT questions about what was discussed.
- You can keep the meeting page private or share it with others. Shared pages give recipients access to the notes, but not the transcript.
Meetings is currently in beta for ChatGPT Pro and Business users in the macOS desktop app. Enterprise access is limited to an alpha group, while support for Windows, iOS, and Android is planned.
Anthropic faces lawsuit over alleged use of 274 Elvis Presley songs
Authentic Brands Group, which controls Elvis Presley’s music and likeness rights, has sued Anthropic, alleging the company used 274 of the singer’s songs to train its AI models without permission. The lawsuit names tracks including “Can’t Help Falling in Love” and “All Shook Up” and seeks up to $41 million in damages.
OpenAI launches GPT-6.1 Terra
The company has upgraded its mid-tier Terra model from GPT-5.6 to GPT-6.1, adding improvements in visual reasoning, computer use and writing while cutting pricing by 50%. The new model is available through ChatGPT Work, Codex, and the API.
Apollo economist says AI has provided “zero financial payoff” to companies outside tech
Apollo Global Management chief economist Torsten Slok says AI has yet to deliver measurable financial gains for major companies outside the technology sector, pointing to flat profit margins and revenue per worker among 493 S&P 500 companies.
Slok argues that while AI can make information processing cheaper, businesses still depend on human judgment, accountability and trust, echoing other studies that have found limited productivity gains from corporate AI adoption.
Elon Musk’s Grok drops first AI-generated “Odyssey” episode
Grok has released the first seven minutes of its AI-generated adaptation of Homer’s Odyssey, three months after Elon Musk promised a full-length, “historically accurate” film.
Apple adds AI assistants to office apps
Apple is adding an AI assistant built on Apple and OpenAI models to Keynote, Pages, and Numbers. It can design slides, draft and restructure text, and build charts and tables from your data.
The apps also support Model Context Protocol on macOS 27, letting outside AI agents read, edit, and automate documents. Keynote even adds an open JSON file format designed for AI agents and developer tools.
The catch: the assistant comes with the $12.99-a-month Apple Creator Studio, not the free versions of the apps. The smart Keynote has a cover charge, and it starts in November as a U.S. English beta.
OpenAI claims major AI science breakthroughs
Weeks after releasing 722 AI-written math papers, OpenAI says its unreleased models made three new discoveries across mathematics, computer science, and physics, tackling some of science’s longest-standing problems.
The company claims its AI disproved the decades-old Exponential Time Hypothesis, solved a 35-year-old wave conjecture, and proved part of a $1 million Millennium Prize problem in physics.
OpenAI used software to verify its math and computing proofs, while outside experts will scrutinize the findings. After 722 earlier papers, the bigger challenge might be keeping up with verification.
Meta plans $95 billion stock sale to fund AI
Meta plans to raise up to $95 billion through new stock sales, its first equity offering since 2013, to finance its massive AI expansion, surpassing similar fundraising efforts by Google and Oracle.
The company is launching $35 billion in offerings now and plans another $60 billion through nonvoting shares, allowing Meta to raise capital while preserving CEO Mark Zuckerberg’s voting control.
The AI race is getting expensive, with tech giants expected to spend roughly $800 billion this year and over $1 trillion next year. Even Meta is turning to investors for cash.
NASA plans three rocket launches for Artemis 3
NASA plans to launch three heavy-lift rockets within two weeks for Artemis 3, turning the mission into an orbital docking test instead of the crewed Moon landing now targeted for 2028.
Blue Origin’s New Glenn will launch a test lander, NASA’s SLS will carry four astronauts aboard Orion, and SpaceX’s Starship will join for a second docking test in orbit.
Astronauts will enter Blue Origin’s lander but not Starship. With Artemis 3 reportedly three months behind schedule, coordinating three massive launches might be the mission’s biggest challenge yet.
Apple responds to iPhone Duo display cracking reports
Apple acknowledged that a small number of iPhone Duo units have developed cracks in their inner displays shortly after launch, and told affected owners to contact AppleCare for replacements.
Users describe cracks along the lower hinge, sometimes after a sticking sensation or popping noise. Apple hasn’t identified the cause, while iFixit points to possible hinge-component failures.
Apple says the issue affects few devices, but limited replacement stock is already causing delays. With some circulating images reportedly AI-generated, even figuring out the scale is tricky.
Anthropic Files S-1, Posts First Profit
Anthropic publicly filed for a Nasdaq listing under the ticker ANT, with Goldman Sachs and Morgan Stanley leading. Q2 revenue hit $11.6 billion, up from $787 million a year earlier, with a first GAAP operating profit of $974 million. The catch: roughly $780 billion in locked-in compute commitments.
