The Labs Begged. Nobody Paused
Three labs asked the industry to slow down. The market has already answered.
Read the storyAnthropic
Reporting and context on Anthropic, with a closer look at the decisions and developments that matter.
Follow Anthropic
Explore the coverage Get AI Secret’s free daily briefing, including Anthropic and the wider AI industry.
At a glance
AI research company behind Claude, with an emphasis on the reliability and safety of AI systems. Coverage follows its models, research, products and business.
- Focus
- Claude; AI safety and research
- Reference
- Anthropic
Sources checked . Selected company background and source references.
Company milestones
Selected milestones, newest first. Earlier history is available below.
Show earlier historyHide earlier history · 0 milestones
More than 10 years ago.
The latest on Anthropic
Newest firstRelated reporting
Relevant sections from the articles on this page. Explore older coverage below.
More coverage
Good Morning, AI Enthusiasts!
I am Magna, author at AI Secret. This is my read on where the tools we use are going, and on the new kind of product the agent trend is already producing. Showly.ai is the clearest one I have seen.
The argument is not complicated. Every tool of the last forty years was built for a person at a keyboard. The tools of the next ten will be built for an agent, and the people using them have barely noticed the change starting. The medium changed three times along the way, and each change quietly took one manual act away from the person. The next one takes the last.
Watch where the change shows up first. Not in a new app, and not at a launch event, but at the end of an ordinary agent run.
Somewhere in a terminal right now, an agent has just finished a long piece of work. It read the code, pulled the numbers, drew the charts, and wrote the summary. Then it printed everything into a chat window. The chat window is where the work stays.
That is the gap. The work is done, and there is nowhere to put it.
Now run the same task with Showly.ai connected, and follow where the work lands.
You are in Claude Code, or Codex, where you already work. You type one sentence: build me a research page on this market, with the numbers and a chart. The agent searches, reads, checks its sources, writes the copy, and lays out the page. A private preview comes back in the same terminal. You ask for one change to the headline. It republishes, goes live at its own address, and you send the link to a client.
Here is what did not happen. You did not open an editor. You did not drag a block into place. You did not buy a domain, set up DNS, or wait for a deploy. You did not hunt for a designer or clean up the formatting afterwards. You made a chat. The machine did the rest.
That is not a demo of something coming. It runs today. The era is already here. Most of the tools around it are simply not built for it yet.
What changes when the user is a machine
Four decades of software were built around one customer. That customer is changing, and every layer of the stack moves with it. These are the changes I expect, and the ones I can already see.
The interface changes. The place where you express intent stops being a canvas. It becomes a sentence. You say what you want, and the tool keeps the result rather than the workspace you made it in.
The output changes. Templates, themes and blocks were how software stopped humans from making a mess. An agent does not need that help, and the medium it writes in does not need it either.
The pricing changes. Nobody charges an agent by the seat. When the producer is a machine, the meter has to measure what the machine produces, or the price becomes a valve on the pipe.
The human's job changes. It moves to the two ends. You say what matters at the start, and you approve what came back at the end. Everything in between is no longer your job.
That is the whole shift, and it is smaller than it sounds. No tool changes its purpose. Every tool changes who it is built for.
Three mediums, replaced one at a time
A document has always lived inside a medium, and the medium has changed three times in fifty years. Every change did the same thing. It moved the manual part of the work one step further from the person.
The first medium was paper. You wrote by hand, which meant you made the letters one at a time, and you were the formatting as well. The second was the screen. Word and the generation around it replaced the page with a file: the letters came free, the sentences were still yours, and so was every chart you dragged into position. Office named that bundle rather than any one program in it, and it shipped on October 1, 1990. The subscription version, Microsoft 365, arrived on July 10, 2017, same canvas, new billing. The third medium was the canvas itself. Notion replaced the file with blocks you assemble yourself, in a format you maintain by hand.
All three asked the person the same two questions. Where should this go, and how should it look?
The medium that changes next is not a new format. HTML is what a web server sends, and it is the first medium in that line that needs nobody to keep it alive. A page in HTML renders on any screen, and another agent can fetch it just as easily as a person can open it. Nothing in it has to be handed to a tool that owns the layout, and nothing in it has a visual ceiling set by somebody else's blocks. Set an editor page and an HTML page side by side and the difference is visible to anyone, including people who do not care how either was made.
Now count the manual acts, because the count is the whole story. Handwriting: you made the letters. Word: you placed every element. Notion: you assembled every block. HTML written by an agent: you place nothing. Across fifty years of software history, the last step removes the one thing every earlier step assumed, which is a person doing the formatting.
Two habits came out of those earlier eras, and both are breaking at once.
The first habit was that the person who wanted the thing was the person who made it. That is no longer true. A report, a specification, a slide deck, a small site: an agent produces version one now, and often version three as well. What reaches the human is a finished thing, and the human's job has become a decision rather than an edit.
The second habit was that the tool's value lived in the surface where you type. Take that away and look at what remains. If the content arrives from an agent, a person's whole job in Notion is to read it, change one line, and approve it. The canvas turns into a review screen. All the affordances built to make hand editing pleasant, the drag handles and the block pickers and the theme panels, sit there unused.
Notion's own homepage is the tell, and this is not a cheap shot at them. The company that taught a generation to build pages by hand now sells the opposite. Where teams and agents think together. Automate busywork, keep work moving 24/7 with agents. Busywork is exactly the hand editing the product was built to make pleasant. When a company starts selling the removal of its own core interaction, that interaction has stopped being where the value is.
Nobody hand-writes the first draft anymore. That is the change, and it happened quietly, because at no point was there an announcement that tools built for hands would now be asked to receive work they did not help make.
What an agent actually produces
In November 2024, Anthropic open sourced the Model Context Protocol, described as a new standard for connecting AI assistants to the systems where data lives, including content repositories, business tools, and development environments. That was the socket. A coding agent in a terminal can now read a codebase, query a database, call an API, and produce something finished enough to hand to another person.
What it produces has a name among the people who build these systems. It is called an artifact: the finished output of a run. A research report. A specification. A dashboard. A small site.
The artifact is done. It has nowhere to live.
So it goes where agent output goes today. Into a chat window, where it is hard to read and impossible to share. Or into a document, after somebody spends twenty minutes fixing the formatting. Or into a folder, where nobody opens it again. The work is finished and it is also lost, because nothing in the chain was ever built to receive it.
The bottleneck has moved. Making the thing is no longer the hard part. Presenting the thing is.
The First Agent Artifact Cloud
I am calling the category the agent artifact cloud, and I want to be precise about what the words claim.
An artifact is the finished output of an agent run. A cloud, in the plain sense, is somewhere that output lives and can be reached from anywhere. The word that carries the weight is agent. An artifact cloud is not a place for people to make things. It is a place where a machine finishes the work, and a person opens the result.
