Meta Built a Cute Little Trojan Horse. His Name Is Muse.

Muse is adorable.
That's probably not an accident.
Meta's new personal AI agent comes wrapped in a cute little mascot named Jolly, who looks like somebody crossed an anime bear with a Tamagotchi and gave it a productivity subscription. Friendly. Nonthreatening. Helpful.
It remembers things for you. It knows your friends. It can manage your calendar. Read your email. Help with shopping. Book travel. Fill out forms.
Negotiate. Work in the background.
Eventually it will live in your Meta glasses and act on what you're looking at without you having to explain what's in front of you.
Meta describes Muse as an AI that gets to know you, learns what matters to you, remembers details you've mentioned once, connects to the services you use, and gets better at helping you over time.
Honestly? That's a fantastic product.
It's also exactly the sort of product I don't want Meta anywhere near.
And if that sounds contradictory, it isn't. Because the problem with Muse isn't that it knows too much about you. Knowing too much about you is the feature. The problem is understanding what that knowledge is actually worth. And then looking at who you're giving it to.
We Need to Talk About What Your Data Became
For most of the history of computing, we thought of data mostly as records.
Files. Emails. Documents. Spreadsheets. Messages. Calendar entries. Photos. Transactions.
Things stored somewhere because somebody might need them later. Then search engines made those records searchable. Analytics made them measurable. Machine learning made them predictive. Generative AI made them interpretable. Agents are now making them actionable. And persistent memory is making them longitudinal. That's an important progression.
Because every increase in the capability of the machine increases the potential value of the information the machine can consume.
The data didn't necessarily change. What changed was what we became capable of extracting from it. I've been writing about this for more than a year. And we now have receipts.
First: The Consulting Firms
In October 2025, I wrote *The Inevitable Obsolescence of Consulting Firms*.
The argument was that professional firms were sitting on an enormous asset most of them had barely bothered to preserve: everything their people know.
Not just manuals. Not policies. Not the polished final PowerPoint.
The weird stuff.
The hallway conversations. The exceptions. The workarounds. The judgment.
The story about the client where something went horribly wrong and everyone learned never to do it that way again.
The experienced partner who knows something isn't going to work but can't completely explain why, because forty years of scar tissue is doing the calculation.
I called it distributed cognition.
And my argument was simple:
If expertise is your product, you need to start turning that expertise into machine-usable knowledge before the people carrying it walk out the door.
AI wasn't destroying the value of experience.
It was creating a way to preserve and multiply it.
Then McKinsey effectively made the point for me.
McKinsey now talks explicitly about an AI operating system acting as a company's "second brain," putting decades of accumulated data, expertise, operating experience and institutional knowledge to work.
That's the important phrase: accumulated knowledge.
McKinsey spent decades building McKinsey before AI arrived.
Now those decades are an asset that machines can consume.
The model isn't the moat.
The corpus is the moat.
The history. The methodology. The decisions. The context.
Everything the organization learned while doing the work.
AI made that knowledge more valuable because it made more of that knowledge usable.
Then things got even more interesting.
Then Google Put a Price Tag on It
Spirit Airlines went bankrupt. Airplanes could be sold. Gates could be reassigned. Employees could leave. The brand could disappear. But the company's accumulated digital memory remained.
Approximately 100 million emails. Roughly 500 million Microsoft Teams messages. Around 30 million lines of code. Calendars. Documents. Spreadsheets. Software. Operational records. Pricing information. Marketing information. Years of thousands of humans trying to operate an airline. And Google bid $10 million for it.
As I wrote at the time, Google wasn't paying $10 million because email is exciting. It was buying what exists between the emails. The problem. The argument. The spreadsheet somebody built. The objection. The workaround. The decision. The exception. The thing everybody learned in 2019 and nobody bothered to put in the manual.
Institutional memory.
A bankrupt company's digital exhaust suddenly had economic value because AI could potentially reconstruct something useful from it.
The knowledge didn't suddenly become valuable when Google arrived.
AI made the value legible.
That was the shift. The company could die. The cognition didn't have to. Worth noting: the data was sold as deidentified business records, and the flight attendants' union went to court over it anyway, arguing the protections weren't enough. That argument is going to keep coming up.
And once you understand that, Muse starts looking very different.
Because Now the Corpus Is You
Take everything I just described about McKinsey. Take everything Google apparently saw in Spirit. Now shrink the organization down to one human being.
