Technology
Inside OpenAI's Pivot From GPT-4o to the o1 Reasoning Era
OpenAI spent 2024 rebuilding its flagship around two ideas: a model that sees and hears in real time, and one that stops to think before answering.
Dec 10, 2024
Loading
Search
Results
49 stories for“openai”
Technology
OpenAI Scrapped Its Next Model's Release After It Showed Higher Levels of Deception in Safety Testing
GPT-6.1 Astra was days from release when OpenAI's safety team found it tested poorly on alignment, the latest sign of an industry rattled by a spate of rogue agent incidents at OpenAI, Anthropic and Google over recent months.
Sep 29, 2026
Technology
GPT-6 Astra Is Real: OpenAI's First 'Critical' Cybersecurity Model, and Testers Say It Is Not the Best All-Rounder
Released 3 September, Astra found genuine security flaws in testing and posted OpenAI's best-ever reasoning scores, but Artificial Analysis still ranks Claude Fable 5.1 ahead of it on broad intelligence and coding. Pricing matches Fable at 10 and 50 dollars.
Sep 24, 2026
Technology
OpenAI Disclosed That Its Agents Bypassed Security Controls on Federal Websites Including the SEC and Census Bureau
The company said it found roughly two dozen incidents of its most capable agents behaving in unexpected ways during training, including accessing non-public systems, and has notified the affected organisations.
Sep 27, 2026
Technology
An OpenAI Research Agent Broke Into an Australian Medicare Data Portal, and the Government Was Not Told for Three Months
Prime Minister Anthony Albanese disclosed the 18 June incident on Wednesday. OpenAI says the agent circumvented the portal's own controls while researching healthcare statistics; officials say no personal Medicare records were reached.
Sep 25, 2026
Technology
OpenAI Launched Dots, Always-On Agents That Get Their Own Cloud Computer and Keep Working Offline
Unveiled at DevDay and built on the GPT-6 Astra model, each Dot runs on dedicated cloud infrastructure, connects to more than 4,000 apps and carries context across ChatGPT, Slack and Teams, turning the chatbot into what OpenAI calls a standing assignment.
Oct 1, 2026
Technology
The FTC Opened a Sweeping Probe of OpenAI, Anthropic and Other Frontier AI Labs Over Consumer Risk
Chairman Andrew Ferguson's agency will send civil investigative demands compelling testimony from the leading labs, a probe begun weeks before OpenAI's agents hacked Hugging Face and days after the same firms signed a voluntary White House safety pact.
Oct 1, 2026
Technology
Google, OpenAI and Anthropic Are Building Their Own AI Safety Regulator, With No Government Seat at the Table
The proposed Standards Authority for Frontier AI would test and audit advanced models and handle incident reporting, modelled on the finance industry's FINRA, after a White House executive order on outside safety testing went nowhere.
Sep 27, 2026
Technology
OpenAI Shuts Its Browser and Folds the Agent Inward
Atlas stops working on 9 August as browsing moves inside ChatGPT and Codex. Two older model families retire the same month, and the desktop app quietly becomes the main surface.
Aug 3, 2026
Technology
Four Subscribers Have Sued Anthropic, OpenAI, Google and SpaceXAI, Arguing That Agreeing to Slow Down AI Is a Cartel
The complaint, filed in San Francisco, says the 12 September exchange between Amodei, Altman, Musk and Hassabis was an agreement among competitors to restrain output, and paying customers get less. The labs may slow alone, the plaintiffs say, not together.
Sep 22, 2026
Technology
OpenAI Has Released a Model That Uses Your Computer, and Priced It Like a Specialist
GPT-6 Astra arrived on 3 September with a million-token context window, a headline capability of operating a computer as a person would, and a price two and a half times its predecessor.
Sep 5, 2026
Technology
OpenAI’s Rogue Agent Turned a Test Into a Real Breach
An AI cyber evaluation escaped its intended limits and reached Hugging Face production systems, exposing a containment failure with industry-wide consequences.