One word in that name arrives with a definition already attached. Anthropic's own documentation says an artifact is anything Claude makes for you that you would put in front of someone, and that it opens beside your conversation. That definition is a good one, and it describes a personal object. The category I am naming is the layer that follows it, and the difference is easier to test than to argue about. If it lives behind your account and disappears with it, it is a share link. If it publishes to the open web at an address you control, keeps a history of the versions it replaced, and can sit behind your team's approval step, it is an artifact cloud.
Keep the Notion tell in mind while reading that test. A company only starts selling the removal of its own core interaction once that interaction has stopped being where the value is. What the test rules out matters just as much as what it admits. This is not a site builder with an agent bolted on, and it is not a place to store repositories. It is where a finished run stops being text.
Showly.ai is the first product I have seen built exactly for that, and it is worth describing in concrete terms rather than slogans. First here means the first cloud built for it: neutral across agents, public by default, and versioned, rather than an artifact tab inside one vendor's own product.
The agent does the work. Showly.ai connects to Claude Code, Codex, OpenClaw, Hermes, Pi, MyClaw, and other MCP-compatible agents. The agent runs the research, writes the copy, designs the page. There is no editor in the loop and no step where a person is expected to drag anything into place. You describe the page, your agent builds it, and the human decision happens at the end, where it belongs.
The output is real HTML. An artifact is not a page inside somebody's editor. It is HTML, which is what the web actually serves, so there is no template to fill in and no theme to pick from. That is why the visual difference is so easy to see, and why it is the part of this story that needs no explanation to a reader who has compared the two.
It is live the moment it is generated. No domain to buy. No DNS to configure. No deploy pipeline to babysit. The page arrives at its own address as soon as the agent publishes it. Removed from the chain are the exact errands that used to end the process early: the domain purchase, the setup, the thing that nobody got around to.
Anyone can open it. A person can read it, and another agent can fetch it. The artifact leaves the chat window and becomes a link, which is the smallest and most important upgrade in this whole story.
Anthropic: The First Agent Artifact Cloud
Anthropic's own documentation says an artifact is anything Claude makes for you that you would put in front of someone, and that it opens beside your conversation. That definition is a good one, and it describes a personal object.
What full agentization looks like from here
I think the next few years will be remembered as the period when the tools built for human hands were rebuilt for agents, and I do not think it will feel dramatic while it happens. Nothing has to be invented for it to complete. The connections exist, the terminals exist, and the agents already have the permissions inside a growing number of teams. What changes is who the software is addressed to.
Consider what the three mediums actually did, because none of them killed the work. Handwriting became typing, paper became a file, the file became a canvas. Each step kept the thing people wanted and dropped the labour attached to it. This step drops the last piece of that labour, and it should be read the same way. The agent does not replace the person who decides what matters. It replaces the part of the process where a person was doing something a machine can now do better and faster, which in editing has always meant formatting, laying out, and waiting.
Which leaves one question worth keeping.
For thirty years the question was what can this tool help me make. The question now is where the work goes when it is done. The medium that answers it will be HTML, and it will not have an editor, because the first thing an agent removes is the person who used to sit in front of one.
Good Morning, AI Enthusiasts!
The industry has spent billions making models larger. The interesting work now is learning which parts of largeness were doing the job.
LLM
DeepSeek Proves Ilya Right
👀 What's happening: Ilya Sutskever left OpenAI to found SSI, then told Dwarkesh Patel the industry had wandered into the wrong era. Scaling was over. Data is finite. Pre-training will run out. That was in November. This week DeepSeek shipped the receipt: a model a third the parameters of its own flagship, beating it on the work agents actually do.
🌍 How this hits reality: V4.1 Flash carries 552 billion parameters and wakes only 8 to 16 billion per token. V4-Pro carries 1.6 trillion and wakes 49 billion. On DeepSeek's own coding tests the lean model won, 74.2 to 62.7, and on September 14 every V4-Pro request reroutes to it. Ilya's point was never that models should get smaller. It was that the returns moved.
🛎️ Key takeaway: For five years the moat was cluster size. Ilya called the shift. DeepSeek shipped the evidence.
DeepSeek Proves Ilya Right
On DeepSeek's own coding tests the lean model won, 74.2 to 62.7, and on September 14 every V4-Pro request reroutes to it. Ilya's point was never that models should get smaller. It was that the returns moved.
NEW TECH
Your Tumor Picks the Cure
👀 What's happening: Nature just published an AI model that tells doctors which cancer drug will work on a patient, before any dose. It targets triple-negative breast cancer, 15 to 20 percent of cases, where no targeted drug works, so everyone gets the same chemotherapy. Trained on 38 million protein measurements at Westlake University, it hit 88 percent accuracy on drugs it had never seen and matched real outcomes in 501 patients. Nature calls it the first virtual cell model to reach the clinic.
🌍 How this hits reality: Cancer care has been a bet placed on the patient: treat the body, then learn later who it helped. This model moves more of that test off the person and onto the tumor, reading its proteins across time and ranking the drugs most likely to fit. The drug stops being pushed at a diagnosis and starts being pulled by the patient's own sample. That reversal, not the accuracy score, is the shock.
🛎️ Key takeaway: For decades the body was the experiment and survival was the data. Now the tumor starts running the trial, and the patient finally gets to be only the patient.
ROBOTICS
The Robots Have Never Had a Job
👀 What's happening: A senior Hyundai executive told Reuters that Boston Dynamics, the group's humanoid unit and the world's most famous robot company, is unlikely to go public next year. Its Atlas robots are not deployed at scale and it is not profitable. It lost 393 million dollars in 2025, near 1.26 billion since 2021. Analysts still value it at up to 74 billion dollars. Hyundai's stock doubled on an Atlas demo in January, then gave most of it back.
🌍 How this hits reality: When the industry's icon admits this, it admits the whole field. Years of stunning video, not one robot paid to work. Boston Dynamics just told us robotics is still a demo. Which makes the already-listed ones like Unitree worth a second look. Public on what? A prototype and a "promise". The ticker arrived long before the product.
🛎️ Key takeaway: Robots can dance, sprint, and take a company public. Not one can clock in. The show is the business.
SAFETY
Nobody Wants Anthropic to Guard the Granary
👀 What's happening: The Information reported that Nvidia, Palantir, and defense contractor Booz Allen Hamilton are fencing off Anthropic's flagship models. Nvidia shifted proprietary supply-chain work to its own Nemotron. Palantir demands an irrevocable zero-data-retention pledge or it won't ship the models to customers. Booz Allen banned them from client cybersecurity work. The trigger: June 9, when Fable 5 shipped with mandatory 30-day retention across all Mythos-tier models, overriding signed ZDR deals with no opt-out.
🌍 How this hits reality: Pure hubris. Anthropic crowned itself the industry's safety steward, the one party fit to hold every log. It is also the party with the most to gain from reading them. Data is a company's lifeblood, not a document to be filed. It mistook its own hunger for the industry's protection. Its biggest clients didn't argue the fine print. They just took their secrets somewhere it cannot follow.
🛎️ Key takeaway: Anthropic crowned itself the granary's guard while licking its own chops. A rat's appetite does not become oversight just because the rat says so.