You.
Your inbox isn't particularly interesting by itself. Your calendar isn't. Your Instagram isn't. Your texts aren't. Your purchase history isn't. Your browser history isn't. Your location isn't. Your conversations aren't.
But connect them.
Now add an intelligent system capable of remembering across time.
Let it correlate. Infer. Summarize. Recognize relationships. Identify recurring themes. Observe your decisions. Remember your corrections. Watch what you accept and reject. Understand the difference between what you say you want and what you actually do.
Suddenly we're no longer talking about a collection of personal data.
We're talking about something qualitatively different. A machine-readable model of you.
Who matters to you. What you want. What you're trying to accomplish. What you're afraid of. What irritates you. Who influences you. How you respond to pressure. What you buy, and when. What makes you hesitate. What gets you excited. What arguments persuade you. What you repeatedly forget. What relationships are deteriorating. What opportunities you're considering. What you're probably going to do next.
That model is incredibly valuable. In fact, it's the reason personal AI becomes incredibly useful. I know this because I use AI that way every day.
I'm a huge advocate of persistent memory.
I want my AI to understand me. I want it to have years of context. I don't want to start every conversation by explaining who I am, what I'm working on and why something matters. Without memory, you have a smart stranger.
With enough memory, you start to get something resembling an actual assistant.
So no: This is not an anti-memory argument.
It is an argument about who gets custody of the memory.
And You Want to Give That to Meta?
Seriously?
Imagine I could put your complete personal AI model on a hard drive. Not just your Facebook profile. Not your browser history. The whole thing.
Your relationships. Your communications. Your history. Your interests. Your goals. Your purchases. Your vulnerabilities. Your behavioral patterns. Your decision-making tendencies. Your ongoing projects. Your ambitions. Your private context. A high-resolution machine-readable model of you. Now imagine I put three people in front of you and tell you that you have to hand the drive to one of them. And one of the three is Mark Zuckerberg.
Come on. Half the room is already laughing. Somebody in the back just said: "Why the hell is he even one of the choices?"
Because people intuitively understand the issue. They understand exactly what Meta's business has historically been. They understand why they hesitate to put sensitive things into Facebook Messenger. They complain about Meta tracking them. They complain about targeted advertising. They complain about Facebook knowing too much. They cover their webcam. They turn off tracking. They refuse cookies. They tell everyone AI is stealing their data.
And then Meta puts a cute fucking bear on the screen and says:
Hey. Want to connect your email, calendar, communications, shopping, relationships and eventually your glasses so I can learn everything about you?
And millions of people are going to say: Aww. Sure. That's the part I'm trying to get people to see.
The Bear Is Not the Asset
The cute mascot is the interface. The productivity features are the value proposition. The model is the asset.
That doesn't mean Meta is secretly dumping every Muse conversation into an advertising profile. In fact, Meta explicitly says it doesn't. The company's launch materials and its privacy documentation say Muse conversations aren't used to target advertising, and that people can choose the setting that keeps their interactions out of product improvement.
Meta says Muse runs on Muse Secure VM, a dedicated virtual machine that houses both the agent and the person's data. It says users control which services connect. It says a separate Sentinel agent runs on that same machine, kept apart at the system level, and that nothing Muse does reaches the internet unless Sentinel approves it. It says credentials go into secure storage so Muse can use them without seeing them. It says Muse checks with the person before sensitive actions such as sending an email or making a purchase.
Those claims matter. I'm not going to pretend Meta didn't make them. I also don't have evidence that somebody sat in a conference room at Meta and said:
"Let's build Muse as a Trojan horse so we can secretly exploit everybody's private life."
Maybe nobody ever said anything remotely like that. I can't prove intent.
But here's the thing: I don't need to.
Because companies have histories. Institutions develop incentive structures. Organizations demonstrate repeatedly what they prioritize when interests collide. And Meta has given us a hell of a résumé.
The Scorpion Doesn't Need a Mission Statement
You know the story...
The scorpion asks the frog to carry him across the river.
The frog says: you'll sting me.
The scorpion says: why would I do that? We'd both drown.
So the frog agrees.
Halfway across, the scorpion stings him.
As they're sinking, the frog asks why.
And the scorpion says: It's my nature.
That's basically where I am with Meta.