Anthropic Commits $120.5B to SpaceX Compute
Anthropic’s IPO filing shows it has agreed to spend up to $120.5 billion on computing power from SpaceX’s AI unit, giving it access to Nvidia chips. Once all capacity arrives, expected in December 2026, Anthropic will pay about $3.1 billion a month, with terms running into 2029.
Meta Sells $95B in Stock for AI
Meta plans to raise up to $95 billion for AI infrastructure and superintelligence, launching $35 billion in common and preferred stock offerings, with another $60 billion potentially coming through a future at-the-market program. Nvidia, AMD and Broadcom have expressed interest in buying up to $5 billion of shares. Meta stock fell about 1.9% on the news.
OpenAI Model Claims Math Breakthroughs
An internal OpenAI model has produced claimed proofs of the Kakeya conjecture in every dimension, a refutation of the Exponential Time Hypothesis, and a proof of Yang–Mills existence, part of a Millennium Prize problem. Several proofs were computer-checked, and leading complexity theorists are organizing an open, independent examination of the results before anyone pops champagne.
GPT-6.1 Terra Halves API Prices
OpenAI released GPT-6.1 Terra, its mid-tier model, at $1 per million input tokens and $5 per million output, half the price of Claude Sonnet 5.5. Cached inputs drop to $0.05. It targets everyday work and computer control, and is live in ChatGPT, Codex and the API.
Microsoft Offers Buyouts to 60,000 Staff
Microsoft opened a voluntary exit offer to employees at level 70 or below in the U.S., Canada, India and the U.K. About 60,000 people, roughly 27% of its 223,000-person workforce, are eligible. Quarterly capex is expected to top $50 billion as AI spending squeezes efficiency.
Bedrock Robotics Raises $600M at $4.1B
San Francisco’s Bedrock Robotics raised a $600 million Series C co-led by CapitalG and Emergence Capital at a $4.1 billion valuation, bringing total funding near $1 billion. The company retrofits excavators for full autonomy and is expanding from Texas into the Midwest and South Atlantic, including data center sites.
Qualcomm Moves to Acquire PickNik
Qualcomm agreed to acquire PickNik, the longtime steward of MoveIt, one of the most widely used open-source robotics manipulation frameworks. Announced in September, the deal would keep MoveIt open source, preserve community roadmaps and third-party hardware support, while bringing its software closer to Qualcomm’s robotics platforms. The acquisition remains subject to closing conditions.
Intel to Buy Classiq for Quantum
Intel agreed to acquire quantum software company Classiq, combining its automated algorithm synthesis platform with Intel’s silicon spin-qubit hardware and software. Classiq helps developers design quantum circuits without deep hardware expertise, potentially giving Intel a stronger position in full-stack quantum computing. The deal is expected to close in the first half of 2027.
Gelsinger Bets on Nagare Memory Startup
Former Intel CEO Pat Gelsinger is betting on Nagare Labs, whose Z2RAM promises SRAM-like bandwidth, DRAM-like density and NAND-like manufacturing. It is a direct shot at the memory wall throttling AI accelerators. The road ahead is arduous and littered with failed memory vendors, which makes the bet worth watching.
China’s AI Stock Bubble Pops
China’s STAR50 index of top AI and chip companies has crashed 40% in under four months, erasing about $480 billion. Unitree has halved since its August listing, and chipmakers MetaX and Moore Threads have also lost half their value since July. A cautionary chart for anyone pricing Anthropic’s IPO.
The RealReal Locks Out AI Shopping Agents
The RealReal upgraded its bot defenses after Meta’s Muse agent was shown re-adding listings to carts every 20 minutes, effectively blocking other shoppers for hours. Agents can still surface listings, but logging in, carting and checkout are now much harder. Agentic commerce just met its first bouncers.
Tesla Stretches Houston Robotaxi to Downtown
Tesla expanded its Houston Robotaxi zone from a 25-square-mile pocket in the northwest suburbs to a rocket-shaped geofence reaching downtown and Bellaire, announced with “Houston, we have liftoff.” Independent trackers count only about 34 driverless Model Ys covering the much larger area.
Cisco Puts AI Agents on Timed Access
Cisco launched Privileged Infrastructure Access, which gives administrators and AI agents network privileges that last only as long as the task at hand. It targets the standing credentials engineers leave on routers and switches, where single sign-on and multifactor checks often stay behind, just as agents start touching production networks.