Attacks target OpenAI CEO home & AI weapons race and drones - Tech News (Apr 13, 2026)

We’ll start with that security story in San Francisco. Police say OpenAI CEO Sam Altman’s Russian Hill home was targeted again early Sunday—just two days after a Molotov cocktail attack at the same property. Investigators allege a car stopped outside, and a passenger fired a round toward the home. Two suspects were arrested, and police say multiple firearms were recovered. No injuries were reported in either incident, but the back-to-back nature is what stands out—AI leadership is becoming high-profile enough to attract not just criticism, but real-world threats.

Now to national security, where the headline isn’t one new gadget—it’s the pace of the arms race. A display of autonomous drones at a Beijing military parade, attended by Xi Jinping and watched alongside Vladimir Putin and Kim Jong-un, triggered fresh concern in Washington that the U.S. is behind China in unmanned combat capabilities. U.S. officials say the Pentagon believes China—and possibly Russia—has an edge in advanced drone production and autonomy. That’s pushing the U.S. to speed up domestic manufacturing, with firms like Anduril ramping AI-enabled drone production in Ohio earlier than planned. Why it matters: as autonomy increases, decisions can happen faster than humans can comfortably supervise. Analysts warn that letting AI steer battlefield choices could make conflicts more unpredictable—and more likely to escalate—especially when everyone’s capabilities are partly hidden behind secrecy.

Staying with AI, but shifting to the business world: at the HumanX conference in San Francisco, the chatter reportedly moved away from OpenAI as the default center of gravity. Anthropic took the spotlight, with attendees describing a wave of demand for Claude-based coding agents in enterprise settings. The interesting wrinkle is that this excitement comes even as Anthropic’s relationship with the Pentagon has been strained by disputes and litigation—yet the company still appears to be gaining traction across other parts of the federal ecosystem and the private sector. The broader signal: companies don’t just want a single “best model.” They’re increasingly building diversified AI stacks—multiple providers, multiple model types—partly for resilience, and partly because geopolitics is now a procurement concern, not an abstract talking point.

On the consumer AI front, Meta is making two moves that fit together. First, it released a new model called Muse Spark, framing it as a step back into the frontier race after a rocky stretch. Early chatter suggests strong performance in text and vision use cases, with weaker results in coding. But Meta’s advantage has never been only model rankings—it’s distribution. If Meta can offer a noticeably better free experience inside apps people already use every day, it can win attention at massive scale. Second, Meta is reportedly developing its own AI-powered search engine to power real-time answers across Facebook, Instagram, WhatsApp, and Messenger. Today, Meta AI leans on Google and Bing for live information. Building its own search layer would reduce dependence on competitors—and it also changes the bargaining power between platforms and publishers. Meta has reportedly also partnered with Reuters for news content, a sign that “who supplies the facts” is becoming a key battleground in AI answers.

The infrastructure behind all of this is getting its own headline treatment—because the money is moving fast. CoreWeave, the Nvidia-backed GPU cloud provider, landed a massive long-term compute deal with Meta covering years into the future, and then piled on a cascade of financing moves: convertible notes, bonds, and major loan facilities. The key point isn’t the paperwork—it’s what it reveals. The AI buildout is being powered by aggressive leverage, justified by long contracts from credible customers. The risk is straightforward: if demand slows, or borrowing becomes more expensive, these capital-heavy expansion plans can get uncomfortable quickly. The opportunity is just as clear: whoever builds capacity first can become the default supplier when everyone else is still waiting in line.

Zooming out to national strategies, Japan is reportedly coordinating a major domestic AI push. SoftBank Corp., NEC, Sony Group, and Honda have helped create a new company aimed at building high-performance, Japanese-made AI and making it broadly available to local industry. Government support could be substantial, and engineers from key players are expected to participate. This is part of a wider pattern: countries don’t want to be permanently dependent on U.S. or Chinese platforms for foundational technology. “Sovereign AI” is turning into a practical industrial policy—especially for economies that worry about supply chain shocks, export controls, or sudden shifts in access.

Now a quick temperature check on how AI is landing in everyday work. Gallup says half of employed U.S. adults now report using AI at work at least a few times a year, with daily use also climbing. More employers are integrating AI into workflows, but the human experience is mixed: people in AI-adopting organizations report more disruption—and they’re more likely to observe both hiring and layoffs. Notably, worry about job elimination within five years is rising, especially inside the companies adopting AI fastest. The takeaway: productivity bumps are real, but for many workers they’re still task-level improvements, not a full reinvention of how work happens. The big question for leaders is whether they redesign workflows thoughtfully—or just bolt tools onto old processes and hope for the best.

Developers are also learning that the AI era changes what “good documentation” looks like. A growing idea called Agentic Engine Optimization argues that coding agents read docs differently than humans do: fewer requests, less patience for heavy pages, and a greater chance of missing context if content is long or locked behind access controls. The practical outcome is that teams are starting to publish cleaner, machine-parseable versions of docs and add agent-oriented indexes—because if an agent can’t reliably find the right instructions, it may improvise, and that’s when bugs and hallucinations show up in production.

And in open source, one of the world’s most important software projects is drawing a line around responsibility. The Linux kernel community clarified that AI-generated code is allowed, but it’s treated as the contributor’s own work. Humans must ensure licensing compatibility, follow contribution rules, and personally sign off. AI tools can’t certify anything. That’s significant because it welcomes AI as a helper without weakening accountability. If something breaks, the chain of responsibility stays intact—and that’s essential for infrastructure software used by millions of systems worldwide.

Switching gears to energy and geopolitics: the war involving Iran is disrupting global energy markets by curbing traffic through the Strait of Hormuz, a key route for oil and gas shipments. Beyond higher prices, the interesting second-order effect is how this reinforces the case for electrification and renewables—especially in import-dependent regions. Analysts point out that China could benefit disproportionately because it dominates supply chains for EVs, batteries, and solar panels. In other words, an oil shock can accelerate the clean-energy pivot—and shift economic leverage toward whoever can ship the alternatives fastest.

Finally, space policy. NASA’s Space Launch System—fresh off sending the Artemis II crew around the Moon—is facing renewed uncertainty as the Trump administration again presses NASA to find commercial replacements. The White House budget emphasizes shifting away from the expensive SLS-Orion approach, and NASA leadership is reportedly pausing or canceling parts of the long-term plan, including elements tied to future upgrades and the Gateway station. The stakes here are political and strategic: SLS supports jobs across the country, but the cost and pace are constant targets. Meanwhile, the U.S. wants a credible path to return humans to the lunar surface on a timeline that competes with China’s ambitions.