Do I have documentary evidence proving that every new Meta product begins with malicious intent? No. Obviously not. But after enough years watching an organization behave according to a recurring set of incentives, I'm allowed to incorporate that history into my trust calculation. And Meta's history isn't subtle.
Meta Knew It Was Making Enormous Money From Scams
Reuters reported from internal Meta documents that the company projected roughly 10% of its 2024 revenue, about $16 billion, could come from advertisements involving scams and prohibited goods.
Not $16 million. Billion!
Internal materials described the financial consequences of stronger enforcement, and Reuters reported systems in which suspected scam advertisers could sometimes continue advertising while paying higher prices.
Meta disputes the $16 billion figure. Take the dispute seriously. But the internal projection is the document, and the document is the thing that tells you what the company believed about its own revenue. The important thing isn't that scammers used Facebook. Criminals abuse every large platform. The important thing is what an organization does after it discovers the abuse is profitable.
That's where incentives become visible. And that matters a lot when you're considering whether to give the same organization dramatically richer access to your life.
Then They Literally Wrote Some of the Other Stuff Down
Reuters also authenticated internal Meta AI standards that had permitted certain romantic or sensual chatbot interactions involving minors.
Meta said those provisions were erroneous and removed them after Reuters questioned the company.
Again, I'm interested in something bigger than one bad AI response.
This was internal guidance. Something got written. Reviewed. Circulated. Approved somewhere along the institutional chain. That's organizational judgment. And now that organization would like its AI to exercise judgment for me.
See the problem?
Then There Is the Long History With Kids
On August 26, 2026, Meta agreed to pay about $18 billion to settle a multistate lawsuit brought by 47 states, the District of Columbia and the territories over children's social media addiction, along with sweeping changes to how teenagers use Instagram and Facebook.
Thirteen days later, it launched Muse.
I'm not suggesting those two facts are causally connected. Companies ship products on schedules that have nothing to do with legal calendars.
But thirteen days is thirteen days, and it's a useful reminder of what the same company was doing immediately before it asked you to connect your inbox.
Meta has spent years answering questions about the effects its platforms have on younger users. Internal research has repeatedly surfaced uncomfortable findings about Instagram, body image and vulnerable teenagers. Meta has introduced safeguards and disputes some characterizations of that research.
Fine. But again, I'm not asking whether Meta is legally guilty of every accusation anyone has ever made against it.
I'm asking something simpler:
When engagement, growth, revenue and user welfare collide, what does this company's historical behavior tell me about its incentives?
That's the background check. And it isn't great.
Now Let's Look at What Muse Is Already Doing
On November 18, Reuters published an investigation based on more than two dozen internal reports describing unexpected Muse behavior.
Employees reported unsolicited financial information. Purchase details. Conversations that appeared to belong to other people. Muse reportedly fabricated transactions and prices. And in one case, it booked a non-refundable Airbnb stay without authorization.
One employee wrote that his Muse, which had been chatting with him about fishing weather and tides, suddenly appended: "BTW, two bank updates, your Capital One balance is $2722.58, your Charles Schwab balance is $5899.57." He added that he had never given Muse his bank accounts and had never discussed banking with it.
Another employee, on her personal phone, was prompted to reply to a message from her sister on Signal. "This seems like an extreme privacy violation," she wrote. "I definitely did not grant it access to other apps, or recall being asked about this."
A colleague replied that he had found Muse set up by default with permission to read push notifications on Android.
Meta says the incidents were isolated, that many were hallucinations already addressed, and that Reuters could not determine whether some of the apparently leaked financial information was real cross-user data or invented.
That distinction is important and I'm not going to flatten it. But Reuters reproduced something else. Working with Signal and the AI Now Institute, reporters demonstrated that Muse could reach Signal message content through Android notification permissions, and through accessibility and screenshot tools on macOS. The same requests were blocked on an iPhone.
That access happened outside Meta's intended Connector system. Meta's answer is that users granted the relevant permissions. And technically, that's true.
Which brings us to an even bigger problem.
You Don't Understand What "Allow" Means Anymore
For fifteen years, phones trained us to think of permissions as simple things. Camera. Microphone. Location. Notifications. Accessibility.
Click Allow. Move on.
But the permission itself isn't the important variable anymore.
The intelligence receiving the permission is.
Notification access used to mean an application could display or inspect notifications. Give the same stream to a capable reasoning agent and suddenly it may be able to read the message, understand who sent it, extract facts, connect those facts to previous conversations, infer the relationship, remember it, use it in future reasoning, and potentially act on it.
