Mark

If you’re someone who worries what the digital future looks like for our grandchildren, this documentary might bring you some hope. There is a better way. The foundation of a better internet.

Fact Friday – @nillion This week’s spotlight is on Nillion, a privacy-focused network building what it calls the “Blind Computer”, a system for private compute, private storage, and private AI, designed to let developers use sensitive data without exposing it. Here is what you need to know: Phase 2 Upgrade Is Live: Nillion rolled out a unified developer portal, bringing nilDB, nilCC, and nilAI into one interface. It also shifted to a credits-based access model, introducing Stripe payments for nilAI usage, making it easier for developers to build and deploy in production. Blacklight Verification Layer Is Live: In February, Nillion launched Blacklight, its verification layer on Ethereum L2. Community-run Blacklight nodes verify workloads across the network, adding a public validation layer to Nillion’s private compute stack. This bridges private computation with transparent verification, a key step toward trustless privacy infrastructure. Anyone can stake 70,000 NIL and run a Blacklight node to earn rewards. Nillion later expanded access with managed nodes, lowering the barrier to participation. ERC-8004 Agent Validation: Nillion also introduced what it describes as the first verification layer for ERC-8004, allowing Blacklight to validate whether registered agents are live and responding onchain. Rather than treating agent registration as a one-time status, Blacklight adds continuous verification over time, improving accountability and transparency for onchain agent systems. Nillion Migrated to Ethereum: Nillion moved NIL from its Cosmos-based nilChain toward Ethereum as part of its L2 architecture shift. This move brings staking, verification, and node participation closer to Ethereum’s liquidity and application ecosystem. At the center of this transition is NIL. Previously more tightly scoped within the Blind Compute, it is now positioned as a core coordination asset across the network, underpinning verification, incentivizing node participation, and supporting emerging validation layers like Blacklight.

nilAI now supports Gemma 4 26B. That means better speed and stronger output for teams building in the Nillion ecosystem. Live on nilGPT 👇

@JohnAlanWoods GPM = no chance!

Collective intelligence, reimagines. Join us as we welcome @metapoll_xyz to the ecosystem and learn how groups can move beyond simple yes/no voting to make better decisions. 🗓️ April 23, 11 AM EST Link below!

And so many regular people are now entering their confidential medical information, including DNA and lab reports, into public AI tools and irrevocably hurting their future employability and insurability. Use Monadic DNA Explorer instead. We use Nillion's confidential LLMs and storage so nobody snoops on your data!

A new privacy-first AI project joins the Nillion ecosystem. @AIagenten is building SikkerKI, a fully private AI chatbot, powered by Nillion. 🧵 👇

@nillion @milesdeutscher @cz_binance The future is blind ⚔️

@milesdeutscher @cz_binance This is exactly why privacy has to become normal. Nillion is building toward that with private storage, private compute, private AI, and verifiable privacy instead of “trust us.”

The CEO of Binance ( @cz_binance ) just said something most people in crypto don't want to hear. The industry is "too transparent," and it's actually a huge problem. Institutions/whales don't want transparency - they want to benefit from blockchain's efficiency, whilst having the same privacy-preserving elements that traditional rails offer. Whoever properly solves this will be a big winner.

At-home sequencing is the future! However, until it becomes cheap, easy and reliable, people will still turn to services which compromise their anonymity and privacy. Our upcoming Batcher app provides an anonymous mixing service using reliable lab partners for your DNA exploration needs!

Every trader on CT has seen this play out; Someone posts a mind-blowing PnL, people copy them, then the calls go cold, the account goes quiet, and everyone who followed is left holding bags. The problem was never the market. It was that there was no way to know if that trader was ever actually good. Read the thread. 🧵

@weezerOSINT The most alarming part is not just that data got exposed. It’s how normal it still is for that data to just sit out there. That’s why the next wave of products needs a different default: private storage, private compute, and systems that can prove they’re still private.

@nillion @buildonnillion The future is blind

The Vercel breach put the problem in plain sight. Too many systems still keep sensitive logic and data stored in one place. nilDB exists to break that pattern. - Sensitive data is split before it is stored - It is distributed across multiple nilDB nodes - No single node holds the complete secret - No single breach reveals the whole dataset That’s why Nillion matters: it gives builders a privacy layer for sensitive data.

Small game tomorrow. COYG!

The complete tech stack for blind compute that web3 never knew it needed… until now. Institutions won’t adopt web3 without this, and Nillion has already proven this with the recent integration of Vodafone. The future is blind.

With Phase 1, we brought together private storage, compute, and AI into a single stack. Since launch, the Blind Computer has scaled to 112K+ users, 641M+ documents, and 1.4M inferences powering real applications. Phase 2 takes that further.

Next, we move from proving usage to running inference, with towers operating collectively as secure and privacy-preserving edge collective compute. Learn more: nillion.com/news/transform…

How Nillion × @web3pairpoint x @chainlink are turning telecom towers into secure AI data centres 👇 Imagine AI is a brain that lives far away. When you use most AI tools today, your request travels to a huge data centre, the brain “thinks,” then the answer comes back. Tower AI compute changes that. It puts smaller AI “brains” on cell towers, close to where people and devices actually are. Now the hard part: You have to pay for the tower hardware upfront, even when usage is still small. So we built the missing piece, a way to count real usage & tie money to it. - Pairpoint (by @vodafonegroup) measures tower compute usage and signs it at the source - Nillion verifies and aggregates it privately - Chainlink publishes the verified usage onchain, so repayments can scale with what’s actually happening So instead of guessing demand, this lets funding follow real usage and repayments scale as usage grows.

A unified developer portal is now live. Developers can access nilDB, nilCC, and nilAI from one place, manage usage, API keys, and workflows without switching between systems.