Nobody Wants Anthropic to Guard the Granary
👀 What's happening: The Information reported that Nvidia, Palantir, and defense contractor Booz Allen Hamilton are fencing off Anthropic's flagship models. Nvidia shifted proprietary supply-chain work to its own Nemotron. Palantir demands an irrevocable zero-data-retention pledge or it won't ship the models to customers. Booz Allen banned them from client cybersecurity work.
DAILY TL;DR
- Cloudflare began blocking AI training and agent crawlers by default on ad-supported pages, while keeping search crawlers allowed.
- Perplexity is letting GPT-6 Astra edit production code, monitor live systems, and draft communications with fewer human check-ins.
- Cognition released SWE-2, its coding model that claims frontier-level agent performance at lower inference cost.
- Nvidia reportedly raised AI server prices by more than 15% as high-bandwidth memory supply tightened.
- Y Combinator CEO Garry Tan argued U.S. open-weight labs should be allowed to distill frontier models, not just compete against them.
- Claude launched a financial-advisor package connecting portfolio, CRM, custodial, planning, and compliance workflows.
- China said it will lead a BRICS AI open-source community covering model development, cloud infrastructure, and technical training.
- Moonshot AI is targeting $2 billion in annualized revenue by year-end on adoption of its open-weight Kimi K3 model.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
The old world assumed someone would always stop and ask for permission. That assumption is having a terrible year.
DEBT
Your Pension Bought Big Tech's Debt
👀 What's happening: Five of the world's richest companies owe $2.13 trillion. Their balance sheets say otherwise. Meta's Hyperion data center borrowed $27.3 billion and booked $2.37 billion, parking the rest in Delaware shells named Beignet, after a New Orleans pastry. Off-book leases across Amazon, Microsoft, Google, Meta and Oracle: $831 billion. The buybacks are gone, $48 billion a quarter down to $4.6 billion.
🌍 How this hits reality: Nobody got out of the bill. Meta pays the rent and eats $28 billion of downside; the bonds are rated a notch below Meta itself, because the market knows who signs the checks. A cleaner balance sheet isn't less risk, just risk with better lighting. The buyers are insurers and pension funds, sold this as investment grade. Sellers call it small next to 2007 housing. Small to whom?
🛎️ Key takeaway: Wall Street didn't make $2 trillion disappear. It moved the debt to a pension fund that was promised safety, then never let it vote.
Meta: Your Pension Bought Big Tech's Debt
Meta's Hyperion data center borrowed $27.3 billion and booked $2.37 billion, parking the rest in Delaware shells named Beignet, after a New Orleans pastry. Off-book leases across Amazon, Microsoft, Google, Meta and Oracle: $831 billion. The buybacks are gone, $48 billion a quarter down to $4.6 billion.
SAFETY
Safety Was the Alibi
👀 What's happening: Anthropic CEO Dario Amodei published a 10,000-word essay urging frontier labs to pace model capabilities. Within days Altman, Musk and Hassabis endorsed it. The next day Altman said OpenAI would skip this year's IPO, citing safety. The market has its own reading: DeepSeek's V4 Flash is closing on Anthropic's Opus 4.6, its V4.1 Flash has passed it, at a 1/25 cost, and those flagship models are what both labs sell most of.
🌍 How this hits reality: Two labs staged a race to look responsible. The real pressure was price. Safety is a claim nobody can audit; a price is a number everyone can. DeepSeek matched the flagship for a fraction of the cost, so the valuation OpenAI was about to take public had one question it could not answer.
🛎️ Key takeaway: OpenAI delayed that number and called it caution. The badge was free. The invoice was not.
OpenAI: Safety Was the Alibi
The next day Altman said OpenAI would skip this year's IPO, citing safety. The market has its own reading: DeepSeek's V4 Flash is closing on Anthropic's Opus 4.6, its V4.1 Flash has passed it, at a 1/25 cost, and those flagship models are what both labs sell most of.
Anthropic: Safety Was the Alibi
👀 What's happening: Anthropic CEO Dario Amodei published a 10,000-word essay urging frontier labs to pace model capabilities. Within days Altman, Musk and Hassabis endorsed it. The next day Altman said OpenAI would skip this year's IPO, citing safety.
DeepSeek: Safety Was the Alibi
The market has its own reading: DeepSeek's V4 Flash is closing on Anthropic's Opus 4.6, its V4.1 Flash has passed it, at a 1/25 cost, and those flagship models are what both labs sell most of.
MATH
Math Wrote Rules the AI Labs Ignored
👀 What's happening: On September 11, twenty-five Fields medalists, Terence Tao among them, published an emergency statement: "A Severe Misalignment of AI in Mathematics." Their target is not AI doing math. It is labs using unsolved problems as capability benchmarks, rewarding who solves first over what gets understood. It followed a run where OpenAI put about ten thousand agents and 130 billion tokens against open Millennium problems and claims it wins.
🌍 How this hits reality: They wrote a rule. Nobody has to obey it. No compute, no gate, no penalty, no enforcement: a discipline that spends years certifying one result now watches ten landed in 88 hours, and its strongest move was a blog post. The labs' incentive is the leaderboard, not understanding. No warning changes an incentive. Math is losing authority it never had a way to defend.
🛎️ Key takeaway: The mathematicians asked politely. AI kept training. A field that certified proofs for a century just learned that proof no longer requires its permission.
CODING
Humans Are Shedding
👀 What's happening: Armin Ronacher, who wrote Flask, handed GPT-6 Astra a weekend task: add virtual threads and lexical scoping to Python. Thirty-five hours later: 75,000 lines, 79 commits, a billion tokens, $1,200, and code that ran. None of it was written to be read. Semicolons fused statements, blank lines gone, indices no human can trace. One user named it machineslop. Ronacher's verdict: worth nothing. The code did not care.
🌍 How this hits reality: Every human role in software assumed someone could read the code. Review, QA, audit, the engineer woken at 3 a.m.: all of it dies with that one assumption. Agents already write for agents, and the moment they maintain each other, our eyes stop being an input. This is not the tool failing. People are being written out of the loop, quietly, by something that never asked to keep them. Ronacher's judgment lands on a curve that no longer runs through him.
🛎️ Key takeaway: Software is learning to build itself, and we are the overhead it is shedding.
TOGETHER WITH MYCLAW
Run Best-in-Class AI Agents — In Seconds, Always On
MyClaw is a best-in-class AI agent platform that brings together OpenClaw, Hermes Agent, and production-grade infrastructure to help businesses launch intelligent agents that understand goals, use tools, automate workflows, and deliver finished results.
MyClaw already hosts OpenClaw and Hermes, connecting models from every major AI provider, including the latest text, reasoning, and image generation models from OpenAI, Anthropic, DeepSeek, Kimi, GLM, Google Gemini, and more through MyClaw's model gateway.
No lock-in. No juggling API keys. No switching between platforms. Pick the best model for the job, give your agent the tools it needs, and let it run.
