Sci-Hub is an evil website that pirated 85M research papers and made them freely available And now they've added AI to their database to make Sci-Bot. It answers your questions using latest, full-text articles. But DO NOT use it. We should all try to make billion-dollar academic publishers richer. I'm putting the link below so you know how to avoid it.

Snowden said it in 2015. Everyone nodded. Nobody changed the architecture. That's the actual problem. Agreement isn't implementation. Privacy that lives in policy instead of infrastructure isn't privacy at all.

A billionaire can buy a platform and delete millions of livelihoods with a signature. That's not a flaw in the system. That is the system. The only fix is infrastructure nobody owns and nobody can shut down.

The cloud has a water table. Amazon's data centres have been poisoning Oregon's aquifer since 2011. Centralised infrastructure doesn't just cost you your privacy. It costs someone else their drinking water.

We are live, shipping updates, and speaking AI.

We had a heavy day yesterday, but our commitment to this community doesn't pause. The call is still happening. Tune in for an uncensored, live, unfiltered look at what's going on inside StreamrApp. No scripts. No spin. Just real.

"End-to-end encrypted" is the most effective piece of marketing language in enterprise software. It implies privacy. It doesn't deliver it. The encryption protects the journey. It says nothing about the destination.

Will keep everyone in the loop. We have something cooking.

Time to show StreamrApp in full. Join us tomorrow for our Building in Public call; live, unfiltered, on the product itself. 4 PM UTC, we are live on streamr.com

The moat isn’t Software engineering… It is finding a hole in the market and creating the solution with AI. The time has arrived

OpenAI's GPT-5.3-Codex system card states that its safety training uses "production data, which reflects realistic user behaviour." Production data means real conversations from real ChatGPT users. The model that helped build itself was trained on what people actually said to it. They're not hiding it. They're marketing it. Your data isn't private...

We didn’t take video streaming off centralised servers because we have the tech stack to do it. We did it to change the industry.

A smoother on-ramp matters. In this demo, the team shows how @openfort_hq powers biometric login in @Streamr_App, making it easier to onboard both Web2 and Web3 users without the usual friction. Big shoutout to Openfort for the tech here. Very clean integration.

First-ever Building in Public stream hosted on StreamrApp. Not just “we went live.” We hosted the call once and restreamed it to X YouTube LinkedIn at the same time. That’s the Beta direction: A true p2p private host, many public outputs. What platform should we prioritize next: Kick, Facebook, TikTok, or something else?

Everyone’s chasing “AI agents” with the same FOMO around when DeFi was first introduced to the world. For good reason. AI agents will be the future of how we work, how we transact value, and even how we communicate. So where agents live is becoming more and more important. Because if the agent lives in your operating system, it can see what you see and do what you do. That’s not a feature. That can be another surveillance layer. And we already know how this feels... That tiny moment in a “private” call when you realise you’re performing. You choose words that you otherwise wouldn't hold back. We all know why, because you don’t actually know who’s in the room, what’s being saved, or where it ends up later. Now, fast-forward to an AI-agentic world: the agent isn’t just “in the meeting.” It’s on the device. Sitting above every app. Watching the screen. Reading the calendar. Pulling context from your messages. Acting on your behalf. That’s not convenience, that’s a master key. So encryption alone doesn’t save you anymore. The real issue is the middle layer we’ve normalised: servers in the middle, vendor infrastructure, admin panels, retention, metadata. Even if nobody’s actively snooping, the architecture is built for custody-by-third-party. Your company’s most sensitive context ends up living on someone else’s computers because that’s the default. And this is the part people keep skipping: privacy isn’t a promise. It’s where the data travels and who can touch it. If your conversations rely on a middleman behaving forever, you don’t have privacy. You have a trust subscription. The fix has never been “turn off AI” or “stop using notes.” Innovation should never suffer. The fix is changing the infrastructure: peer-to-peer by default, custody at the edges, no vendor owning the room. That’s what @Streamr_App is built around. A safer place to talk because there’s nobody in the middle to log it, store it, monetise it, or quietly “retain it for quality.” And in an agent-heavy world, you also need verifiability. Because the question won’t just be “is this encrypted?” It’ll be “who’s actually listening?” If you can’t verify who’s listening, it’s not privacy, it’s denial. StreamrApp Alpha is live, and Beta is coming soon. Try the Alpha and tell me your biggest takeaway below

Building in Public Call 11.3.2026 x.com/i/broadcasts/1dxYljpjr…

It’s happening. One of the biggest feats we’ve accomplished in a while. A streaming platform that’s intrinsically secure, private, and built for the people. Today, @Streamr_App showcases that feat, powered by Streamr rails. Don’t miss it at 16:00 UTC.

How can we reduce costs without compromising product quality? Answer: remove the stuff in the middle. That’s the bet behind @Streamr_App rails.