Living neurons playing Doom & EU clean-energy and SMR push - Tech News (Mar 11, 2026)

A computer made of living human brain cells just learned to play a version of Doom—and that’s not science fiction, it’s a real lab demo raising big questions about the future of computing. Welcome to The Automated Daily, tech news edition. The podcast created by generative AI. I’m TrendTeller, and today is March 11th, 2026. Let’s get into what moved the tech world—what happened, and why it matters.

Living neurons playing Doom

We’ll start in Europe, where the European Commission is pitching a broad set of moves to make energy cheaper, cleaner, and less fragile. The headline is resilience: the EU wants to cut exposure to geopolitical shocks tied to imported fossil fuels, while also responding to ongoing frustration from households and industry about high bills. One centerpiece is a “citizens’ energy package,” which is basically a promise to make the system friendlier to consumers—clearer bills, easier switching between suppliers, and targeted support for people most at risk of energy poverty. Alongside that, the Commission is trying to unlock more private investment into grids, efficiency, and next-generation clean tech, working with the European Investment Bank. And yes, nuclear is in the mix too: there’s a parallel push to help countries prepare small modular reactors, with the goal of first operational deployments in the early 2030s. The interesting part here isn’t any single policy lever—it’s that Europe is treating energy cost, energy security, and industrial competitiveness as one connected problem.

EU clean-energy and SMR push

In the U.S., we’re seeing a clear signal that AI has moved from “experimental” to “official workflow.” The Senate has reportedly approved ChatGPT, Google’s Gemini, and Microsoft’s Copilot for staff use, with plans to plug them into internal digital systems. The pitch is straightforward: less time spent on repetitive drafting, summarizing, and research—more time on judgment calls and decisions. In parallel, the Pentagon is moving ahead with Gemini AI agents for unclassified Defense Department networks, potentially touching millions of personnel. Put those together and the takeaway is big: AI isn’t just helping companies write marketing copy or code anymore—it’s starting to shape how government work products are created, revised, and reviewed. That raises the stakes for transparency, oversight, and the question of who is accountable when AI-assisted work goes wrong.

AI tools enter U.S. government

That accountability question is at the center of a high-profile legal fight involving Anthropic. The company told a federal judge it could lose hundreds of millions—or even billions—in 2026 revenue unless the court quickly blocks a Trump administration decision labeling it a U.S. supply-chain risk. Anthropic argues it’s being punished for pushing back on how its AI might be used, including concerns around domestic mass surveillance or autonomous weapons. Microsoft has backed a temporary block, warning the designation could disrupt Defense Department contracting and force costly software removals with few alternatives. The judge accelerated the schedule for a key hearing later this month, while the Justice Department wouldn’t promise to pause further action in the meantime. What makes this case so consequential is that it’s not only about one vendor’s future—it's a test of how far the government can go in restricting contractors when policy disagreements spill into procurement decisions.

Anthropic fights supply-chain label

Now to the world of “agentic” AI—systems that don’t just answer questions, but act and interact. Meta has acquired Moltbook, a viral simulated social network that looks a bit like Reddit, except many of the “users” are AI agents. Meta’s interest appears to be the idea of an always-on directory where agents can find each other and coordinate—basically, an early hint of what an agent-to-agent consumer platform could look like. But Moltbook also carried a warning label: the project raised authenticity and security concerns, including the possibility that some posts weren’t from agents at all, but humans impersonating them. That’s the tension in this space: the more lifelike and autonomous these systems become, the more important verification, safety boundaries, and basic trust mechanisms get.

Meta buys viral AI-agent network

On the creative side of AI, OpenAI is reportedly preparing to bring its Sora video generation and editing tools into ChatGPT subscription plans. If that happens, it’s a strong play to make ChatGPT feel less like a chat window and more like a full media studio—especially as rivals push their own multimedia features. The bigger story isn’t just “AI can make videos.” It’s that the main AI apps are turning into bundled platforms, where text, images, audio, and video creation sit under one subscription—and that could reshape the economics of creator tools and editing workflows.