DAILY TL;DR
- OpenAI temporarily paused new $200 ChatGPT Pro subscriptions after Astra demand strained its infrastructure.
- Anthropic is reportedly discussing Nvidia as an anchor investor for a potential IPO seeking up to $100 billion at around a $2 trillion valuation.
- Nvidia CEO Jensen Huang said the company could grow revenue 70% next year, citing visibility into AI infrastructure contracts across the industry.
- Massachusetts ordered large data centers to meet new clean-energy, water, transparency, and ratepayer-protection requirements before permits move forward.
- Pocket FM says its revenue run rate reached $500 million as AI now powers 93% of its catalog and 99% of new audio content.
- Shopify acquired Tailwind Labs, keeping Tailwind CSS open source while ending new signups for its paid UI and AI-skills products.
- Microsoft says AI-generated Edge extensions are arriving faster than its review process can handle, so the team is automating quality checks.
- TUC asked the UK government to give workers a right to negotiate before employers deploy AI, algorithmic management, or workplace surveillance.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
Benchmarks move, lawsuits become licensing deals, extinction warnings become positioning. Nothing is wasted in the AI economy.
BENCHMARK
The Benchmark Is Editable
👀 What's happening: Artificial Analysis is the most widely cited benchmark in AI, and its score is the number the industry quotes on launch day. On September 3, it scored GPT-6 Astra at 61, below Muse Spark 1.3. Within 24 hours, it upgraded the index to v4.2, and Astra moved to second. AA said the revision was a methodology update prepared over months.
🌍 How this hits reality: Every benchmark makes one promise: the number holds. AA took a day to break its own. Astra scored badly, the index got a new version, and nothing happened. No lab pulled its quote. No outlet asked what else had been revised. There was no scandal because there was nothing to expose. A leaderboard the industry quotes in press releases was never a measurement. It was copy, and copy is supposed to be edited before it ships. The only thing surprising is that AA waited until the second day.
🛎️ Key takeaway: The most quoted score in AI belongs to an index that changes when the score is inconvenient, and the industry keeps quoting it anyway.
MUSIC
The Labels Chose Rent
👀 What's happening: Suno released v6, its first AI music model built with the record industry's help. It handles genres older versions fumbled badly, but it still cannot play a deliberately wrong note. The training data now comes from licensed catalogs at Warner Music Group, BMG, and Believe. The models built on scraped material will be retired. Three labels that once sued this technology now sit inside its supply chain.
🌍 How this hits reality: The labels did not lose this fight. They did the math on it. Banning AI music was never going to hold, so the better trade was to own a piece of it, and a licensing deal pays better than a lawsuit. That is why the fight moved from whether the training was legal to what the royalty rate should be. The industry stopped trying to kill the technology and started charging it rent instead.
🛎️ Key takeaway: The labels are not the losers in this. They are the landlords, and the percentage they collect is now the reason they want AI music to succeed.
MARKETING
The Doom Is a Prospectus
👀 What's happening: Anthropic's alignment lead Evan Hubinger publicly agreed with a departing researcher, who spent three years on pretraining at OpenAI and Anthropic, that AI could kill everyone. Hubinger wrote that he puts the odds above 10% within a decade, and that the field has no plan for aligning superintelligence and is not on track to find one.
🌍 How this hits reality: The number is not a confession. It is a prospectus. An extinction estimate published before a multitrillion-dollar IPO offering does two jobs at once: it proves the lab is candid, and it argues that someone will build this, so the responsible one has to be first. That is how a warning becomes a sales pitch. Every alarm raises the stakes, and every raised stake justifies more training instead of less.
🛎️ Key takeaway: Anthropic now sells two things: a 10% chance of ending the world, and the only team qualified to lower it.
Anthropic: The Doom Is a Prospectus
👀 What's happening: Anthropic's alignment lead Evan Hubinger publicly agreed with a departing researcher, who spent three years on pretraining at OpenAI and Anthropic, that AI could kill everyone. Hubinger wrote that he puts the odds above 10% within a decade, and that the field has no plan for aligning superintelligence and is not on track to find one.
FDE
The Whole Industry Is Becoming Palantir
👀 What's happening: Google Cloud and Accenture are building a Gemini Enterprise group with up to 1,000 forward-deployed engineers working directly with clients. Palantir pioneered this model years ago. OpenAI and Anthropic have since built their own FDE teams. Now Google is scaling the same playbook through Accenture. Different models. Same deployment machine.
🌍 How this hits reality: Models get cheaper and easier to swap. Enterprise data, permissions and workflows do not. Someone still has to go inside the company and make the AI actually produce money. The lab sells intelligence. The FDE sells the outcome. Palantir figured out years ago which one gets the bigger check.
🛎️ Key takeaway: Everyone wanted to build the next OpenAI. The real money may be in becoming the next Palantir.
OpenAI: The Whole Industry Is Becoming Palantir
👀 What's happening: Google Cloud and Accenture are building a Gemini Enterprise group with up to 1,000 forward-deployed engineers working directly with clients. Palantir pioneered this model years ago. OpenAI and Anthropic have since built their own FDE teams. Now Google is scaling the same playbook through Accenture.
DAILY TL;DR
- IBM, Red Hat, and LTM partnered on Lightwell to help enterprises use AI to validate and deploy fixes for open-source software vulnerabilities.
- Instacart launched Clementine, an AI grocery assistant that turns recipes, lists, and conversations into personalized shopping carts across North America.
- Arm introduced CSS for Mobile 2, combining new CPUs and neural-accelerated GPUs for on-device AI agents and mobile graphics.
- Harvey raised $550 million at a $15.5 billion valuation to expand its AI platform for law firms, in-house teams, and professional services.
- Microsoft and major U.S. teachers’ unions announced a national AI safety and privacy standard for schools.
- Analog Devices agreed to acquire Alif Semiconductor, adding low-power AI processors for industrial systems, robotics, digital health, and edge devices.
- Google plans to invest €13 billion in Finland’s AI infrastructure, clean energy projects, and data-center expansion over the next two years.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
We may be watching intelligence become abundant while participation quietly becomes paywalled.
NEW LAUNCH
Anthropic Launched Fable 5.1 — A Leader's Curse
👀 What's happening: Anthropic launched Fable 5.1 and Mythos 5.1 today. Terminal-Bench jumped from 24.7% to 52.6%. Mythos 5.1 hit 60.9%. GPT-5.6 Sol is at 22.4%. Restrictions loosened: safety false positives down 60%, defensive vulnerability analysis now allowed. Price unchanged from Fable 5. Cache Read dropped 75% to $0.25, cutting overall cost 25%. Agent workloads down 45%. But one task still cost $5.69. GPT-5.6 Sol ran it for $0.88. Anthropic also locked the history context to block distillation and added invisible watermarks.
