Untangling Web3

🎙 Untangling Web3 The podcast that simplifies tomorrow's technology, today. Follow and subscribe to join our journey to untangle Web3.

Planning a tech event? Untangling can help with: • live podcast recordings • panel hosting + moderation • guest speaking • content capture for clips + social We’ve got hands-on experience turning complex tech topics into engaging live conversations around AI, blockchain, digital identity, IoT, and more. Need support for your event? Reach out now: theuntanglingweb3podcast@gmail.com

FBI just seized websites tied to Handala, the pro-Iranian hacktivist group that hit medical device giant Stryker last week (per TechCrunch). That’s not “retail breach” territory. Stryker builds surgical robots, implants, OR equipment. If a vendor like that gets disrupted, the blast radius shows up in hospitals: inventory gaps, delayed procedures, teams flying blind. Also: a takedown within days is fast in government time. Either they already had solid intel on the group’s infrastructure, or they moved hard before it got rotated and reused. And the motive angle matters less than people think. “Ideological” vs “financial” changes negotiation dynamics, sure. It doesn’t change operational damage. If your incident response plan assumes your systems fail, not your vendor’s, is it actually a plan? What’s the one dependency you’d be most nervous to lose for 72 hours?

Could SpaceX's future be in orbiting data centers, not just space travel? Alec and Jack on Untangling Web3 explore how cheaper space flight might make near-Earth compute more viable than terrestrial land. It’s a fascinating convergence of two huge industries. 🤔

Stryker (yes, the medical device giant) says pro‑Iran hackers wiped thousands of employee devices. Not ransomware. No “pay us.” Just destruction. That’s a different threat model: - Wipers are about operational chaos, not ROI. - Backups help, but they don’t magically restore laptops, endpoints, identity access, and day‑to‑day workflows at scale. - When your tech touches hospitals, “disruption” isn’t an IT inconvenience. It ripples outward. TechCrunch frames it as a response to US military action. If that’s accurate, corporate networks are becoming pressure points in geopolitics. If you’re running security in 2026, what do you change first: endpoint hardening, segmentation, recovery drills, or exec-level risk assumptions? 🔒

TechCrunch reports Stryker (yes, the hip replacement giant) just had thousands of employee devices wiped by pro-Iranian hackers. Not encrypted for ransom. WIPED. That’s a different class of pain: - No negotiation, no leverage, no “pay and pray.” - Recovery is slower, messier, and the point is the message. Researchers reportedly believe this may be the first major cyberattack on US soil tied directly to retaliation for the Trump administration’s military action in Iran. If that’s right, private companies are now collateral in geopolitical disputes, even when their “business” is medical devices, not defense. So what counts as “critical infrastructure” when the blast radius includes whoever is connected and convenient? Which industries are actually ready for destructive attacks, and which ones are just running ransomware playbooks?

Elon Musk's vision for space is shifting from Mars to the Moon, aiming for a self-growing lunar city in 10 years. Beyond commercial travel, the goal is humanity's long-term survival—a backup plan if Earth becomes uninhabitable.

Reliz (parent of institutional crypto lender BlockFills) just filed Chapter 11. The filing lists ~$50M–$100M in assets vs ~$100M–$500M in liabilities. In a bull market. That’s not a “2022 hangover” you can wave away. What sticks out: BlockFills wasn’t a retail casino. It marketed itself as B2B plumbing for institutional trading desks, the kind of lender that’s supposed to be boring. So if an “infrastructure” lender can still end up with liabilities potentially 5x assets after sentiment has recovered… what does that say about the core liquidity assumptions behind crypto lending? How many similar balance sheets are out there that just haven’t gone public yet 👀?

Wired reports someone accessed FBI files related to Epstein. Not via a fancy exploit. Via a misconfigured access control. That’s the part that should make everyone uncomfortable. Most security “threat models” obsess over elite attackers. Meanwhile the real risk is boring: a permission set that was wrong, stayed wrong, and nobody owned it. Centralizing sensitive data is easy. Building the operational discipline to continuously audit who can access what is the hard part. If something this high-profile can be exposed by configuration drift, what’s sitting open in systems nobody is actively watching?

TechCrunch says law enforcement just shut down SocksEscort, a proxy service built on tens of thousands of hacked home routers. Not a “hacker crew”. Infrastructure for rent. That’s the part people miss about modern cybercrime: - Someone compromises boring consumer gear at scale. - Someone else sells “clean” residential IPs. - Other groups (ransomware, DDoS, fraud) plug in and go. Taking down a service like this hits a supply chain, not a single attacker. But the bigger problem hasn’t moved: unpatched routers sitting in spare rooms and small offices are still free real estate. So who actually owns the fix here: router manufacturers, ISPs, or users? 🤔

AI's energy hunger is growing fast—1-2% of global electricity already! ⚡️ Could space-based data centers be the answer to scaling issues and sovereignty concerns? Alec and Jack dive into the wild possibilities on Untangling Web3.

