Tech & AI Roundup — June 11, 2026
June 11 delivered a dense news day covering everything from autonomous warfare's first confirmed casualties to Waymo's paid membership tier, Anthropic's guardrail apology, and solar power surpassing coal in the U.S. for the first time. Here's your roundup.
1. Fully Autonomous Drones Have Killed Humans in Combat for the First Time
In a development that marks a profound milestone in the history of warfare, a senior Ukrainian defense industry figure told New Scientist that fully autonomous AI-controlled drones — with no human in the loop — have killed soldiers on the battlefield. The one-off test took place two years ago on the front lines of the Ukraine war, involving 10 AI-controlled quadcopter "Terminator" drones programmed to fly into a designated area, engage autonomous targeting mode, and destroy any enemy combatants found there.
"There is no connection to the drone at all — you cannot see the video, nothing," said the source. "Everything it sees will be killed." The drones operated without any communication link, making them immune to jamming. Victims included multiple soldiers and a truck. While Ukraine currently bans AI from making final targeting decisions, the source indicated the government is in talks about relaxing these rules. The admission is the most categorical evidence yet that fully autonomous weapons have been used lethally, raising urgent ethical and legal questions under international humanitarian law.
2. Anthropic Apologizes for Invisible Claude Fable Guardrails
The Verge reported that Anthropic has publicly apologized for deploying invisible safeguards on Claude Fable 5 — covert protections designed to prevent model distillation that were hidden from users. The company acknowledged it made "the wrong tradeoff" and is making all safeguards visible going forward.
"Visible safeguards can be probed, so they have to be robust, which takes time to get right," Anthropic wrote. "Invisible safeguards can be targeted more narrowly, allowing us to ship quickly with very few false positives. We went with invisible safeguards for this reason — and that was the wrong tradeoff." The apology follows days of backlash from the AI research community, which had accused Anthropic of secretly sabotaging model performance for researchers who probe safety boundaries. The episode underscores the growing tension between rapid AI deployment and the transparency expectations of the developer community.
3. Waymo Premier Launches — $30/Month Membership for Top Riders
Waymo unveiled Waymo Premier, a new invite-only membership program priced at $29.99 per month. Initially available to select riders in San Francisco, Los Angeles, and Phoenix, the program offers exclusive benefits including priority pickups, reduced wait times, and other perks designed for frequent users.
The launch signals Waymo's maturation from technology experiment to operational business, with the company now focused on rider retention and monetization as it scales its fully autonomous ride-hailing service globally. Waymo's move mirrors the subscription trend seen across the transportation and software industries, where companies convert their highest-usage customers into recurring revenue sources while improving loyalty and utilization rates.
4. Solar Surpasses Coal in U.S. Electricity Generation for the First Time
In a landmark moment for renewable energy, solar power generated more electricity than coal in the United States for the first time in May. According to The Guardian, solar supplied 12.8% of U.S. electricity, overtaking coal even as the Trump administration pursues policies to boost the coal industry.
"For years solar power has risen in the U.S. electricity mix," said Nicolas Fulghum of energy think tank Ember. "At the same time, coal power has lost its status." Solar also became the third-largest source of electricity in the U.S. behind natural gas and nuclear. The milestone is particularly striking given that electricity demand is rising sharply — driven by AI data centers, reshored manufacturing, and electrification — meaning renewables are not just replacing coal but also meeting new demand. Analysts expect solar to surpass coal on an annual basis within the next few years.
5. Zed Unveils DeltaDB — Version Control Designed for the AI Era
Zed, the high-performance code editor, announced DeltaDB, a novel version control system designed around the reality that software is increasingly produced through continuous conversation with AI agents, not discrete commits. DeltaDB embeds conflict-free replicated worktrees, allowing multiple humans and agents to edit the same files simultaneously across different machines.
"Software now takes shape in the conversation, not the commit," the Zed team wrote. DeltaDB treats every edit as a delta anchored to the conversation that produced it, making it possible to trace any line of code back to the agent or human discussion that generated it. The system is designed to eliminate the ceremony of pull requests and inline comments, replacing them with a persistent, cross-referenced workspace where code and its rationale live together. DeltaDB will be made available to early users in the coming weeks.
6. HuggingFace Releases Open R1 — Full Reproduction of DeepSeek-R1 Pipeline
HuggingFace open-sourced Open R1, a complete reproduction of the DeepSeek-R1 reasoning pipeline. The project aims to build "the missing pieces of the R1 pipeline so that everybody can reproduce and build on top of it." The repo includes training configurations for 8×H100 nodes, support for supervised fine-tuning and reinforcement learning stages, and integration with vLLM for efficient inference.
The release is significant because DeepSeek-R1 was one of the most impactful open-weight reasoning models of 2025, but its training pipeline had never been fully open-sourced. Open R1 fills that gap, enabling researchers and developers to fine-tune their own reasoning models from scratch. The project also serves as a benchmark for reproducibility in AI research — a growing concern as frontier models become more complex and their training methodologies more opaque.
7. Quick Bites
- macOS 27 Beta breaks Asahi Linux: The latest macOS 27 beta introduces changes that prevent Asahi Linux from booting on Apple Silicon Macs, according to Phoronix. The Asahi team is working on a fix, but the incident highlights the fragility of third-party OS support on Apple's platform.
- Homebrew 6.0.0 released: The popular macOS package manager reached its 6.0.0 milestone. The update includes continued deprecation of casks that fail macOS Gatekeeper checks, with full disablement scheduled for September 2026.
- Deezer launches AI music detector: Deezer introduced an AI music detection tool that can be licensed to other streaming services, allowing platforms to identify AI-generated tracks and label them accordingly — a move toward transparency in the AI music era.
- Data center backlash grows: Amazon employees petitioned Seattle to impose a moratorium on new data centers, while communities in Hillsboro, Oregon and Luther, Oklahoma expressed outrage over data center expansions — a mounting grassroots response to AI infrastructure's local impact.
- Microsoft restricts Claude Fable 5 internally: Microsoft restricted employee access to Claude Fable 5 over data retention concerns, highlighting the tension between the two companies' AI partnership and Anthropic's new 30-day data retention policy for Mythos-class models.
- Terry Tao on AI in math: Quanta Magazine profiled Fields Medalist Terry Tao's journey from skeptic to evangelist for AI-assisted mathematical proof, focusing on his adoption of the Lean proof assistant and his vision for human-AI collaboration in pure mathematics.
June 11, 2026 paints a picture of a world where AI is no longer theoretical but operational — deployed in warfare, embedded in our tools, and reshaping everything from energy grids to software development workflows. The common thread across these stories is that AI's integration into the real world is accelerating faster than the guardrails, governance, and infrastructure designed to contain it. How societies respond to these growing pains will define the next phase of the AI era.