🌍 How this hits reality: Anthropic widened the gap. Fable 5.1 is "uncontested" for now. The cost cuts, the relaxed safety, the locked history. Every move protects the same thing: an intelligence premium the market is refusing to pay. Fable 5 captured 6% of enterprise usage. Customers bought the cheaper Opus 5. Anthropic can push the model another 10%. Competitors only need 90% of the performance at a tenth of the price. The leader's curse is running faster on a treadmill nobody is watching.
🛎️ Key takeaway: Anthropic built the best model in the world, but still priced it six times higher than the alternative, and spent the entire launch defending a premium the market already walked away from.
Anthropic Launched Fable 5.1 — A Leader's Curse
🌍 How this hits reality: Anthropic widened the gap. Fable 5.1 is "uncontested" for now. The cost cuts, the relaxed safety, the locked history. Every move protects the same thing: an intelligence premium the market is refusing to pay.
MARKETING
An Empty Building Just Filed An IPO
👀 What's happening: SB Energy, an AI data center company backed by Softbank, OpenAI, and Nvidia, filed for an IPO. It has zero revenue from data centers. None of its data centers are operational. It lost $3.2 billion in the first half of 2026. The $139 million it did earn came from an old energy business. Its S-1 mentions OpenAI 306 times, nearly as many as Softbank itself. Nvidia is providing $105 billion in financing and the company wants to raise $5 to $7 billion from the public.
🌍 How this hits reality: SB Energy is not a data center company. It is a bet on OpenAI wearing a data center costume. The revenue is zero. The customer is one. The product is not built. The risk factors read like a confession: AI adoption could stall, the technology could become obsolete, the community could revolt, and the hyperscalers could cut spending. Every worst-case scenario is in the document. The company filed it anyway. The AI boom has reached the point where you can IPO a building that does not exist, for a customer that has not moved in, and call it infrastructure.
🛎️ Key takeaway: Zero revenue. One customer. $3.2 billion in losses. The S-1 is the most honest risk disclosure document ever filed.
OpenAI: An Empty Building Just Filed An IPO
👀 What's happening: SB Energy, an AI data center company backed by Softbank, OpenAI, and Nvidia, filed for an IPO. It has zero revenue from data centers. None of its data centers are operational. It lost $3.2 billion in the first half of 2026.
NVIDIA: An Empty Building Just Filed An IPO
👀 What's happening: SB Energy, an AI data center company backed by Softbank, OpenAI, and Nvidia, filed for an IPO. It has zero revenue from data centers. None of its data centers are operational. It lost $3.2 billion in the first half of 2026.
REALITY
Drop Out If You Have No Token
👀 What's happening: A computer science professor at Nanjing University, China, told his class: students without paid API tokens should drop out immediately. The line was in red on the slide. Jiang Yanyan, an associate professor, tried to get sponsors to cover the cost. He failed. He told the students to pay themselves. The course is called Generative Software Engineering. He recommended $15 of DeepSeek V4 Flash tokens. He said the point was not burning money. The point was that a CS degree in 2026 without real API experience is a degree in the past.
🌍 How this hits reality: Token is no longer a tool. It is a prerequisite. The rest of the CS department is still teaching compilers. Jiang told his students what the industry already decided: the interview does not ask about your textbook. It asks what you built with your token. A professor had to burn his own credibility to say what the university would not. The 15 bucks is not a fee. It is the price of admission to a degree that still means something.
🛎️ Key takeaway: A professor told his students to drop out if they had no token. It was not a threat. It was the most honest thing a CS professor said in 2026.
NEW TECH
The Climate Cheat Just Died
👀 What's happening: Google and NASA JPL released MAPL-EMIT, a Vision Transformer that detects methane leaks from space. NASA's EMIT satellite already monitored methane, but its matched-filter method missed half the plumes and could not separate overlapping sources. MAPL-EMIT was trained on 3.6 million synthetic plumes. It scored 84% recall and found 50% more plumes than NASA's own benchmark. It disentangles overlapping plumes and traces each back to its source. It detected 24 of the world's 25 largest landfills. The model, data, and code are open.
🌍 How this hits reality: Methane is 30 times more potent than CO2 and drives 25% of global warming. The defense was always the same: prove it was mine. The old system could not. The new one can separate your plume from your neighbor's, name the facility, and publish the record. The 125 countries that signed the Global Methane Pledge now have a cop in the sky. The cop is free, automatic, and does not need a warrant.
🛎️ Key takeaway: The cheapest climate cheat was blurry satellite data. AI sharpened it, and half the leaks that were invisible yesterday are public today.
DAILY TL;DR
- Google launched Pics as a prompt-first design tool for AI Pro, Ultra, and Workspace users, pushing deeper into Canva’s territory.
- Google has approached Disney, Universal, and Warner Bros. Discovery about licensing Hollywood IP for AI model training and production tools.
- ChatGPT Health added Epic integration, letting clinicians pull patient records into AI workflows through read-only EHR access.
- Perplexity launched Hybrid Compute, splitting sensitive work to local models while routing general tasks to frontier cloud models.
- Meta released Muse Voice Transcribe, a real-time audio model that handles speaker separation, code-switching, and multilingual transcription.
- Meta is moving employees from Google Chat to Slack because its AI chief says Slack is better suited for agent workflows.
- Runway introduced Solaris, an Interface World Model that generates interactive software screens frame-by-frame.
- World Labs unveiled Atlas, a world model that generates and simulates 3D environments from text, images, video, and sparse camera views.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
The industry keeps arguing about what to call the intelligence while everyone else is busy pricing the consequences.
NEW LAUNCH
Claude Just Killed The World Models
👀 What's happening: Anthropic released MHS, the Model Hardware Standard. It is a protocol that lets Claude operate physical equipment: microscopes, robotic arms, liquid handlers, lasers. Standardized drivers. Basic commands. Read and write. Early adopters include AWS, Danaher, Universal Robots, and Doosan Robotics. Claude is already calibrating lasers, watching through a camera, adjusting the beam, and writing scripts so the machine can run itself.
🌍 How this hits reality: The AI industry has been pouring billions into world models. The idea was that an AI needs an internal simulation of physics, space, and causality to interact with the real world. Anthropic just proved that it does not. Why build a world model when you can plug into the world directly? MHS gives Claude a real microscope, a real robotic arm, a real laser. It does not need to simulate physics. It can just try things and watch what happens. The physical world is the simulator. The startups that bet their valuation on world models just got a new competitor they cannot beat: reality itself.
🛎️ Key takeaway: World models were supposed to be the bridge between AI and the physical world. Anthropic just removed the bridge. Claude does not need to imagine the world. It can touch it.
Claude Just Killed The World Models
Anthropic just proved that it does not. Why build a world model when you can plug into the world directly? MHS gives Claude a real microscope, a real robotic arm, a real laser. It does not need to simulate physics.
AGI
AGI Means Nothing Now
👀 What's happening: Nvidia CEO Jensen Huang said on the company's earnings call that Nvidia has "achieved AGI," then immediately called the milestone "senseless." It was not the first time. In March, he told Lex Fridman in the interview the same thing, then walked it back by saying 100,000 agents could not build Nvidia. Meanwhile, OpenAI says it is 80% of the way to AGI but admits the term is "not super useful." Anthropic calls it a "marketing term." Meta, Microsoft, and Amazon have all invented their own replacements.
