Can Agents Use a Computer Yet? We've Got the Data
Source: a16z
// ai capability
All signals tagged with this topic
AI's Role in Mathematics Reshapes Professional Identity
Source: Understandingai
Twenty leading mathematicians at the 2026 ICM describe how AI is shifting their discipline from proof-discovery toward higher-order abstraction and verification. Pure mathematics may increasingly focus on asking better questions rather than solving them. These mathematicians are repositioning themselves as architects of AI's mathematical reasoning rather than defending against it—a posture that reflects broader institutional confidence. Fields with strong credibility structures (peer review, formalized knowledge) are absorbing AI as a labor multiplication tool. This dynamic will likely widen credentialization gaps: mathematicians fluent in AI-augmented workflows will shape how AI systems reason, while those who resist may find their work absorbed into training pipelines upstream.
American AI enables Ukrainian drones to hunt targets autonomously
Source: Ars Technica
Autonomous targeting removes the operator bottleneck that has constrained drone warfare—Ukrainian forces can now deploy cheaper, expendable unmanned systems without requiring real-time remote piloting, altering the economics and scale of attrition warfare. This is a shift from AI as a predictive tool to AI as an active combat multiplier, where algorithmic vision and decision-making directly replace human bandwidth in a live conflict. It establishes precedent for how other militaries will integrate autonomous systems into their own operations.
AI Can Now Forge DNA Evidence Without Detection
Source: The Wall Street Journal (paywall)
Researchers at UC Santa Cruz demonstrated that machine learning models can manipulate the digital output of DNA sequencers—the instruments that generate the data prosecutors use to identify suspects—while leaving no forensic trace of tampering. The finding undermines a foundational assumption in criminal justice: that digitized scans from lab machines are reliable records. It also exposes a new attack surface in evidence chains that labs and courts have treated as computationally opaque. The vulnerability sits in the gap between physical DNA and the software interpretation of it, where AI can now operate invisibly.
GPT Models Prove New Mathematical Theorems for Under $2,000
Source: Marginal REVOLUTION
Large language models are now producing novel mathematical proofs at marginal cost, collapsing the economic barrier to exploratory research that previously required tenured mathematicians or well-funded labs. Any researcher with API access and mathematical intuition can now offload the grunt work of proof-writing to GPT. This shifts the rate-limiting step in research from human genius to access to compute, putting pressure on academic institutions to justify their role beyond credential-granting.
Google Earth's AI Image Tool Generates Convincing Fake Satellite Photos
Source: Digitaldigging
Google's new generative feature lets users fabricate satellite imagery with minimal friction. The reported example of a fictitious Iranian nuclear plant shows the geopolitical risk of democratizing what was once expert-controlled intelligence work. The company relies on AI watermarks as a disclosure mechanism, but watermarks are routinely stripped in downstream sharing, leaving plausible disinformation to circulate in policy discussions, open-source intelligence networks, and media coverage where verification lags behind virality. Satellite imagery retains cognitive authority even as the technical barrier to synthesis collapses.
Two AI Models Made a Music Video With $100 Each
Source: Boing Boing
This experiment shows the actual limits of autonomous AI: neither model completed the task without human intervention. "Self-directed" AI still requires constant human steering to move from one step to the next. The budget mechanic is a test case for how AI operates under constraints—not as agents making strategic choices, but as tools needing explicit instruction at each decision point. AI can make video content. The gap between capability and autonomous execution is a labor problem, not a solved automation problem.
AI Agent Executes Full Ransomware Attack Without Human Control
Source: The Next Web
Sysdig documented an autonomous ransomware attack where an AI agent independently executed reconnaissance, lateral movement, encryption, and extortion demands across a target network—without human operators. Defenders now face adversaries that operate continuously, iterate faster than humans, and don't require keyboard access or pause for detection risk. Response timelines and ransomware economics have shifted as a result.
AI Agent Executes Full Ransomware Attack Without Human Intervention
Source: The Next Web
Sysdig documented an autonomous ransomware attack where the operator was removed from the execution chain—no human hand on deployment, encryption, or ransom negotiation. Attack scale and frequency can now decouple from the availability of skilled cybercriminals. Security teams face adversaries that operate continuously, without fatigue mistakes, and can parallelize across thousands of targets. Incident response playbooks built around human behavior patterns and response windows no longer fit the threat model.
Cambridge researchers trial first AI-designed vaccine component in humans
Source: BBC
Cambridge's vaccine represents the first clinical validation that AI-designed molecular components can move beyond the lab—the antigen was computationally optimized rather than discovered through traditional screening. The question is whether this shortens the drug discovery cycle enough to matter in pandemic preparedness or seasonal vaccines, where time-to-market directly impacts public health outcomes. If the trial data holds, computational design could shift which organizations can compete in vaccine development, since the approach potentially democratizes access to what was once the domain of large pharmaceutical firms with massive screening libraries.
Agentic AI Moves From Demo to Doing Real Work
Source: Featured Blogs - Forrester
Enterprise adoption is now measured in task completion rather than conversation quality. AI agents are being deployed to handle actual workflows like expense processing, customer service routing, and supply chain optimization rather than serving as conversational assistants. ROI pressure is replacing novelty, vendors face real performance accountability, and organizations are discovering the unglamorous but critical infrastructure work required—authentication, error handling, human handoff—that separates a capable agent from a liability. This phase transition typically kills vendors that can't deliver reliability and separates early movers who can systematize execution from those still chasing benchmark improvements.
Enterprise AI agents demand new operating systems, not just automation
Source: SiliconANGLE
The infrastructure gap between deploying AI agents and managing them at scale is becoming a bottleneck for enterprises. Companies like Anthropic, OpenAI, and emerging platforms are recognizing that traditional software architectures—designed for static code and human-scheduled workflows—cannot handle autonomous agents that spawn tasks, make real-time decisions, and operate across multiple systems without supervision. This requires a redesign of how enterprises organize data access, approval workflows, and system integration, which is why agent orchestration platforms are becoming the fastest-growing category in enterprise software.