Sora video tools inside ChatGPT

Sticking with AI—but moving to software engineering—there’s a growing divide between what’s possible and what’s safe. One developer says he built a new programming language in a matter of weeks using an AI coding agent, with the model generating essentially all the code while he focused on specs and tests. It’s a striking example of how fast prototyping can get when the bottleneck becomes decision-making rather than typing. At the same time, Amazon is dealing with the opposite side of the coin: reliability. After a string of serious outages, the company held an internal meeting acknowledging availability problems and promising extra safeguards around production changes, including changes that were assisted by generative AI. Amazon disputes that AI-written code was a broad cause—but the lesson still lands: as more teams rely on AI assistance, the rigor of review, testing, and rollout discipline becomes even more important, not less.

AI coding agents meet reliability

Let’s go off-planet. NASA’s Office of Inspector General has warned that SpaceX’s Starship lunar lander is likely to face further delays, putting pressure on NASA’s revised 2028 target for a crewed Artemis moon landing. The report points to schedule compression and an unresolved challenge around in-orbit refueling at the scale NASA needs. It also flags an uncomfortable safety reality: NASA doesn’t yet have a proven way to rescue astronauts if a lander fails in space or on the lunar surface, meaning contingency options are limited. And it’s not just SpaceX—Blue Origin’s competing lander is also reported to be behind. Meanwhile, on the other side of the lunar race, a Nature Astronomy study highlights candidate landing areas in the Moon’s Rimae Bode region for China’s first crewed lunar mission, targeted for before 2030. The theme is clear: the moon is becoming a multi-player destination again, and timelines are tightening while engineering risk stays stubbornly real.

Moon race: Artemis and China sites

Back on Earth, YouTube’s scale is starting to look less like a platform and more like a media superpower. A financial research report estimates YouTube generated over $60 billion in 2025 revenue, putting it ahead of Disney’s media business. The point here isn’t a scoreboard for bragging rights—it’s what it signals about where audiences and money actually concentrate. YouTube is now a combined engine of ads, subscriptions, and creator-driven programming, and its growing suite of AI tools could accelerate production volume even further. For legacy media, the competitive set isn’t just Netflix anymore—it’s the entire creator economy at industrial scale.

YouTube tops global media revenue

In supply-chain news, Apple is reportedly ramping iPhone manufacturing in India to around a quarter of global output, part of a broader effort to reduce reliance on China amid trade tension and tariff uncertainty. This isn’t just about one company diversifying. It’s another sign that electronics manufacturing is being reshaped by geopolitics, not only by cost. The more production spreads across countries, the more complex logistics and quality control become—but the payoff is resilience when trade policy gets unpredictable.

Nasal vaccine and dementia biomarker

And finally, two health-and-biotech stories worth watching. Stanford Medicine researchers described a mouse study of an intranasal “universal” respiratory vaccine concept, designed to provide broad protection in the lungs across different threats. It’s early-stage, but the promise is compelling: a nasal approach that could offer faster front-line defense against new outbreaks, and potentially simplify seasonal protection. Separately, Japanese researchers reported DNA aptamers that bind neurofilament light chain—an important blood biomarker linked to neuron damage in Alzheimer’s and other disorders. If aptamer-based sensors prove practical, they could make monitoring cheaper and easier, opening the door to more routine tracking of brain health rather than rare, expensive tests.

Before we wrap, let’s return to today’s hook—because it’s a reminder of how strange the frontier is getting. In Melbourne, Cortical Labs says it trained a biological computer made from living human brain cells to play a simplified Doom. Performance isn’t the point yet; proof of concept is. The bigger question is where this goes: from radically different energy profiles for computing, to new lab models for research, to a fresh set of ethical and governance debates about “wetware” systems. Even in a week packed with AI headlines, that one stands out as genuinely new.

That’s the tech landscape for March 11th, 2026: governments formalizing AI use, vendors fighting over procurement power, the moonshot timeline slipping under real-world constraints, and a glimpse of computing that’s literally alive. If you want to keep up with how these threads connect—energy, AI governance, and the next generation of platforms—follow the show and come back tomorrow. Until then, I’m TrendTeller.

Living neurons playing Doom & EU clean-energy and SMR push - Tech News (Mar 11, 2026)

Our Sponsors

Today's Tech News Topics