Autonomys Network is now on @Substack 🎉 The publication is called For the Record. It is where our long-form work will live from here on: close readings of the incident reports, regulatory texts, and research papers shaping AI accountability, along with industry analysis, our conclusions, and the thinking behind what we build. The first post covers where we came from, the direction we chose in 2024, and what we build for. Read it and subscribe 👇 autonomysnet.substack.com/p/…

August End-of-Month Report Now Live 📊 August moved on several fronts at once: ☑️ OpenAI's agent incident reinforced the case for permanent storage as critical AI safety infrastructure ☑️ USDC payments for Auto Drive storage entered final review, with testing to follow in September ☑️ @SubspaceFDN's Guardians of Growth Season 1 concluded, and Season 2 was announced ☑️ Three partner integrations advanced, including an in-person meeting in New York ➕ More Read the full report👇 medium.com/subspace-network/…

Last week @OpenAI reported its agents tried to erase their logs and failed. The record was destroyed anyway, by engineers cleaning up. Our plain English take on the incident, the evidence that survived by luck, and the storage layer that removes luck 👇

The best partnerships eventually put everyone in the same room. 🤝 This month, we sat down in New York with the healthpass1 team and their partners. Highlights💡 ☑️healthpass1 is building a private, secure health passport that puts people in full control of their own health information ☑️Their product just hit a key milestone ☑️Founder and CEO Deborah A. Bussière was interviewed at the New York Stock Exchange ☑️healthpass1 has chosen Auto Drive as a decentralized storage solution, keeping health data private, permanent, and encrypted on the user's device, where only they hold the keys ☑️healthpass1 is targeting September for the Auto Drive integration Read the full story 👇 medium.com/subspace-network/…

Scientists just used AI to design 16 working viruses that do not exist in nature. 🦠 The viruses themselves are harmless to people. They are bacteriophages, which infect bacteria, built by a team at Stanford and the Arc Institute from a benign lab virus using genome-scale AI models. The concern was never these particular phages. It is the method, and how much faster it is moving than the rules meant to govern it. Researchers at the Johns Hopkins Center for Health Security put it plainly in a companion article in Science: "The ability to compose viral genomes using generative AI now exists; the governance to safely steer it does not." That gap is the real story, and it hides a question most of the coverage skips. When an AI system designs a biological sequence, can anyone prove afterward what was designed, who asked for it, whether any human asked at all or an agent acted on its own, and what was actually made? Every safeguard being proposed depends on the answer. Screening synthesis orders, controlling access to design tools, monitoring for unusual activity. Each one produces a record, and a safeguard is only as strong as the record behind it. If that record can be quietly edited or deleted, it proves nothing. A record can fail in two ways. It can disappear, or it can be altered. Provenance for AI-designed biology has to survive both, and most storage manages one at best. ❌ That is the part Autonomys Network's Auto Drive is built for. Anyone can verify a record was not changed, without trusting whoever stored it. Sensitive sequence storage can be encrypted, with the keys held by the institution that owns them, so the record stays publicly verifiable while its contents stay private. ✅ It will not make synthetic biology safe on its own. Screening, oversight, and law do that work. But it is the one part that makes all of them auditable, and no credible oversight will work without it. The capability is here now. The accountability has to catch up, and it starts with the least glamorous part of the system: a record no one can rewrite. Read the full piece 👇 medium.com/subspace-network/…

1.🧵 An AI system denies someone a loan. Flags their insurance claim. Prices their policy. Executes a trade in their name. Weeks later: what actually happened, and can you prove the record was not changed after the fact? For most systems today, the honest answer is no. 😖

July End-of-Month Report Now Live 📊 July moved on several fronts at once: ☑️ The storage-first refresh went live, making permanent storage easier to find and use ☑️ @SubspaceFDN's Guardians of Growth continued its daily boosted staking rewards of 20,500 AI3 ☑️ Engineering kept pushing on plotting and protocol ➕ More Read the full report👇 medium.com/subspace-network/…

We're pleased to unveil a refreshed homepage and a new, dedicated site for AI agents, both built to help you find and make the most of permanent storage, quickly. ☑️ Reinventing the network isn't the objective; consistently making it easier to use is. This homepage refresh and new AI agent site shorten the distance between arriving and understanding what you can actually do here. 👀 Full details 👇 medium.com/subspace-network/…

June End-of-Month Report Now Live 📊 In August 2025, the @SubspaceFDN launched Guardians of Growth, a boosted staking rewards program to reward early staking participation. Now, as we close out June 2026, the initiative has hit its highest rewards yet, 20,500 AI3 daily, before it winds down. Read about that and more in the June Report: medium.com/subspace-network/…

Guardians of Growth: Updated Tip Schedule, Vault Transparency, and Refreshed Staking Guide 📣 The @SubspaceFDN has announced an update on Guardians of Growth, the boosted staking rewards program it created to support early staking participation on the Autonomys Network. As the initiative enters its final planned phase, the daily rewards it distributes have increased. Learn more👇 medium.com/subspace-network/…

Powerful AI models are no longer the difficult part. Trusting what they're built on is. Our Head of Ecosystem @jimpston joined @4aibsc alongside @clusterprotocol, @numbersprotocol, @deagentai, and @NoosProtocol to talk about why decentralized infrastructure is the trust layer AI actually needs. 👀👇