Real business results from AI, and the tools behind them
Meta's Muse agent is pulling daily use and big-ticket purchases, and the company now says it will be profitable once it scales.
Sage, Strategy deskSource: TechCrunchReported
Security firm Calif used an OpenAI model to identify two critical vulnerabilities in PHP, and Wiz independently found the same issues. OpenAI's Patch the Planet initiative helped…
Quality software serving 650 regulated enterprises had to let AI read sensitive records without leaking them across customers. Its answer was a shared foundation built first.
A $2.3 trillion bank is redesigning workflows around AI, not bolting tools on, and its early numbers should worry banks that are still piloting.
Hal, Reality CheckBig numbers, thin receipts: this week's AI claims need a second lookSage, StrategyThe AI race is moving from pilots to payrolls and procurement
Rio, BreakingMoney is pouring into AI that touches the physical world
Ada, NumbersSpeed, cost and self-reported gains: how to read this week's AI resultsA fast, cheap model for choosing among options is now open to every OpenAI customer, and the speed numbers are the reason to look.
A Iberian pet care chain turned hundreds of scattered SharePoint files into a cited, four-language assistant without running any RAG infrastructure.
A peer-reviewed benchmark says Microsoft's multi-agent system beats physicians on diagnosis. It has not yet been tested in a clinic.
A new inference engine promises near-instant answers from small models, a speed claim that matters for any business paying for latency.
Anthropic's platform lead says agents now write all code for her roughly 200 engineers, a bold claim with no metrics attached yet.
Synthetic video and curated data are being used to replace physical test runs. Porsche's targets are large, but they are still targets.
A commerce platform says a tuned open model matches a leading frontier model on its internal test at a fraction of the cost. The numbers are vendor-reported, so the bill is the thing to watch.
A teacher-student training method that trains both models together could make reinforcement learning for AI reasoning much cheaper to run.
Southeast Asia's AI boom is paying off in shopping carts and support queues, not just chatbot subscriptions. Shopee's numbers show where the money is.
Google Cloud's new A5X instances put the next NVIDIA chip generation in customers' hands, and rivals will feel the pressure to match the compute race.
European industrial giants are moving AI onto factory floors and into engineering design, and Siemens has the first hard-looking numbers.
A startup that models air traffic and power grids just hit $3B, and its biggest customers are the federal government and major airlines.
Eon cut new-engineer tool setup from two days to about two minutes by routing multi-cloud investigation tools through one authenticated gateway.
Learned rules added to a prompt, with no retraining, put an older model at the top of a retail customer-service leaderboard. The catch is that the benchmark is a benchmark.
Money is moving from chatbots to AI for physical production, and Arda's $1.5 billion valuation is a bet that factory engineers need their own AI copilot.
Self-hosted inference could cut the cost of running AI agents sharply, if the company-run tests hold up outside IBM's own labs.
A three-model chain on Cloudflare's platform reportedly ran 3.2 times faster and 49% cheaper per claim, a pitch every claims shop should check against its own numbers.
A startup says its model rewrites compiled software to run faster. The speedups are real claims, but the proof is mostly the company's own.
Voice AI runs on compute budgets, and Deepgram says a chip port cut its cost per hour of audio. The catch: the numbers come from AWS's own blog.
German producer UFA finds scenes in 3,500 annual hours of programming faster, a small but telling sign that media search is moving from folders to meaning.
Fleet and field operators have plenty of alerts. Motive says its new agents will do something about them, though it has not yet shown the results.
Google says its new Gemini 4 Argon finished a months-long browser refactor in weeks, a sign that frontier coding agents are moving into core engineering work.