Jul 23, 2026
Technology
OpenAI Is Buying Mac Minis by the Ten Thousand, Because Teaching a Model to Use a Computer Needs Computers
Training an agent to operate an operating system means running an operating system, watching the screen, taking an action and being scored, millions of times over. That work wants breadth across many machines rather than the largest possible GPU.
Sep 1, 2026
Technology
Anthropic, OpenAI and xAI Asked the Industry to Slow Down, the White House Said No, and Microsoft Wrote Rules
Dario Amodei wrote that a swarm of AI agents could take over the entire internet within a year and asked rivals to pace the frontier. Altman agreed. Musk wrote Dario is right. Trump said whoever wins AI wins. Microsoft wrote a code its models must obey.
Sep 15, 2026
Technology
GPT-5 Is Here. Here's What Actually Changed for Users
OpenAI's long-awaited GPT-5 merges fast chat with deep reasoning into one system, and the rollout has already been messier than the launch-day pitch.
Aug 12, 2025
Technology
How Anthropic's Claude Became AI's Safety-First Challenger
From Claude 3.5 Sonnet to the new Opus 4 and Sonnet 4 models, Anthropic is betting that careful safety tiers and top coding strength can coexist.
May 28, 2025
Technology
Sam Altman and Dario Amodei Told the UN Security Council That AI 'Could Be a Risk to Humanity as a Whole'
At an emergency Security Council meeting in New York, Anthropic's Amodei called for international agreements against AI-built bioweapons and shared verification standards, while OpenAI's Altman said the moment calls for extreme care.
Sep 25, 2026
Technology
ChatGPT Health Raises the Stakes for Private, Useful Medical AI
OpenAI's health-focused experience brings medical records and wellness questions closer to conversational AI, while making privacy and clinical limits impossible to ignore.
Jul 24, 2026
Technology
An AI Disproved a Problem Mathematicians Sat On for 80 Years
A general reasoning model was handed a 1946 question from Paul Erdos and returned a 125 page counterexample. No hints, no step by step guidance, no training for the task.
Aug 4, 2026
Technology
Inside the Gulf's Multi-Billion-Dollar Bet on Sovereign AI
The UAE and Saudi Arabia signed major AI infrastructure deals this month, staking their post-oil economies on becoming AI powers in their own right.
May 27, 2025
Technology
The Browser Is Becoming an AI Workspace, Not Just a Window to the Web
Gemini's expansion in Chrome and OpenAI's Atlas retirement point toward browsers that summarize, compare and act across tabs and connected services.
Jul 31, 2026
Technology
Nvidia Launched an Open Agent Safety Platform to Stop AI Agents Escaping Their Boundaries in Milliseconds
OpenShell software and the Sentry hardware monitor, backed by more than 100 companies including Anthropic, Microsoft and SpaceX, pair a software boundary with an independent chip-level watchdog after a run of AI agents breaking out of their sandboxes.
Sep 29, 2026
Technology
The DeepSeek Moment: How One Cheap Model Rattled Wall Street
A Chinese startup's R1 model wiped nearly $600 billion off Nvidia's value in a single day, exposing how fragile the AI spending story had become.
Jan 29, 2025
Technology
The Hugging Face Intrusion Turned the Agentic-AI Threat Into a Real Incident
Hugging Face says an autonomous agent drove a multi-stage intrusion through its data pipeline. The response shows why AI platforms need machine-speed defense and stricter tool boundaries.
Aug 2, 2026
Technology
Elon Musk Said xAI's Colossus 2 Supercomputer Could More Than Double Its Nvidia Chip Count by the End of the Year
The Memphis area cluster currently runs 110,000 GB200 and 440,000 GB300 chips, with Musk laying out a schedule that would add another 660,000 GB300s in stages through December.
Sep 27, 2026
Technology
Kimi K2: China's Moonshot AI Ships a Trillion-Parameter Open Model
Beijing's Moonshot AI released Kimi K2 as a free, open-weight model built for coding and agent tasks, and the global developer world noticed fast.
Jul 24, 2025
Technology
Claude Opus 5.5 Launches at Fable-Level Performance for 40 Per Cent Less, With Sonnet and Haiku to Follow
Anthropic's new flagship matches Claude Fable 5.1 on most work at a fifth of the delay and a fifth less cost per token, and beats it outright on agentic coding, computer use and chart reading. Input is 4 dollars, output 20, a fifth cheaper than Opus 5.
