Serendipity35

If you’ve found yourself both fascinated and/or unsettled by the accelerating pace of artificial intelligence, THE AI DOC; OR HOW I BECAME AN APOCALOPTIMIST offers one way to lean into that tension rather than avoid it.

This week it was my film for the Film Matinee Club with Montclair Film. Our discussion after viewing the film was "spirited." Artificial intelligence certainly pushes people's intellectual and emotional buttons.

Directed by Daniel Roher and Charlie Tyrell and hosted by Roher. It is about making a documentary, and it is about AI, and it is also a personal narrative centered on Roher’s own fears about the future, especially as he and his wife contemplate having their first child in an AI-driven world.

Across all sources, the documentary’s expert roster includes top AI CEOs, pioneering researchers, alignment and ethics leaders, and public intellectuals. The film intentionally spans “doomers,” “optimists,” and “apocaloptimists,” giving a wide-angle view of the AI debate.

Roher becomes an Apocal (as in apocalypse) optimist because (spoiler alert) as he learns more and understands AI's capabilities, he begins to see more positive possibilities. And yet the answer to whether AI will cause the end of us or make our lives very much improved is still an open question. Even the experts don't know no matter what side they take on the AI debate.

The film very deliberately does not settle on a single answer about the future of AI. The film’s whole structure is built around the tension between optimism and existential risk, and it ends by embracing that unresolved state rather than resolving it./p>

The phrase that if an app is free, you are the product means that when an app doesn’t charge you money, it usually makes money from you instead. They do that mainly by collecting your data or selling your attention to advertisers.

If that is true, then how is AI changing what that means? It is a question that deserves several posts here to really answer.

Your behavior, preferences, and time become what is being monetized. Your data becomes the product. Free apps often gather your demographics, browsing or in-app behavior, location, interests, and habits. This information is then used to target ads or sold to third parties.

The addictive nature of app design keeps you scrolling, tapping, or watching so they can show you ads. You pay with time, not dollars. “Free” is a business model, not a gift.

I will give these companies a nod that running an app costs money (servers, engineers, storage). If you are not paying, the company must earn revenue another way. Ad-free options are becoming more common as a premium. You have probably noticed that on apps and also on video streaming services. You thought that paying for Amazon Prime meant no ads on the videos. Wrong. Free is often an illusion.

In the world of AI, the difference between free and paid tiers is more than a matter of convenience. It is also about identity and privacy.

Privacy becomes the hidden cost. Data is currency. Companies track you across apps and devices, build detailed behavioral profiles, and use algorithms to influence what you see. This raises concerns about autonomy and consent.

Is there no stopping them? As long as you agree to their terms, they have a lot of power. BUT you can read those terms and privacy settings more carefully. (They rely on the fact that many users don't read the terms or adjust their settings at all.) Educate yourself and understand how digital ecosystems make money. You can choose paid or privacy-focused alternatives. And you can remove the app entirely from your life.

I see comparisons of using AI to using social media platforms. I don't think AI data is the same as social media data. Social media platforms monetize your attention. The longer you scroll, the more ads they can show. AI chatbots operate on a different axis. Your prompts aren’t just content; they’re training signals. They reveal how people think, what they struggle with, what they’re curious about, and how they phrase questions. Maybe it is anonymized (a good thing) but it is still valuable and often sensitive data.

Alarmist articles will remind you that many free AI chatbots use your prompts, your corrections, and your uploaded files. They have that photo of your family that you let them enhance. What will they do with what you give them? I can't answer that as of now, and certainly not for the future. I know that your conversation history is used to train or fine-tune future versions of the model. Hey, you are part of the product pipeline - but don't expect to be paid for your contributions.

I also concede that the business model matters and that different AI companies monetize differently. For example, Microsoft provides its own privacy commitments and policies, and those govern how your data is handled. For details, they always direct users to their Privacy Statement.

Here are 4 business models currently out there:
Ad-supported = Your attention is monetized.
Freemium = Free tier gathers usage; paid tier subsidizes development.
Enterprise licensing = Your data may be isolated; the company earns from businesses.
Open source =  The model is free; the company may sell hosting or support.

So "if an app is free, you are the product" still applies, but not always in the same way. When an AI tool is free, you’re not just the product — you’re also the collaborator. You’re an unpaid teacher, tester, and a source of fuel for improvement.

Remember when Google was just a search company? Now, Google has stopped being just a search company.

Here are 13 AI tools.

• Google Classroom AI Tutor is a personalized learning assistant giving you adaptive education at scale
• Gemini Live is Voice + screen sharing, AI conversations, and real-time problem solving
• NotebookLM takes any source and gives instant summaries to cut your research time.
• Veo 3 takes text and creates high-quality video without any editing experience required
• Gems in Gemini will build task-specific AI assistants to automate your workflow
• Google AI Studio can test and compare AI models side by side
• Google App Builder goes from prompt to working app with zero coding skills needed
• Firebase Studio is also for AI-powered app development
• Imagen turns text into production visuals
• Gemini Ask on YouTube can extract answers from any video
• Gemini in Google Sheets will auto-generate formulas and insights faster than your spreadsheet speed
• Google Cloud Vision API offers OCR + image detection
• Nano Banana is used for AI image editing with variations, allowing you to refine visuals instantly

 

 

Manus is a newer AI from Meta. Before the acquisition of Manus, Meta’s primary AI system was Meta AI, powered by the Llama (Large Language Model Meta AI) series of models. While Meta AI was a robust "chatbot," Manus represents a shift toward "agentic AI."

Why did they make the move? Meta AI was integrated across WhatsApp, Instagram, and Messenger. It was designed to do several things: answer questions, provide information, generate images, compose text, and summarize long conversations or documents. Pretty much what every other AI chatbot was doing.

Meta AI was primarily a "conversation layer." If you asked it to "book a flight" or "build a website," it could give you advice or write code, but it couldn't actually go into a browser, navigate a website, or complete the transaction for you.

In late 2025, Meta acquired the Singapore-based startup Manus for a reported $2 billion to solve the "execution gap" between talking and doing.

So while you would ask a chatbot to "Write a travel itinerary for a road trip from San Diego to San Francisco," you could ask an agentic AI to also "Book the hotels for this itinerary."

Manus can plan and execute multi-step tasks. It can open a virtual browser, research a topic, create a spreadsheet, and then email that spreadsheet to a colleague without human intervention. Competitors like OpenAI (with Operator) and Google (with Gemini Agents) were moving toward AI that can control a computer. Meta needed Manus's "execution layer" to stay competitive.

Where's the money? Meta is integrating Manus into its Ads Manager and business tools. This allows businesses to automate complex marketing workflows—like building entire landing pages and running ad reports—simply by asking.

More history: 
September 2023: Meta AI first debuted publicly, initially on devices like smart glasses.
April 2024: Wider rollout across Meta’s social apps.
April 29, 2025: Standalone Meta AI app released.

March 6, 2025: The autonomous AI agent Manus was officially released to the public. It gained attention as an early example of an autonomous AI agent that could operate without continuous human guidance, and in late 2025, it was acquired by Meta.

Manus is an example of what’s often called an “agentic” AI system. Rather than simply responding to prompts, it is built to take a high-level goal and carry out the steps needed to achieve it. That might include researching information, planning a workflow, writing and executing code, analyzing data, and producing a finished output. In other words, Manus is structured to complete multi-step tasks with a high degree of autonomy. It functions more like a digital project manager or operator than a chatbot.