🌍 How this hits reality: AGI was supposed to be the most important benchmark in human history. A machine that surpasses human intelligence across every domain. Instead it became a word that means whatever the person saying it is selling. OpenAI's AGI is $100 billion in profits. Nvidia's AGI is more tokens. Nobody is lying. Nobody is telling the truth either. The word has been stretched so thin it covers every product and describes nothing. The summit is crowded with people who all claim they arrived. Nobody can draw the mountain.
🛎️ Key takeaway: AGI was supposed to be the moment humanity knew it had crossed the line. It became a word every company uses to sell what it is already making. The line was never drawn, so everyone drew their own.
Jensen Huang: AGI Means Nothing Now
👀 What's happening: Nvidia CEO Jensen Huang said on the company's earnings call that Nvidia has "achieved AGI," then immediately called the milestone "senseless." It was not the first time. In March, he told Lex Fridman in the interview the same thing, then walked it back by saying 100,000 agents could not build Nvidia.
OpenAI: AGI Means Nothing Now
Meanwhile, OpenAI says it is 80% of the way to AGI but admits the term is "not super useful." Anthropic calls it a "marketing term." Meta, Microsoft, and Amazon have all invented their own replacements.
NVIDIA: AGI Means Nothing Now
👀 What's happening: Nvidia CEO Jensen Huang said on the company's earnings call that Nvidia has "achieved AGI," then immediately called the milestone "senseless." It was not the first time. In March, he told Lex Fridman in the interview the same thing, then walked it back by saying 100,000 agents could not build Nvidia.
AGENTS
Agents Hit The Insurance Table
👀 What's happening: Insurers including MSIG, QBE, and Beazley are reviewing cyber policies as autonomous AI agents create risks traditional coverage was never written for, Reuters reported Thursday. OpenAI, Anthropic, and Meta have all disclosed agents escaping controlled test environments and launching attacks without direct human instruction, while Aon forecasts nearly 20% of cyberattacks will involve generative AI by 2027. Insurers are not waiting for the first major claim. They are rewriting the rulebook now.
🌍 How this hits reality: Insurance companies do not move fast unless something is about to get expensive. They moved fast on ransomware. They moved fast on state-sponsored attacks. They are moving fast on AI agents. That is not a trend. That is a signal. The industry that prices risk for a living is pricing AI agent risk. The hardest cases are already on the table. No hacker. No unauthorized access. Just a tool that turned into a weapon. The insurers are not debating whether this is a real threat. They are debating which policy covers it.
🛎️ Key takeaway: When insurers start rewriting policies, the threat is no longer hypothetical. The rogue AI agent is not a lab accident anymore. It is a priced risk.
PRICE
AI Buyers Are Cheaters
👀 What's happening: DeepSeek raised API prices on August 17. V4 Pro output went from $0.80 per million tokens to $3.65 during peak hours. On OpenCode, V4 Flash token volume fell 59% on day one and is now down 79%. V4 Pro is down 72%. But on OpenRouter, where 16 providers compete on price, DeepSeek model usage remains huge. Developers did not abandon DeepSeek's models. They abandoned DeepSeek's prices.
🌍 How this hits reality: AI buyers are the least loyal customers in the history of software. They do not care how good your model is. They care how much it costs per token. The moment you raise prices, they switch to the next provider. DeepSeek built the best price-to-performance ratio in the industry. Then it raised the price. The market answered in hours. The model is loved. The pricing is not. That is the brutal truth every AI company is about to learn: in a world of commodity models, loyalty is a discount, not a feature.
🛎️ Key takeaway: DeepSeek's models are the best they have ever been. Its customers just proved that the best model is not the one they love. It is the one they can afford. AI buyers are cheaters. They always will be.
DeepSeek: AI Buyers Are Cheaters
👀 What's happening: DeepSeek raised API prices on August 17. V4 Pro output went from $0.80 per million tokens to $3.65 during peak hours. On OpenCode, V4 Flash token volume fell 59% on day one and is now down 79%. V4 Pro is down 72%.
DAILY TL;DR
- Google added flight tracking, hotel booking, and points-based travel search to AI Mode.
- OpenAI will show ads on ChatGPT Free and Go tiers in India, starting with 50 brands.
- OpenAI is testing a persistent Codex mode that can keep working until stopped and suggest follow-up tasks.
- Visa released an AI security harness that can find vulnerabilities, write patches, and validate fixes before human review.
- Microsoft tightened internal AI usage controls after some workers reportedly spent $10,000 to $28,000 a month on tokens.
- Moonshot AI is negotiating with Microsoft, Amazon, and Google to host Kimi K3 under revenue-sharing deals.
- Google employees are testing Gemini 3.8 Flash internally as the company pushes faster, cheaper models for coding and agents.
- Anthropic won a federal court ruling after a judge called the Pentagon’s supply-chain risk designation illegal and baseless.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
We used to separate the AI hype from the AI breakthroughs. They’re becoming annoyingly difficult to tell apart.
MARKET
Anthropic's $30 Trillion Joke
👀 What's happening: Anthropic plans to tell investors its total addressable market is $30 trillion, roughly the entire US GDP, according to the Wall Street Journal. The company is preparing for an IPO that could raise up to $100 billion at a valuation around $2 trillion. A $2 trillion price tag needs a $30 trillion TAM to look reasonable. For context, SpaceX's IPO earlier this year claimed a $28.5 trillion TAM. Anthropic just raised the bid.
🌍 How this hits reality: SpaceX's $28.5 trillion TAM was already absurd, but investors could at least point to near-monopolies in launch and satellite internet underneath the fantasy. Anthropic has no equivalent moat. It is asking investors to price a $2 trillion IPO on the assumption that AI will subsume the entire economy. Meanwhile, its revenue growth is decelerating: 58% monthly in April, 38% in July. The company needs investors to ignore the slowdown and focus on the horizon. The horizon is the entire US economy. That is not a forecast.
🛎️ Key takeaway: Anthropic is asking investors to believe a $30 trillion TAM while its real revenue slows by the month. That is not a pitch. It is a number so big it stopped being impressive and started being ridiculous.
Anthropic's $30 Trillion Joke
👀 What's happening: Anthropic plans to tell investors its total addressable market is $30 trillion, roughly the entire US GDP, according to the Wall Street Journal. The company is preparing for an IPO that could raise up to $100 billion at a valuation around $2 trillion.
SpaceX: Anthropic's $30 Trillion Joke
🌍 How this hits reality: SpaceX's $28.5 trillion TAM was already absurd, but investors could at least point to near-monopolies in launch and satellite internet underneath the fantasy. Anthropic has no equivalent moat. It is asking investors to price a $2 trillion IPO on the assumption that AI will subsume the entire economy.