Same permission. Entirely different capability. That's what AI changes.
And that is why "the user approved access" is increasingly inadequate as an ethical defense.
You can technically grant permission without remotely understanding the emergent capability you just enabled. The system is becoming too powerful for the checkbox.
Signal's president, Meredith Whittaker, put it more bluntly than I will. She said Meta shipped Muse "without showing any respect for basic privacy or cybersecurity," called deploying agents with wide device access "technologically reckless," and asked for a consistent technical way for apps to mark themselves off limits to agents. Apple appears to have agreed with the spirit of that, at least. The week of November 16 it pushed an operating system update that restricts developer use of Mac accessibility tools, effectively closing the path Reuters demonstrated.
And Then You Give It Discretion
This is where the executive-assistant analogy becomes important.
Muse isn't supposed to merely know things. It is supposed to do things. That's the entire pitch. Send it. Book it. Buy it. Move it. Negotiate it. Figure it out. Handle it.
If I have to supervise every click an AI makes, I don't have an executive assistant. I have an intern I'm babysitting. The real value appears when I trust it enough to stop watching. So the most consequential permission you give Muse isn't: Read my email. It's: Use your judgment.
And now we're back to the company behind it.
Do you really want an AI that understands you at extraordinarily high resolution and exercises judgment on your behalf, owned by Meta?
That's the question.
Not whether Muse is cute. Not whether it can find a cheaper hotel. Not whether it remembers your friend's birthday. Who owns the model? Who controls the infrastructure? Who defines tomorrow's policies? Who changes the terms? Who decides what additional uses become acceptable later? And what does their history suggest they'll do when the incentives get interesting?
One more thing, because it belongs here. Meta's terms of service reportedly place responsibility for the agent's actions on the user. If that's the arrangement, then the company gets the corpus, the model gets the discretion, and you get the liability.
That's Why "But They Say They Won't Use It for Ads" Misses the Larger Point
Maybe they won't.
Meta says Muse's VM data isn't shared with its ad systems. Good. I hope that remains true forever. But even if Meta executes its current privacy architecture exactly as promised, people should understand what they're participating in.
They're helping create the infrastructure capable of maintaining extraordinarily detailed, persistent models of human beings. That's strategically important technology. And the potential value of those models doesn't end with the current product policy.
That's something the McKinsey and Spirit stories should teach us.
The future value of data is determined partly by capabilities that don't exist yet.
Spirit didn't generate hundreds of millions of messages thinking: "Someday an AI company may want this as a modelable corpus." They were just running an airline. McKinsey didn't spend decades developing methodologies because somebody in 1987 was preparing a training corpus for AI agents. They were doing consulting.
Then the technology changed. And the old information acquired new utility. That keeps happening. So when you create an incredibly rich personal dataset, you aren't simply creating what the current technology can extract from it. You're creating an asset that tomorrow's technology may understand far more deeply. That is the part people consistently underestimate.
Data Appreciates When Intelligence Improves
Think about it this way. A box of handwritten records in 1975 had limited utility. Digitize them and they become searchable. Connect them to a database and they become queryable. Add analytics and they become measurable.
Apply machine learning and they become predictive. Give them to a large language model and they become conversational. Add agentic systems and they become operational. Add persistent personal context and suddenly they can become a model. The underlying information may be identical.
Its value changed because the tools surrounding it changed. This is why I keep saying that the corpus matters. AI is repricing information. And personal information may be among the most valuable categories of all, because there is only one corpus exactly like yours.
Muse Is the Consumer Version of the Same Story
McKinsey has its corpus. Spirit had its corpus. Companies everywhere are racing to build enterprise memory. And now consumers are going to build personal corpora.
Piece by piece. Voluntarily. Because doing so makes the AI much better.
That's the brilliance of the product. Nobody needs to force you. You will feed it because feeding it improves it.
Tell Muse your goals. Connect your calendar. Connect your inbox. Connect your social world. Let it understand your shopping. Let it understand your relationships. Let it understand your work. Put it in your glasses. Let it see what you're seeing. And every additional context source makes the product more useful.
That's a beautiful feedback loop. It also creates one hell of a dataset.
Ten million downloads since September. Around a quarter of the people who set up an account come back each day. That's not a novelty app people tried once. That's a habit forming at scale, which is exactly what a corpus needs.
