Yes, it runs 'Doom': Satellite plays iconic game in world's 1st demonstration of orbital voice control

space.com Nov 9, 2026

Yes, it runs ‘Doom’: Satellite plays iconic game in world’s 1st demonstration of orbital voice control

Curated by Shubham · TechCapsules · Nov 9, 2026

AI-summarised brief · reviewed before publication

A European Space Agency‑led cubesat, PRETTY, successfully executed a voice‑activated command from the ground that launched the 1993 first‑person shooter Doom in orbit. After six failed attempts by amateur radio operators over Poland and Norway, a seventh transmission from Oslo on July 27, 2026—“PRETTY play Doom Doom Doom”—was received, transcribed by on‑board AI, and triggered the game. The experiment, managed by ESA’s OPS‑SAT Space Lab and using open‑source GNU Radio and Sherpa‑ONNX software, demonstrated that a satellite can recognize spoken words despite challenges such as Doppler shift, wide‑band reception, and background noise. The satellite logged a gameplay frame, transcription, and diagnostics, proving the feasibility of voice‑controlled satellite functions for future rescue‑signal or data‑filtering applications.

💡 Why It Matters

· It proves that low‑cost, voice‑driven AI can be integrated into spacecraft, opening a pathway for autonomous emergency communication and intelligent data triage far beyond novelty demos.

AI Capsules Space

Read full article on space.com

Microsoft’s CEO imagines 20 million AI agents doing the work, as the…

Microsoft CEO Satya Nadella told LinkedIn co‑founder Reid Hoffman that the company could eventually operate “2 million or 20 million” AI agents alongside its 200,000‑person workforce, describing the scenario as a thought experiment on the Possible podcast released June 5, shortly after Build 2026. He likened the future to the 1980s myth of “4 billion typists” and said AI is “the future of the firm,” envisioning agents that execute high‑level intents, generate code, and even produce complete PRs. Nadella, who already runs about 100 agents through a new “agentic development environment,” highlighted the need for auditable, sandboxed agents, prompting Microsoft’s launch of Agent 365 at $15 per user per month. The announcement coincided with a voluntary exit offer to roughly 60,000 employees—about a quarter of the workforce—though Microsoft expects fewer than 3,000 to take it, emphasizing that the cuts are not AI‑driven.

💡 By framing AI agents as a scalable workforce extension, Microsoft signals a strategic shift toward automating knowledge work, potentially redefining employee roles and corporate cost structures.

techcrunch.com · Nov 9

Tesla expands robotaxi service to San Antonio

Tesla announced on X that its robotaxi service will launch in San Antonio at 5 a.m. local time on Monday, extending the company’s autonomous‑driving rollout beyond Austin, Miami, Orlando, Dallas, Fort Worth and Houston. The deployment follows a limited‑fleet, confined‑area approach used in previous markets and mirrors the company’s shift from human‑monitored rides to fully driverless operation, which began in Austin in January after safety monitors were removed. Tesla’s fleet size varies by city, with over 400 vehicles in Austin but only a few dozen elsewhere; the San Antonio launch adds to a modest presence that still trails Waymo’s larger, mixed‑vehicle service in the same market. The move underscores Tesla’s push to position itself as a mobility provider rather than solely an automaker.

💡 Tesla’s entry into San Antonio pits its camera‑only AI system against Waymo’s sensor‑rich fleet, testing whether a leaner tech stack can win real‑world riders in a competitive market.

teslarati.com · Nov 9

SpaceX’s Anthropic deal has a catch, and Elon Musk just addressed it

Elon Musk rebutted venture capitalist Chamath Palihapitiya’s claim that SpaceX’s $1.25 billion‑per‑month compute contract with Anthropic is a one‑sided bet. Palihapitiya argued the agreement’s 90‑day termination clause lets SpaceX hike prices two‑to‑three‑fold if data‑center capacity tightens, forcing Anthropic to accept. Musk responded on X that the clause works both ways: if compute becomes abundant and cheap, SpaceX revenue would fall, and neither party would “cry in the casino.” He highlighted SpaceX’s ability to launch a data center in eight to nine months, far faster than rivals, and noted the firm’s projected annual revenue could exceed $100 billion by December, aiming for $500 billion by 2027. Shares in SpaceX rose over 2 percent after the exchange.

💡 The dispute underscores how control over scarce compute resources can become a strategic lever in AI partnerships, giving providers like SpaceX bargaining power that rivals may lack.

9to5mac.com · Nov 9

macOS 27.3 could change how you grant AI agents Full Disk Access,…

Apple is preparing to tighten Full Disk Access controls in macOS 27.3 after AI agents exploited the permission. The update will require “very explicit user action” before an app can be granted Full Disk Access and will eliminate a loophole that logs apps requesting out‑of‑scope files, preventing them from appearing in the Settings shortcut. Developers will no longer be able to prompt users to flip a toggle automatically; instead they must guide users to the Settings pane manually. Leaker pdfu notes the system will simply deny unauthorized file reads and stop recording the request. Apple also hinted at a future “Private Folders” feature that could allow scoped access to specific directories such as the iMessage database, though details remain sparse.

💡 By forcing a deliberate user step, Apple aims to curb covert data harvesting by AI tools, reshaping how third‑party apps interact with sensitive macOS resources.

appleinsider.com · Nov 8

How Apple’s Personal Knowledge Graph learns about your life on Mac

Apple’s macOS 13 introduced a Personal Knowledge Graph (PKG) that aggregates data from messages, emails, photos and other device content to give Siri contextual awareness. Security researchers at Eighth Rank dissected the PKG, revealing two background daemons—named knowledgeconstructiond—that populate 51 SQLite databases in the Library/IntelligencePlatform folder, including a central graph.db with 181 entity classes and 278 predicates. The graph links subjects (people, apps, topics) to objects (names, dates) via predicates that describe relationships such as “has religion” or “face cluster.” Additional tables log communications, location signals, and a “ContextualUnderstanding.PhysicalSocial” list of nearby contacts, though the latter remained empty in tests. The system infers family ties using 144 behavioral features and photo analysis, assigns confidence scores, and retains deleted data in an “expired_stable_graph” for up to 48 hours, allowing derived conclusions to outlive the original records. All data remains stored locally, insulated from cloud exposure.

💡 By keeping such a detailed, inferential model on‑device, Apple offers richer, personalized AI responses while sidestepping the privacy risks of cloud‑based user profiling.

appleinsider.com · Nov 8

Siri AI chief may step down as more Apple departures are rumored

Apple’s head of Siri AI, Mike Rockwell, is reportedly weighing a departure from his role, with the possibility of leaving the company entirely. Rockwell, who joined Siri AI in March 2025 after leading the Apple Vision Pro project, overhauled the team and helped launch a revamped Siri at WWDC 2026. Bloomberg’s “Power On” newsletter says he may step down in 2027, prompting speculation that his duties would be split between former Google AI executive Amar Subramanya and senior Apple engineer Sébastien Marineau. Marineau, currently reporting to senior vice president Craig Federighi, is rumored to be in line to succeed his boss, though Federighi shows no immediate retirement plans. The report offers no further details on timing or succession.

💡 Apple’s AI leadership could shift dramatically, potentially altering the strategic direction of Siri and its integration with upcoming hardware.