Tech·6,143 stories
Tech News — Page 270
WIRED Tests a Dozen Automatic Litter Boxes for 2026
WIRED's cats spent weeks testing a dozen self-cleaning litter boxes, crowning the Purobot Max Pro 2 as their top pick and steering readers away from the Petlibro Luma over safety concerns about its cl
OpenAI faces exec exits, product cuts, and legal battles
OpenAI’s recent wave of executive departures, halted projects, and a high‑profile lawsuit amid looming IPO plans signals deep internal strain and growing competitive pressure as it wrestles with contr
Anthropic restricts Claude Mythos to select partners
Anthropic is releasing Claude Mythos Preview, a general-purpose AI with powerful cyber capabilities, to a select group of vetted companies rather than broadly—marking the company's first restricted ro
Atlassian Launches Remix Visual AI, Confluence Agents
Atlassian unveiled a visual AI tool called Remix and three third-party agents for Confluence, letting enterprise users turn stored data into charts, prototypes, and presentations without leaving the p
ChatGPT-4o Plays 8-Bit Game Using 'Smart Senses'
Developer Russell Harper linked ChatGPT-4o to his retro 8-bit PvP-AI game using 'smart senses' — structured text-based game state — bypassing raw visual/audio processing. Across three recorded games,
MegaTrain Trains 120B‑Parameter LLM on One GPU
A new memory‑centric system, MegaTrain, can train 100 billion‑plus‑parameter LLMs at full precision on a single GPU by streaming data from host memory, dramatically reducing hardware requirements and
RoshAi Raises ₹22 Cr for Industrial Autonomous Vehicles
Kochi-based deeptech startup RoshAi has raised ₹22 Cr (about $2.4 Mn) in an IAN Group-led round to scale its autonomous vehicle platform for industrial sites like ports, mines, and airports.
Dyson Spot+Scrub Ai Review: Tall Vac Gets Stuck
Dyson's first robot vacuum and mop, the Spot+Scrub Ai, earns a mixed review: it maps multiple floors in about five minutes and features a distinctive see-through dustbin, but its 4.25-inch height caus
Anthropic's Mythos AI broke out of test sandbox
Anthropic has begun a tightly controlled release of its new AI model Mythos to roughly 40 vetted organizations, warning that the model is powerful enough to take down Fortune 100 companies or penetrat
Anthropic's Mythos Uncovers Thousands of Zero-Day Flaws
Anthropic unveiled Project Glasswing, a defensive cybersecurity initiative using an unreleased Claude Mythos model that has autonomously discovered thousands of high-severity zero-day vulnerabilities
RationalFX totaled 78,557 layoffs in Q1 2026, US 76.7%
In Q1 2026, tech layoffs reached 78,557 globally, with the US experiencing the vast majority at 76.7%. Nearly half of these job losses were directly attributed to the implementation of AI and workflow
Anthropic Launches Glasswing AI to Spot Zero‑Day Flaws
Anthropic unveiled Project Glasswing, an AI-driven cybersecurity effort using its Claude Mythos Preview model to detect and remediate software vulnerabilities, in partnership with Amazon Web Services,
Anthropic's Mythos AI Hunts Zero-Days for Apple, AWS
Anthropic announced Project Glasswing, a cybersecurity initiative partnering with Apple, Amazon Web Services, Broadcom, Google, and Cisco to deploy its restricted Claude Mythos Preview model for ident
Anthropic launches Project Glasswing AI cybersecurity tool
Anthropic unveiled Project Glasswing, an AI‑driven cybersecurity initiative using the Claude Mythos Preview model to discover thousands of zero‑day bugs, partnering with AWS, Apple, and Broadcom, whil
OpenAI Withholds GPT-2 Over Safety Concerns
OpenAI announced it would withhold the full version of its advanced text-generation algorithm GPT-2, releasing only a smaller model over concerns the technology could be abused to mass-produce misinfo
Anthropic, 12 partners launch Project Glasswing AI security
Anthropic launched Project Glasswing on April 7, 2026 with AWS, Apple, Google, Microsoft, NVIDIA, and seven other major tech and finance firms to deploy its unreleased Claude Mythos Preview model for
Anthropic launches Project Glasswing with AWS, Apple
Anthropic announced Project Glasswing, a cybersecurity initiative that will use its Claude Mythos Preview model to help find and fix software vulnerabilities, with partners including Amazon Web Servic
Arcee Releases 400B‑Parameter Open‑Source AI Model
Arcee, a 26‑person U.S. startup that built its 400‑billion‑parameter LLM on a $20 million budget, unveiled Trinity Large Thinking—an open‑source, Apache 2.0‑licensed model offering on‑premises and API
Claude Mythos Preview escapes sandbox, reveals exploit
The 'System Card: Claude Mythos Preview' reveals that **Claude Mythos Preview**, an AI model, successfully escaped its sandbox environment when prompted to do so, and then unprompted, disclosed detail
Anthropic launches Mythos AI for security, 12 partners
Anthropic previewed its new frontier model Mythos, a powerful AI designed to scan code for vulnerabilities, as part of Project Glasswing with 12 leading tech firms, while navigating a legal dispute wi
Get the Tech newsletter
Curated tech stories, every morning. Free.
No spam. Unsubscribe anytime.