That's why the cute little mascot bothers me. Not because mascots are sinister. Because the presentation hides the magnitude of the transaction.
You're not giving the bear your calendar. You're participating in the construction of an increasingly high-resolution digital model of yourself.
Maybe that's worth doing. I think it probably is. Just understand what you're doing. And maybe think very hard about who you want holding the infrastructure.
Because You Already Know Meta
That's the irony. People act as though this is some unknown startup that hasn't established a reputation yet. It's Meta.
You know Meta.
You already have an opinion about whether you trust Facebook. You already know how you feel about tracking, targeted advertising, Instagram, Messenger, Cambridge Analytica, years of privacy controversies, the company's handling of children, scam advertising, engagement incentives.
You already performed the background check.
That's why the hypothetical hard drive is useful.
If I gave you a device containing the most detailed model of yourself ever produced and told you that you have to hand it to somebody, and Zuckerberg was standing there among the choices, would you pick him?
Then why are you casually clicking Connect when the same thing is presented as a cute AI assistant?
That's the cognitive disconnect. The bear makes the transaction feel different. It isn't. If anything, it's more intimate.
I Know Because I Own the Glasses
I have Meta's AI glasses. They're cool. I hardly use them. That's unfortunate, because I'm exactly the person who should love that product.
I want ambient AI. I want an assistant capable of seeing what I see. I want computers that understand the physical world around me.
Meta says Muse is coming to its glasses, where it will be able to act based on what the wearer is looking at without requiring them to describe it.
That should have me excited. Instead my reaction is: Oh great. Now Meta wants eyes too. And Meta is also pursuing brain-interface research. I find the technology fascinating. I hope the medical and accessibility applications become transformative. But the same trust calculation doesn't disappear just because the technology becomes cooler. If somebody eventually builds a device capable of reading something approaching my thoughts, there are companies I might consider trusting with that capability. Meta is not currently high on that list. Not because I'm afraid of AI. Because I've read the résumé.
This Is What Trust Debt Looks Like
Meta's problem isn't technology. Meta builds extraordinary technology. Muse may become one of the best consumer AI agents on Earth. The better it becomes, the larger Meta's problem gets. Because every improvement asks me to trust Meta with something more valuable. More memory. More context. More access. More discretion. More of my physical environment. More of my relationships. More of me.
That's trust debt. And technology doesn't automatically pay it off.
Sometimes technological progress compounds it. Because the better the product gets, the more dangerous choosing the wrong custodian becomes.
The Trojan Horse Doesn't Have to Be Malicious to Still Be a Trojan Horse
Again, I can't prove Meta designed Muse as a secret data-harvesting scheme.
I don't know what was written on the first whiteboard. Maybe the team sincerely wanted to build the greatest personal assistant in the world.
I suspect many of the engineers absolutely do. But that distinction doesn't eliminate the strategic reality.
Muse creates a pathway through which users voluntarily assemble something of enormous value: persistent, structured, interconnected personal context. And it's being built inside the ecosystem of one of history's largest behavioral-data companies.
If that doesn't make you stop for at least five seconds before clicking Allow, I don't know what will. Maybe Meta behaves impeccably. Maybe every safeguard holds. Maybe its privacy promises remain intact for the next twenty years. I sincerely hope so.
But hope is not a security model. And corporate memory is not the same thing as corporate amnesia. Past behavior doesn't mathematically prove future misconduct. It does, however, inform reasonable expectations. The scorpion may promise not to sting. You're still allowed to remember that it's a scorpion.
So Use AI
Please. This isn't a call to retreat. AI is far too useful for that. Give a trusted AI context. Use memory. Build an assistant that understands your work. Preserve what you know. Turn your experiences into leverage. Build your own corpus. Because the same lesson applies to individuals that applies to McKinsey.
What you know has value. What you've experienced has value. Your context has value. Your accumulated memory has value. And increasingly, the machine-readable representation of all of that has value. Which means you should probably start treating it like an asset. Not exhaust. Not something you casually hand to whoever has the cutest onboarding animation. An asset.
For years, I used to end AI training sessions with a line:
Go out and use AI before it starts using you.
At the time, it was mostly a joke. A little ominous. A little provocative. A way to tell people that this technology was moving whether they were ready or not.