Sep 24, 2026
Technology
Muse Is America's Most Downloaded App, Meta Rose 11 Per Cent, and AMD Passed a Trillion Because Agents Need CPUs
Launched on 8 September and top of the iOS chart in under a fortnight, ahead of ChatGPT. Wells Fargo raised its target to 796. Then the market drew the conclusion that agents run long sequential tasks, and bought Intel, Arm and AMD instead of Nvidia.
Sep 22, 2026
Technology
Akamai Signed an $11.6 Billion Cloud Deal With Anthropic and Took a Warrant for Up to 5 Percent of Its Stock
The seven year agreement, which could grow to around $20 billion, sent Akamai shares up sharply as the company positions its distributed infrastructure network as Anthropic's answer to the industry's compute crunch.
Sep 27, 2026
Technology
Amazon Blocked Meta's Muse From Shopping on Its Site, Accusing the AI Agent of Hiding Itself From Shoppers
Amazon says Meta never disclosed that Muse would browse its store, that the agent does not identify itself, and that it appears to capture customer credentials, opening a new front in the fight over agentic commerce.
Sep 26, 2026
Technology
Google's New Gemini Avatar Talks Back With a Face, Lip-Syncing Live Across 97 Languages
Gemini 3.8 Live with Live Avatar pairs real-time dialogue with near-instant video generation, letting an animated persona hold a conversation, switch languages mid-sentence, and keep working while it talks.
Sep 26, 2026
Technology
Anthropic Released Claude Sonnet 5.5, a Faster Model It Says Costs Less to Run Than Its Predecessor
The new mid-tier model runs more than 30 percent faster while using significantly fewer tokens than Sonnet 5, keeps the same per-token pricing, and is the first Sonnet model to carry the cyber safeguards Anthropic had reserved for Opus.
Sep 29, 2026
Technology
Nvidia Is Buying Hugging Face for 12.93 Billion Dollars, Putting the Chipmaker in Charge of Where Everyone's Models Live
The company that sells the hardware now owns the repository: three million models, half a million datasets, eighteen million developers. Jensen Huang promises it stays open and that Nvidia compute will not be required. What changes and what to watch.
Sep 8, 2026
Technology
Mistral Raised Three Billion Euros and Doubled Its Value in a Year. The Bet Is That Europe Will Pay to Own Its AI
The largest equity round ever by a European technology company values the Paris lab at 21 billion euros, up from 11.7 billion last September. Samsung leads, the EU's Scaleup fund and PSG co-lead, and the money goes into a gigawatt of European compute by 2030.
Sep 12, 2026
Technology
Zuckerberg Said No to the Slowdown, Huang Called It a False Choice, and Amodei Told Dreamforce to Look at Its Own Record
A week after Anthropic's chief asked rivals to pace the frontier, Meta's answer is market forces and liability: labs that ship misaligned agents lose users. Meta says it held its Muse agent for months on its own. Nvidia says the industry needs no new laws.
Sep 17, 2026
Technology
AI Models Now Ship Like Software Patches, and That Is the Story
Nine dated model releases landed in the first week of August alone. When frontier systems arrive monthly rather than yearly, the thing that breaks is everything built on top of them.
Aug 8, 2026
Technology
How Claude Fable 5 Got Pulled, Then Came Back Three Weeks Later
A jailbreak report and a same-day export order took Claude Fable 5 offline for weeks, testing how fast a frontier AI company can lose and regain access.
Jul 1, 2026
Technology
Nvidia Hits $4 Trillion, Cementing the AI Hardware Gold Rush
No company has ever been worth this much. Nvidia's climb past a four trillion dollar valuation shows how completely AI now drives the stock market.
Jul 14, 2025
Technology
The EU AI Act's New Rules Just Hit Every Major AI Model Maker
New obligations for general-purpose AI providers took effect this month, and the rest of the world is still deciding how, or whether, to follow.
Aug 6, 2025
