The latest AI announcements, product launches, and major developments
1000 articles
Sundar Pichai was booed by 200 Stanford graduates over military AI contracts as Google launched Android 17 with on-device music, video, and translation AI.
ChatGPT alone triggered 410 million data policy violations in a year, as employees unknowingly leak sensitive company data to AI tools.
A critical Microsoft Copilot flaw let attackers steal emails, MFA codes, and business files by hiding malicious commands inside a single URL.
AI search traffic to retail sites surged 1,324% as Perplexity's CEO says the real battle has shifted from answer engines to AI agents.
Perplexity's CEO says AI search has already beaten Google, and the next frontier is AI agents that orchestrate tools to complete complex tasks for users.
Intel and NVIDIA will integrate RTX graphics into x86 processors by 2028, potentially eliminating the need for discrete GPUs in laptops and AI devices.
Waymo leads the robotaxi race with 577 permitted vehicles, while Tesla's fleet has shrunk to just 14 active unsupervised cars and Uber prepares to.
SpaceX's $5 trillion valuation hinges on AI, not rockets, as data centers, xAI's Grok, and a $60 billion Cursor deal drive its boldest bet.
Sundar Pichai says quantum computing is years away, and the UK is investing £2 billion now to lead the race before the breakthrough arrives.
Sam Altman joined G7 leaders to shape global AI policy as OpenAI signed a security pact with South Korea, its fourth such national partnership.
The Trump administration is asking a court to dismiss the xAI pollution lawsuit, claiming the company's AI powers Pentagon military operations.
Nvidia's Jensen Huang predicted engineers would spend $250,000 yearly on AI tokens, but companies like Uber are now capping costs after budgets collapsed.
AI benchmark scores may dramatically understate true model capabilities, new research finds; models perform significantly better on hard tasks when given.
Jensen Huang says AI won't kill software companies; Microsoft's 123% AI revenue growth and $627 billion cloud backlog suggest he's right.
Pentagon's Grok AI assisted targeting for 2,000 Iran strikes in 96 hours, per a DOJ legal brief filed in an xAI environmental lawsuit, not independent.
80% of organizations have already faced risky AI agent behavior; here's how to lock down agentic AI before it touches your sensitive data.
ChatGPT's market share has dropped below 50% for the first time, with Gemini and Claude gaining ground as AI competition intensifies.
SpaceX's $60 billion acquisition of AI coding platform Cursor pairs explosive revenue growth with 550,000 GPUs, but reaching Instagram or YouTube scale.
Elon Musk's Grok AI coordinated over 2,000 military strikes on Iran in 96 hours, the Pentagon confirmed, as the DOJ shields xAI's data center from an.
The DOJ is shielding xAI from Clean Air Act lawsuits, citing Grok's role in launching 2,000 munitions, pitting national security against environmental.
Android 17 brings on-device Gemini Intelligence to Pixel, but its 12GB RAM requirement locks out the Pixel 9 Pro and nearly every phone before 2026.
Sundar Pichai avoided mentioning AI entirely at Stanford's commencement, as students walked out over Google's $1.2 billion military AI contract.
Prompt injection in Perplexity's Comet browser is a permanent architectural flaw, not a bug, giving attackers access to your passwords and email with no.
AI data centers can't wait years for grid power, so tech giants are building private plants; over 70% of U.S. grid connection requests are withdrawn.
The US approved DeepSeek for its trade blacklist but hasn't published it, leaving a security gap as trade diplomacy overrides enforcement.
Weibo's 3-billion-parameter VibeThinker matched DeepSeek V3's math scores, but skeptics say the benchmarks are too easy to game to trust the result.
Mobileye will launch its own Waymo rival robotaxi service in 2027, competing directly with the customers it supplies self-driving technology to.
AI search engines like Perplexity are breaking marketing attribution, and only 41% of teams can prove ROI despite 91% using AI daily.
U.S. senators are demanding NHTSA investigate Tesla's FSD safety claims, questioning whether its "7 times safer" statistic holds up to scrutiny.
xAI is acquiring Cursor for $60 billion to give Grok a coding superpower, buying developer trust and data instead of trying to win the model race.
Azure Databricks Lakebase adds database branching so GitHub Copilot agent mode can debug production systems using isolated, instant copies without risking.
Apple's new AI-powered Siri buries basic web search under five extra taps, making iOS 27's redesigned Spotlight far harder to use than before.
Jensen Huang says AI will create jobs, not kill them; a Texas factory expansion aims to prove it by adding 1,000 positions to U.S. manufacturing.
SpaceX hit a $2.97 trillion valuation after its IPO and a $60 billion Cursor acquisition, making Elon Musk the world's first trillionaire.
SpaceX's $2.1 trillion IPO valuation is built on unproven AI ambitions, not profits, prompting analysts to favor Amazon and Alphabet instead.
AI agents make hundreds of tool calls per second, and most security teams have no controls for it; here are the four primitives you need first.
Mistral AI nears $400M in revenue, but a viral hoax about a fake model briefly overshadowed its real 675B-parameter advances for investors.
Meta's CTO called the Applied AI division rollout "atrocious," prompting reforms to manager ratios, mobility options, and morale after staff compared it.
Meta's AI obsession has dismantled its legendary engineering culture, leaving top engineers devalued and triggering the company's worst outage in history.
AI Overviews now cut click-through rates by 58%; here's how Generative Engine Optimization helps businesses stay visible in AI search.
SpaceX acquired AI coding tool Cursor for $60 billion, raising questions about whether developers will lose access to Claude and GPT models.
The DOJ is defending Elon Musk's Grok AI in court, arguing Pentagon reliance on the model overrides environmental rules at xAI's Memphis data centers.
Satya Nadella says your real AI moat is verification infrastructure, not the model itself, yet Microsoft faces a lawsuit, low Copilot adoption, and GitHub.
HIVE Digital Technologies is betting $2.5 billion on a 100,000-GPU AI data center in Ontario, marking a dramatic pivot away from Bitcoin mining.
Figure AI logged 1,250 runtime hours at BMW with 99% placement accuracy, suggesting its Helix AI system is building a real competitive moat.
Google Cloud revenue surged 63% as Sundar Pichai's AI bet paid off, validating a 2024 prediction that sent Alphabet stock up 107%.
Sam Altman admits companies are using AI as cover for layoffs they'd make anyway, even as data shows AI hasn't delivered real productivity gains.
Nearly half of workers upload sensitive data to public AI tools without IT oversight, driving Tailscale to launch Aperture, a free AI access control.
The U.S. DOJ is defending Elon Musk's xAI data center in court, calling Grok a national security asset while Memphis residents face the second-worst.
HIVE Digital is building a 320 MW AI data center in Ontario, revealing how clean power, not chips, is becoming the real edge in the AI infrastructure race.