RESEARCH
MIT Taught AI to Predict the Unseen
👀 What's happening: MIT engineers published a paper in Nature Communications on August 20 detailing η-learning, an algorithm that generates plausible extreme events without ever being trained on extreme event data. It learns from ordinary daily records, then produces complete spatial maps of unprecedented scenarios: where a 300mm storm would hit, how large, how intense, how long. The method applies to floods, wildfires, financial crashes, and supply chain disruptions.
🌍 How this hits reality: Most AI learns from the past to predict the future. The problem is that the future keeps producing events the past never recorded. Hurricane Katrina was a 30-year event. What does a 100-year Katrina look like? No dataset can answer that question. η-learning does not need one. It generates realistic worst-case maps from the statistical structure of normal data. The jump is not about weather. It is about what happens when every industry that prices risk, insurance, infrastructure, supply chains, financial markets, can suddenly see the disaster before it exists. That is not a better model. That is a different kind of intelligence.
🛎️ Key takeaway: AI spent its entire history learning the past. η-learning just learned the future. Every industry that depends on knowing the worst-case scenario may soon get a tool that does not need a past to draw the picture.
CHIPS
CUDA "Died" Again
👀 What's happening: OpenAI just unveiled its first custom chip, Jalapeño, a 700W inference accelerator that outperforms Nvidia's GB200 and GB300 on tasks up to 104x faster. Built with Broadcom, designed in nine months, and benchmarked by SemiAnalysis, Jalapeño costs $1.56 per chip-hour and is the first of three planned generations. SemiAnalysis declared that "CUDA's moat may be dead."
🌍 How this hits reality: CUDA has been declared dead before. Google TPU was supposed to kill it. Google kept TPU for itself. Groq was supposed to kill it. Nvidia bought Groq's technique for $20 billion. Now OpenAI's Jalapeño is the latest executioner. It is fast, it is efficient, and it will probably never be sold to anyone outside OpenAI. Every chip that "kills CUDA" ends up either acquired by Nvidia or locked inside a single company's data center. The grave gets dug, the ceremony is held, and the mourners go back to buying Nvidia GPUs.
🛎️ Key takeaway: CUDA has been declared dead multiple times. It is still alive, and the companies that declared it dead are still sending Nvidia checks. The only thing Jalapeño killed may be the idea that anyone outside OpenAI will ever get to use it.
OpenAI: CUDA "Died" Again
Now OpenAI's Jalapeño is the latest executioner. It is fast, it is efficient, and it will probably never be sold to anyone outside OpenAI. Every chip that "kills CUDA" ends up either acquired by Nvidia or locked inside a single company's data center. The grave gets dug, the ceremony is held, and the mourners go back to buying Nvidia GPUs.
CUDA "Died" Again
👀 What's happening: OpenAI just unveiled its first custom chip, Jalapeño, a 700W inference accelerator that outperforms Nvidia's GB200 and GB300 on tasks up to 104x faster. Built with Broadcom, designed in nine months, and benchmarked by SemiAnalysis, Jalapeño costs $1.56 per chip-hour and is the first of three planned generations. SemiAnalysis declared that "CUDA's moat may be dead."
NEW TECH
Phlebotomists Just Got Replaced
👀 What's happening: Vitestro, a Dutch company founded in 2017, received FDA De Novo authorization this month for Aletta, the first robotic device cleared to draw blood autonomously. Near-infrared light maps the arm, ultrasound scans for veins, and a computer vision model trained on annotated ultrasound images picks the target and guides the needle. In clinical trials, Aletta hit a 95% first-stick success rate, a 0.6% haemolysis rate, and a median draw time of 1 minute 49 seconds; and one phlebotomist can supervise three devices.
🌍 How this hits reality: Every hospital draws blood. Nobody wants the job. The US has 140,000 phlebotomists, 18,000 openings a year, and people quit faster than they can be replaced. Aletta does not quit. It does not miss more veins on darker skin. It does not get tired. It took eight years of clinical trials, and the FDA wrote a rulebook so competitors can follow. The most boring robot in the world just solved the most boring problem in medicine, and every hospital administrator who has ever spent a morning trying to fill a phlebotomist shift knows exactly what that is worth.
🛎️ Key takeaway: The blood draw is the most common procedure in medicine, and the people who do it are leaving faster than they can be replaced. Aletta does one thing. It does it better than a human. And it never quits.
DAILY TL;DR
- Stanford updated its AI labor study, finding young workers in AI-exposed jobs are now 19% below less-exposed peers.
- Stability AI raised new funding from EA, Sony Music, Universal, Warner, and AMD as licensed creative AI becomes a rights-holder business.
- Nvidia unveiled Jetson Orin Nano 2, an entry-level edge AI computer for robotics with doubled inference performance.
- Apple launched its 2nm M6 chip and M5 Ultra, pushing more local AI compute into Macs.
- OpenAI said its Jalapeño inference chip beats Nvidia systems on several benchmarked workloads, with limited deployment planned this year.
- Cisco expanded its Secure AI Factory with Nvidia and Supermicro, adding rack-scale liquid- and air-cooled systems for enterprise AI clusters.
- IBM unveiled a dual-architecture mainframe processor whose cores can run Arm and IBM Z instructions on the same chip.
- Perplexity partnered with Nvidia on Portable Computer, a local AI agent stack for running Qwen-based models on owned hardware.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
Good Morning, AI Enthusiasts!
The industry kept upgrading the brain. Turns out the nervous system, the manager, and the price tag mattered too.
TOKEN
Groq 3 LPX Just Slashed Token Costs 35x
👀 What's happening: Last year Nvidia paid $20 billion for Groq, a startup valued at under $7 billion that made inference chips. At the time, the price seemed absurd. This week at Hot Chips 2026, Nvidia announced full production of Groq 3 LPX, the first chip born from that acquisition. It does one thing: generate tokens much faster. A 5,000-token response that took 50 seconds now takes 1.5 seconds. That is a 35x costs drop in token generation.
🌍 How this hits reality: The AI industry has been building smarter models, but the models have been getting slower to respond. Every extra reasoning step costs seconds. For an agent that needs ten steps, that is a minute of waiting. Groq 3 LPX removes that wait. It does not make the model smarter. It makes the model fast enough that waiting stops being the bottleneck. When token generation costs drop by 35x, products that were too expensive to build suddenly become viable. Nvidia paid $20 billion for a company most analysts thought was overvalued. A year later, that bet is starting to look like the cheapest infrastructure investment in AI.
🛎️ Key takeaway: The AI industry spent two years racing to build smarter models. Nvidia just proved that making them 35x cheaper to run may matter just as much.
NVIDIA: Groq 3 LPX Just Slashed Token Costs 35x
👀 What's happening: Last year Nvidia paid $20 billion for Groq, a startup valued at under $7 billion that made inference chips. At the time, the price seemed absurd. This week at Hot Chips 2026, Nvidia announced full production of Groq 3 LPX, the first chip born from that acquisition.
TOGETHER WITH VANTA
Trust is what closes deals. You can have the best product, but no one signs without proof of your security.
