AI is getting smarter every day. But the more it learns about us, the bigger the question becomes: who actually controls what it remembers?
Think about how we use AI today.
We ask it to help with work, explain complicated topics, plan our schedules, and sometimes even discuss personal matters.
Over time, AI can start learning our preferences, habits, and how we like things done.
That sounds useful, right?
But here's something worth thinking about.
Just because AI remembers something about you doesn't mean that information is true. And just because it knows something doesn't mean it should be allowed to keep it forever.
That's why the recent collaboration between Verona (@verona_dev) and ZetaChain (@ZetaChain) caught my attention.
They're looking at two important pieces of the AI puzzle: private memory and verified information.
ZetaChain: Giving Users Control Over AI Memory
Imagine spending months teaching an AI assistant how you work.
It remembers your writing style, favorite tools, ongoing projects, and personal preferences.
Then one day, you decide to switch to a better AI model.
Wouldn't it be frustrating to start teaching everything from scratch?
This is one of the problems ZetaChain is trying to solve.
Through its Private Memory Layer for AI, ZetaChain is working toward AI experiences where users control their information and can carry their memory across different AI models.
Its consumer application, Anuma, already demonstrates this approach.
According to ZetaChain, more than 320,000 Anuma accounts have been created since February 2026, with over 1.3 million AI requests across 38 models.
The interesting part isn't just the numbers.
It's the idea that your AI memory should belong to you, rather than being locked inside a single AI platform.
Verona: Remembering Something Isn't the Same as Proving It
Now, this is where Verona enters the picture.
Let's say you tell an AI assistant that you're over 18 years old.
The AI might remember that statement.
But remembering your claim doesn't automatically prove that it's true.
Verona approaches this problem differently.
As an intelligence layer for AI, Verona focuses on turning information verified at its original source into reusable proof that users control.
Instead of revealing sensitive personal information every time an application needs verification, authorized AI agents can use that proof without accessing the underlying raw data.
For example, imagine an application needs to confirm that you're old enough to access a particular service.
The goal is to prove that you meet the age requirement without unnecessarily exposing your full date of birth or identity documents.
That's the difference between information an AI remembers and information it can actually verify.
Memory makes AI more personal. Proof makes the information it uses more trustworthy.
And I think that distinction is going to matter a lot as AI agents become more involved in our daily lives.
Why This Collaboration Matters
What makes Verona × ZetaChain interesting is how their approaches could complement each other.
ZetaChain focuses on private, user-controlled AI memory.
Verona focuses on verified information that can be reused without exposing the sensitive data behind it.
Now imagine an AI assistant that remembers your preferences while also being able to verify specific facts when necessary.
It could potentially offer a more personalized experience without requiring users to hand over unnecessary private information.
That's a much more interesting direction for AI than simply building models that remember more and more data.
To be clear, the teams are currently exploring how these capabilities could work together. This isn't an announcement that a fully integrated product is already live.
But the direction itself is worth watching.
The Most Interesting Part? They're Asking Users First
Instead of deciding everything behind closed doors, Verona and ZetaChain have launched a joint research initiative to understand what people actually want from AI memory.
They're asking questions such as:
• What information should AI remember about you?
• What information should it never keep?
• Should your AI memory follow you when switching models?
• Who should decide what gets stored or deleted?
• Would you allow AI to verify certain facts without sharing the underlying data?
• What personal information would you allow an AI to use when connecting you with other people?
These aren't just technical questions.
They're questions that could shape how people interact with AI in the future.
The research includes the Anuma community, Verona's community, and anyone who uses AI.
Both teams plan to publish the findings in November 2026.
My Take: The Future of AI Isn't Just About Intelligence
For a long time, much of the conversation around AI has focused on which model is faster, smarter, or more powerful.
But I think we're entering a stage where another question matters just as much.
Can we actually trust what AI knows about us, and can we control how it uses that information?
An AI assistant that remembers everything might sound impressive.
But an AI assistant that respects our privacy, gives us control, and can verify important information without exposing unnecessary personal details?
That's a future I'd be much more interested in.
And this is why I'm paying attention to what Verona and ZetaChain are exploring.
Because the next big improvement in AI might not just be a smarter model.
It might be a better relationship between AI and the people using it.
Your Opinion Matters Too
Verona and ZetaChain are inviting AI users to participate in their research.
The survey is anonymous, requires no account, takes around five minutes, and closes on October 27, 2026.
If you use AI, consider sharing your perspective.
👉 Official Survey:
https://docs.google.com/forms/d/e/1FAIpQLSdjzQjUk_q0wfL5fEQ4xq1Z86N_rEfDD9zL2o8oGMZe8BApyg/viewform
One final question:
If AI could remember everything about you, what's the ONE thing you would never want it to keep?
I'd love to hear your thoughts.
Official Sources:
Verona announcement: https://x.com/verona_dev/status/2110008524596298093
ZetaChain article: https://www.zetachain.com/blog/zetachain-x-verona-asking-300000-users-what-ai-should-remember