AI, machine learning and LLM insights.
Security researcher Frank Chu found that the AI-note service tl;dv left its "meetings" collection unprotected in Firebase, allowing any logged-in user to download every transcript – a flaw that stayed unpatched for six months.
The audit of 2,465 public AI-agent skills revealed 57.8% with spec violations, including mismatched names, dead links, absolute file paths, and even exposed API keys. Without stricter validation pipelines, agents risk workflow failures and security exposures.
The GA release, announced without a changelog, slashes reasoning-token usage by up to 62%, turning into immediate cost savings for metered users, but it also drops the model’s built-in refusal behavior and requires the “thinking” flag to be off for correct JSON output.
Relay will cut off free users on Aug 15 and paying customers on Sep 14, ending its AI-automation service. Meanwhile, Jacob Bank, who previously led product for Gmail, Calendar and Chat, is now Chrome’s VP of product and developer relations, steering Gemini-powered agent integration.
The Series B will finance Wispr’s new meeting-note-taker, backed by its proprietary Canto model targeting sub-10% error rates, and a partnership with the Oasis wearable ring to bring voice-first automation into noisy, privacy-sensitive workplaces.
The watchdog changes come after a Washington Post investigation uncovered 46 instances of officers using Flock’s 120,000-camera network for personal stalkings and other unauthorized searches, prompting the company to force a case-number field that currently isn’t validated.
Apple plans to pay news outlets a usage-based fee—potentially hundreds of millions of dollars—to let Siri pull verified, up-to-the-minute headlines into answers, shifting from flat-fee licensing to a performance-linked model.
Anthropic’s probabilistic watermark embeds a subtle pattern in every token, letting regulators spot AI-generated text but also risking “sticky” marks that linger in contracts and could expose lawyers to liability.
The new model will power the upcoming “Apple Intelligence” features in the next iOS update, after Apple officially registered its on-device generative AI service with Chinese regulators last month, marking a historic compliance milestone for a U.S. tech giant.
The multiyear deal lets Abbott’s Lingo sensor stream glucose readings every few minutes straight into the Gemini-powered coach inside Google Health, where AI translates spikes and dips into actionable lifestyle tips in real time.
Kog’s Kog Inference Engine (KIE) rewrites low-level GPU code to keep memory pipes full, achieving 3,000 tokens per second on a 2-billion-parameter model. Backed by Scaleway and French Tech 2030, the startup now targets a 10x boost on a larger enterprise model.
By publishing Glimmer’s model weights, Meta lets engineers run AI on smartphones, IoT devices, or on-prem servers, sidestepping cloud latency and privacy concerns, while its Muse Spark API remains the premium, closed-door offering.
The new "Media watermark" toggle lets users turn off the lower-right sparkle on Gemini-generated images and videos, but still embeds SynthID patterns and C2PA metadata so the content can be verified by Google tools and downstream platforms.
Anthropic’s new SynthID-Text embeds a covert binary pattern in Claude’s token choices, satisfying the EU AI Act’s Transparency Code without altering readability. The watermark survives casual edits, adds less than a tenth of a percent latency, and is disabled for private deployments.
The Epoch AI-Ipsos poll of 1,106 employed adults also shows that data analysts (46%) and document readers (39%) are rapidly following, while full-task automation stays under 10% across all roles.
Computer History records window focus, button presses and typed text locally, then feeds only the relevant event slice to ChatGPT. The opt-in system lets users blacklist apps, prune entries, and toggle logging, promising richer assistance without visual capture.
The move follows a wave of exits among OpenAI’s safety and ethics leaders and signals a strategic pivot toward scaling consumer-ready AI products ahead of a massive IPO.
An analysis of over 1,200 campaign sites shows AI mentioned in nearly two-fifths of races, outpacing foreign policy and crypto. Candidates now must explain how power-hungry data centers will affect electricity rates, water supplies and local land use.
Okta’s new identity-scoped Model Context Protocol delivers only the tool schemas an agent is authorized to use, dramatically lowering token consumption and inference costs while enforcing least-privilege access for enterprise AI workflows.
A February 2026 Pew survey of 10-to-18-year-olds shows they view AI as a functional aid—57% for research, 54% for school tasks—while fearing it erodes creativity, fuels environmental waste, and spreads misinformation.
DeepSeek replaced the preview build behind its unchanged `deepseek-v4-pro` endpoint with the officially released V4 Pro 0813 model, altering weights, instruction handling, and JSON tool-call formatting—potentially breaking production pipelines that rely on stable output.
The benchmark on a 32 GB M2 Pro MacBook measured token-per-second rates, JSON success, and total call time, showing Muse Glimmer’s 33-second per-call latency versus Llama 3.2’s sub-second speed, even though accuracy was identical.
The rebrand follows Cognition's $2.4 billion licensing deal with Google and a scrapped $3 billion acquisition, shifting focus from a proprietary model to an IDE that can orchestrate Claude Code, OpenAI Codex, Google Gemini and any ACP-compatible agent.
The 160-intersection rollout, brokered by traffic-control integrator Econolite, pushes Ouster’s Utah presence to nearly 300 sites and underscores a broader industry shift toward AI-embedded public-infrastructure projects.
