Tech·4,618 stories
Tech News — Page 72
Nobody Knows How the US Greenlights Frontier AI Models
OpenAI is rolling out its latest frontier model, Sol, to the public — but nobody outside the company or the White House can explain how the US government decided it was safe to release, and the proces
OpenAI launches GPT-5.6 publicly, debuts ChatGPT Work
OpenAI secured the Trump administration's approval for a full public rollout of GPT-5.6 after a two-week "limited preview" limited to government-approved organizations, and on the same day unveiled Ch
Meta's New AI Chips Enter Production in September
Meta is on track to begin production of its latest AI-specific MTIA chips in September, per an internal memo cited by Reuters, as the company looks to cut GPU costs during a component shortage and $12
Nvidia Stock Drops 15%, Micron Nearly Triples
Nvidia's stock has fallen 15% from its May peak despite continued revenue growth, as memory chip makers like Micron capture the AI infrastructure windfall with DRAM prices rising tenfold and no compet
Claude Reflect turns usage stats into a retention play
Anthropic rolled out Claude Reflect, an analytics dashboard that visualizes users' AI habits, prompts mindfulness questions, and steers them toward deeper Claude integration — a retention tool dressed
SpaceX, Anthropic, OpenAI Exits Eclipse 25 Years of VC IPOs
The combined SpaceX, Anthropic, and OpenAI exits will generate more value than all U.S. VC-backed exits since 2000, per the NCVA-Pitchbook Venture Monitor, with the trio likely topping $4 trillion aga
Meta Muse Spark 1.1 API at 25% of OpenAI, Anthropic
Meta released Muse Spark 1.1 to US developers via a public API preview, with pricing at roughly 25% of OpenAI and Anthropic, positioning the model as a "step-change" in coding capability aimed at the
Meta Opens Muse Spark 1.1 Coding API to US Devs
Meta has launched Muse Spark 1.1, an upgraded AI model with stronger coding and agentic capabilities, opening it to US developers through a new Meta Model API preview with $20 in free credits.
Apple Eyes PrismML to Run 27B AI Models On-Device
Apple has met with PrismML, a Khosla-backed startup claiming it can run a 27-billion-parameter AI model directly on an iPhone 17 Pro without servers — a potential boost to Apple's on-device AI strateg
Anthropic Launches Claude Reflection Dashboard in Beta
Anthropic launched a "reflect" feature for Claude that analyzes users' conversation patterns and prompts them to reconsider what they delegate to AI. The dashboard requires Claude's memory to be enabl
Character.AI Launches AI-Generated Interactive Microdramas
Character.AI has launched c.ai Series, animated generative-AI microdramas that come bundled with post-episode character chatbots, pushing into a projected $26 billion vertical-video market while keepi
SpaceXAI Debuts Grok 4.5 Built With Cursor
SpaceXAI unveiled Grok 4.5, its first model built in partnership with coding startup Cursor and aimed at handling long-running legal, finance, and coding tasks for enterprise users.
Nilekani Steps Down as GP at Fundamentum's $200M Fund
Nandan Nilekani is stepping down as general partner at Fundamentum Partnership while remaining its anchor investor, as the venture firm launches a ~$200 million third fund targeting Indian early-stage
Texas Data Centers Used Dry-Cleaner Permits to Skip Reviews
A Texas loophole has let 38 data centers—including OpenAI's Stargate campus in Abilene—secure minor air permits meant for dry cleaners, sidestepping environmental review while building gas plants that
Meta's Muse Image Uses Public Instagram Posts by Default
Meta's new Muse Image AI model, built by its Superintelligence Labs, uses public Instagram posts and reels by default to generate AI content, and users won't be notified when their images are remixed.
SpaceXAI Launches Grok 4.5 With Cursor
SpaceXAI unveiled Grok 4.5, its first model co-developed with coding startup Cursor, positioning it for long-running legal, finance, and coding tasks at a price multiple outlets describe as undercutti
AI Code-Scan Agents Tricked Into Running Attacker Code
Researchers at the AI Now Institute demonstrated that autonomous AI coding agents from Anthropic and OpenAI can be tricked by booby-trapped README files into executing attacker code on the user's mach
Meta's 'Super Sensing' Glasses Would Record Continuously
Meta is developing prototype 'super sensing' smart glasses that would continuously record audio and photos, but under a proposed design, raw footage would be processed as metadata for Meta AI to query
SpaceXAI releases Grok 4.5, which Elon describes as an ‘Opus-class model’
SpaceXAI launched Grok 4.5, its first model since going public, positioning it as a cheaper, more token-efficient alternative to Anthropic's Opus class — with OpenAI's GPT 5.6 release following a day
OpenAI Launches GPT-Live Voice Models
OpenAI launched GPT-Live, new voice models that listen and speak simultaneously in real time, rolling out two versions globally as the IPO-bound startup expands its audio portfolio.
Get the Tech newsletter
Curated tech stories, every morning. Free.
No spam. Unsubscribe anytime.