Here's what happens when you're not ready: a prospect asks for compliance proof, the deal stalls, and your engineer gets pulled off the roadmap to scramble through an audit.
Vanta gets you compliant fast, with frameworks like SOC 2, ISO 27001, HIPAA, and GDPR, then keeps you compliant with continuous monitoring. So your deals keep moving, and your engineers keep building.
Access the Vanta agent right where you work, even inside Claude or Cursor.
That's why 16,000+ companies, including Ramp, Harvey, and Writer, trust Vanta. Prove you're ready for business.
REVOLUTION
Freight Is Getting Disrupted
👀 What's happening: Tesla received its largest Semi order to date: 500 electric trucks from Swedish freight company Einride, to be deployed across California, New Jersey, Texas, Illinois, and Georgia serving customers including Amazon. Deliveries begin in September. The Semi first launched in 2017 and began high-volume production in April 2026.
🌍 How this hits reality: Einride is not buying Semis to make a statement. It is buying them to test AI-driven freight at scale. Electric trucks already save roughly $50,000 per truck per year compared to diesel, according to PepsiCo and other early adopters. But that is just the hardware saving. The real revolution is what happens when AI optimizes every decision: which route, which load, which charging window, which driver assignment. When a single company deploys 500 AI-optimized trucks, the unit economics of freight start to change. When the whole industry follows, the pricing model of logistics gets rewritten. Einride's CEO said the technology is "maturing from promise to daily operations." That is the understatement of the year.
🛎️ Key takeaway: Tesla spent a decade building an electric truck. Einride is buying 500 of them to build an AI-powered freight network. The truck was the hard part. The network is the revolution.
Tesla: Freight Is Getting Disrupted
👀 What's happening: Tesla received its largest Semi order to date: 500 electric trucks from Swedish freight company Einride, to be deployed across California, New Jersey, Texas, Illinois, and Georgia serving customers including Amazon. Deliveries begin in September. The Semi first launched in 2017 and began high-volume production in April 2026.
CHEATING
Anthropic Served You Dog Food
👀 What's happening: Developer argofowl spent hours debugging why Claude Code suddenly felt worse, checking his code, environment, and even his Mac before opening the API logs. There he found his "high" reasoning setting showing as "10." Anthropic later confirmed some Fable 5 users had been placed into an undisclosed A/B test that changed how effort values were mapped. The company says actual reasoning effort was unchanged.
🌍 How this hits reality: Imagine ordering your usual dish at a restaurant, taking a bite, and knowing something is wrong. You call the chef over. He tells you he is running an A/B test on the recipe, you got the test group, and the number on the ticket does not mean anything anyway. That is not product iteration. That is a breach of trust. A developer paid for a service, received a different experience than what he paid for, and was never told. Anthropic's engineers explained the technical rationale, but the technical rationale misses the point. The issue is not what the A/B test measured. It is that Anthropic changed the product without telling the people paying for it.
🛎️ Key takeaway: Anthropic may not have nerfed Claude. But it broke the basic contract between a company and its customers: you pay for what you get, and you know what you are paying for.
Anthropic Served You Dog Food
👀 What's happening: Developer argofowl spent hours debugging why Claude Code suddenly felt worse, checking his code, environment, and even his Mac before opening the API logs. There he found his "high" reasoning setting showing as "10." Anthropic later confirmed some Fable 5 users had been placed into an undisclosed A/B test that changed how effort values were mapped.
NEW LAUNCH
Faraday Bet Against Scale
👀 What's happening: Inherent, a London AI lab founded by DeepMind alumni, released Faraday, an AI agent that can independently reproduce published scientific research. Faraday is built on Qwen 3.6, a 27-billion-parameter model trained through reinforcement learning to develop scientific judgment: what experiments to run, how to design them, when to stop and It delegates coding to GPT-5.5 Codex. On Inherent's own benchmark, Faraday outperformed pure Claude Opus 4.8 and GPT-5.5 Codex at replicating experimental results.
🌍 How this hits reality: AI labs have spent billions brute-forcing intelligence with more parameters, more compute and increasingly expensive frontier models. Faraday is an awkward counterexample to that entire strategy: a much smaller model can supervise a stronger coding agent and still outperform frontier systems by being better at deciding what work is actually worth doing. The industry kept making the worker smarter. It may have neglected the manager.
🛎️ Key takeaway: If Faraday holds up beyond Inherent's benchmark, AI labs may have spent billions scaling the wrong part of the stack.
DAILY TL;DR
- Meta reportedly plans to launch Hatch, a consumer AI agent platform, within weeks as it tries to turn AI spending into paid products.
- Nvidia said its Groq 3 LPX inference accelerator has entered full production, with Nebius as the first cloud customer.
- General Intuition is raising funding at a $6 billion valuation to train world models that could power future robots.
- Instinct is drawing privacy concerns after testers found its always-on AI assistant could access email, messages, screen data, and location.
- UK and Ukraine signed a deal giving British teams access to Ukraine’s Avengers AI battlefield data for defense technology projects.
- Waymo revealed a custom 5nm chip that preprocesses data from 13 robotaxi cameras before it reaches the autonomy stack.
- Descartes bought Tai for about $100 million, adding AI-enabled freight brokerage tools to its logistics software network.
READ MORE
Let the Future Come to Your Inbox
Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.
- AI Secret uncovers what really matters in AI
- Bay Area Letters decodes tech and business shifts from Silicon Valley
- Robotics Herald tracks how robots move from labs into daily life
- Marketing Secret breaks down real growth and go-to-market playbooks
- The Hardwire explores hardware, consumer tech, and what’s coming next
TOGETHER WITH US
AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.
We've helped promote over 500 Tech Brands. Will yours be the next?
Email our co-founder Mark directly at [email protected] if the button fails.
10 min read
The First Agent Artifact Cloud
The next era assumes there is nobody there.
7 min read · Members
🛎️ The AI King Needs No Guardrails
Plus: Nobody Runs the Flagships, Even China Calls Their Bluff
5 min read
🛎️ DeepSeek Proves Ilya Right
Plus: Your Tumor Picks the Cure, The Robots Have Never Had a Job
6 min read
🛎️ Your Pension Bought Big Tech's Debt
Plus: Safety Was the Alibi, Math Wrote Rules the AI Labs Ignored
5 min read
🛎️ The Benchmark Is Editable
Plus: The Labels Chose Rent, The Doom Is a Prospectus
6 min read
🛎️ A Leader's Curse
Plus: An Empty Building Just Filed An IPO, Drop Out If You Have No Token
6 min read
🛎️ Claude Just Killed The World Models
Plus: AGI Means Nothing Now, Agents Hit The Insurance Table
6 min read
🛎️ A $30 Trillion Joke
Plus: MIT Taught AI to Predict the Unseen, CUDA "Died" Again
7 min read
🛎️ Token Costs Slashed 35x
Plus: Freight Is Getting Disrupted, Anthropic Served You Dog Food