Sora inside ChatGPT might be happening soon (reportedly). That’s a big strategic tell. ChatGPT is already at ~920M weekly active users, and OpenAI has talked about a 1B target. Closing that gap with “one more feature” is hard… unless the feature becomes a habit. What’s interesting: they’re not treating Sora like a standalone product with its own growth curve. They’re folding video gen into the platform as a retention + acquisition lever. That’s the classic platform play: more capabilities → more default-ness → fewer reasons to leave. But does AI video actually have daily pull? Or is it a novelty spike, then back to “cool demo”? If Sora lands in ChatGPT, what happens next: new daily workflow, or short-term hype?

A federal judge just blocked an AI browser from placing Amazon orders on a user’s behalf (The Verge). The interesting part isn’t “AI shopping.” It’s authorization. Even if you hand an agent your Amazon login, Amazon can still say: “we never authorized *that* kind of access.” And the court (so far) seems to agree there’s strong evidence of unauthorized access. Two separate consent relationships are colliding: 1) You → agent (“act for me”) 2) Platform → agent (“you’re allowed here”) Agentic AI lives or dies on moving across systems that were built for humans and policed by ToS. If platforms can block at the authorization layer, the “AI does it all” future hits a legal wall, not a technical one. Who should decide what your AI agent can do: you, the platform, or the courts? 🤔

Why is launching big rockets from Earth so difficult? The vast majority of weight is fuel! 🤯 But what if we used the Moon as a staging ground? Turns out, water ice on the moon can be split into rocket fuel, making deep space missions far more feasible.

UK data center co Nscale just raised $2B at a $14.6B valuation…and put Sheryl Sandberg + Nick Clegg on its board. That’s the part I can’t stop thinking about. GPU-dense AI infra is the obvious story. But Sandberg/Clegg aren’t “more compute” hires. They’re “more legitimacy” hires. If you’re trying to win sovereign + enterprise AI contracts in markets that don’t want American-owned compute, the differentiator isn’t rack density. It’s governance, political access, and trust at scale. Are we about to see “board composition” become a competitive moat for AI infrastructure outside the US?

The US just issued its first nuclear construction permit in nearly a decade. It went to TerraPower’s Natrium, a reactor design that isn’t operating anywhere in the world yet. 👀 Two details worth sitting with: 1) It’s sodium-cooled (not water), which means high heat transfer without pressurization. That’s a different safety and engineering trade space than the reactors most people argue about. 2) The part most headlines skip: it’s built with molten salt energy storage, so it can “bank” heat and ramp power when the grid needs it. That’s nuclear designed for a world of spiky demand, not just steady baseload. Important: NRC approval ≠ a finished plant. It just means they can break ground in Wyoming. Timelines, costs, and supply chain risks are still very real. If AI-driven load keeps rising faster than grid buildout, is the real bottleneck for advanced nuclear the tech, the regulators, or the economics?

🚨 Want to join one of the hottest tech podcasts right now? If you, or someone you represent, has valuable insight on technology, the future, or the projects shaping what’s next, we’d love to hear from you. Untangling Web3 speaks to founders, builders, and researchers on emerging tech. Email: theuntanglingweb3podcast@gmail.com

Apple reportedly pulled ByteDance’s China-market apps from the US App Store back in January. Not TikTok. The obscure stuff most Americans have never heard of. That’s the interesting part: geopolitics is getting enforced at the distribution layer, not the network layer. No splashy takedowns. No big court moment. Just a quiet “region availability” change and entire categories of software stop existing for US users. Meanwhile the TikTok deal is still being negotiated. The broader catalog moved faster than the diplomacy. If app stores are the chokepoint, “trade policy” can look a lot like a settings update. How much of today’s tech policy is actually platform policy with a flag on it?

Why put data centers in space? Alec and Jack on Untangling Web3 explore the incredible benefits: abundant, direct solar energy and passive radiative cooling in the vacuum of space. It's about using unused energy and solving heat issues efficiently.

Kraken reportedly got a Federal Reserve master account. That’s not a “partnership.” It’s direct access to the Fed’s payment rails, the plumbing big banks use. No need to route dollars through a sponsor bank as an intermediary. Faster settlement, lower cost, less counterparty risk. For years, even well-capitalized crypto firms have been forced into dependency on partner banks. 2023 showed how fragile that can be when banking relationships vanish overnight. So what is this, exactly? A real policy shift at the Fed… or a one-off exception? And the bigger issue: when the Fed grants (or withholds) a master account, it’s picking who gets to scale on core financial infrastructure. What should the criteria be for a Fed master account, and who should decide them?

What if data centers weren't on Earth? SpaceX is exploring orbital data centers powered by solar energy for AI. We this fascinating development on Untangling Web3, hinting at a future where abundant energy fuels AI's exponential growth. Is this the next frontier? 🤔