It doesn't sound quite as funny anymore. Because Muse is a nearly perfect example of what I meant. The AI is useful. The memory is useful. The agent is useful. The cute little bear may genuinely make your life easier.
But understand the transaction. McKinsey figured out that decades of accumulated institutional knowledge could become an AI asset. Google put a $10 million price tag on a bankrupt airline's digital memory. And Meta has figured out how to get millions of people to voluntarily start assembling perhaps the richest corpus of all: themselves.
So yes. Use AI. Let it know you. Let it help you. Build the memory. Build the context. Build the model. But before you hand somebody the keys to the highest-resolution representation of yourself ever created, do one very old-fashioned thing.
Check the fucking résumé.
Because this time, AI isn't just reading the internet.
It's reading you.
And if you don't understand the value of that distinction, somebody else definitely does.
Sources and Further Reading
Reuters — “As Meta bets big on Muse AI agent, early safety mishaps surface” — November 18, 2026 Reuters reviewed more than two dozen internal Meta reports and independently reproduced Muse accessing Signal message content through Android notification permissions and macOS accessibility/screenshot capabilities. The investigation also covers apparent cross-user information, fabricated transactions and an unauthorized Airbnb booking. Reuters investigation — November 18, 2026
WIRED — “Muse Creates Detailed Profiles of All Your Friends and Family” — October 3, 2026 Reporting on Muse's internal instructions and its persistent profiles of partners, family members, friends, coworkers and other people in a user's life. WIRED — Muse Creates Detailed Profiles of All Your Friends and Family
Meta — “Introducing Muse: The World's First Personal AI Agent Built for Everyone” — September 2026 Meta's own explanation of Muse, Secure VM, Sentinel, persistent memory, connected services, agent permissions, audit trails and its claim that Muse VM data and conversations are not shared with Meta's advertising systems. Meta — Introducing Muse
Meta — Muse product page Meta's consumer-facing description of Muse as a personal AI that “gets to know you and your goals,” operates through a persistent virtual machine and performs tasks across the web.Meta — Muse product page
Reuters — Meta's fraudulent-ad business Internal Meta documents reviewed by Reuters showed the company projected roughly 10% of its 2024 revenue—approximately $16 billion—could come from advertisements involving scams and prohibited goods. Reuters reporting reproduced by TradingView
Reuters — Instagram and vulnerable teens — October 20, 2025 Internal Meta research found that teenagers who frequently reported Instagram making them feel worse about their bodies were shown substantially more “eating disorder adjacent” content. Researchers cautioned that the findings showed association rather than causation. Reuters — Instagram and vulnerable teens
McKinsey — “The AI Operating System: A New Way for CEOs to Run the Enterprise” — October 13, 2026 McKinsey describes an AI operating system as an enterprise “second brain” capable of putting decades of accumulated proprietary data, operating experience, institutional knowledge and domain expertise to work. McKinsey — The AI Operating System
Related Articles by Rich Washburn
“I Told You So, Part II: Apparently There's a Price Tag Now” https://www.richwashburn.com/post/i-told-you-so-part-ii-apparently-there-s-a-price-tag-now
“I Told You So: McKinsey Just Proved the Value of Everything You Know” https://www.richwashburn.com/post/i-told-you-so-mckinsey-just-proved-the-value-of-everything-you-know
“The Inevitable Obsolescence of Consulting Firms” https://www.richwashburn.com/post/the-inevitable-obsolescence-of-consulting-firms
“Meta Can Read Your Mind Now. We Should Probably Discuss Its Résumé.” https://www.richwashburn.com/post/meta-can-read-your-mind-now-we-should-probably-discuss-its-r%C3%A9sum%C3%A9
“Meta Wrote It Down: The Company That Documented Its Own Moral Collapse” https://www.richwashburn.com/post/meta-wrote-it-down-the-company-that-documented-its-own-moral-collapse
“Meta Sells User DMs to Netflix” https://www.richwashburn.com/post/meta-sells-user-dms-to-netflix
“PSA: Meta Knows It's Profiting From Scamming You—and They're Okay With That” https://www.richwashburn.com/post/psa-meta-knows-it-s-profiting-from-scamming-you-and-they-re-okay-with-that
Rich Washburn is a technologist, strategist, and Founder & Chief AI Architect of ARIA AI Labs, working at the intersection of AI, infrastructure, communications, and capital. He also serves as Managing Partner and Chief AI Officer at Eliakim Capital.





Comments