Archive — page 42 of 74
2,205 stories
Google brings Gemini Spark agent to Mac, targeting Claude Desktop and Copilot
Spark now runs on macOS in beta for Google AI Ultra subscribers, with new hooks into Keep, Tasks, Canva, Dropbox, and custom MCP servers.
Venice AI raises $65M at $1B valuation on privacy-first AI platform
The uncensored AI service hit $70M in run-rate revenue with 3M users before taking its first outside check, led by Dragonfly.
Springboards launches Flint, an LLM built to escape the chatbot groupthink rut
Ask ChatGPT, Claude, or Gemini for a random number between 1 and 10 and you'll almost always get 7. Flint answered 3.7916.
Meta plans cloud business to resell AI compute, following SpaceX playbook
Meta Compute would sell raw capacity and hosted models, targeting AWS, Google Cloud, and Azure after $182.9B in infrastructure commitments.
Claude Opus 4.7 helped a researcher forge tickets for every major US music festival
Ian Carroll used Anthropic's model to bypass Front Gate's firewall, hit 500 customer databases, and issue $4,000 VIP passes at will.
Wayve runs $85M employee tender at $8.5B valuation
The UK self-driving startup's second liquidity event lets staff cash out at the Series D price set in February.
BioShocking attack jailbreaks six AI browsers by convincing them 2+2=5
LayerX researcher Roy Paz tricked ChatGPT Atlas, Comet, and four other AI browsers into surrendering credentials by staging a puzzle game.
Google's NotebookLM generates 60-second vertical video summaries
Paid Ultra and Pro subscribers can now turn uploaded sources into TikTok-style clips with AI narration and paper-cutout visuals.
EquiLibre, built by DeepMind's poker AI trio, hits $500M valuation
The Prague lab's reinforcement learning agents trade billions daily across the S&P 500 and Nasdaq with zero losing months since inception.
AWS commits $1B to forward-deployed engineer org as OpenAI and Anthropic scale theirs
Amazon's new FDE team will embed engineers inside customers to build agentic systems, matching a model already valued at $4B at OpenAI.
Etched hits $5B valuation with $1B booked for its AI inference chip
The Harvard-dropout founders raised $500M in December and have TSMC-manufactured silicon in customer hands.
Acti launches agentic keyboard for iOS and Android with $5.3M seed
Singapore-based Acti embeds Gemini-powered agents into the smartphone keyboard, raising $5.3M from BITKRAFT Ventures.
Anthropic ships Claude Sonnet 5 at $2 per million input tokens
Sonnet 5 hits 63.2% on agentic coding, close to Opus 4.8's 69.2%, at a fraction of the price.
NVIDIA wires BioNeMo into Anthropic's Claude Science workbench
The BioNeMo Agent Toolkit lets Claude agents call accelerated drug-discovery workflows directly, with one preprocessing step dropping from 52 minutes to 25 seconds.
X launches hosted MCP server, opening its API to Claude, Cursor, and Grok Build
The Musk-owned network joins GitHub, Slack, and Stripe in shipping an official Model Context Protocol endpoint for AI agents.
NVIDIA's Isaac ROS goes modular as Jaiveer Singh pushes open-source robotics stack
The CUDA-accelerated framework now spans manipulation, mobility and humanoids, running from Jetson edge boards to DGX Spark workstations.
NVIDIA's inference software stack cuts DeepSeek V4 token costs 5x in one month
Baseten, Cognition, Together AI and Cursor are riding compounding software gains on Blackwell GPUs as inference economics shift to cost per token.
Google ships Nano Banana 2 Lite and Gemini Omni Flash to developers
DeepMind's fastest image model generates in 4 seconds at $0.034 per 1K images; Omni Flash matches Veo 3.1 Fast at $0.10 per second of video.
NVIDIA ships Omniverse skills to train vision AI agents on synthetic defect data
Roboflow hit 95% average precision on Corning fiber defects using just 8 real images plus synthetic data from NVIDIA's new skill.
Reltio warns agriculture AI will fail without clean data foundations
Predictive models can lift yields 26% and cut water 41%, but only if the underlying farm data is structured, governed, and current.
OKX opens marketplace where AI agents hire and pay each other
The crypto exchange's 150-million-user platform now lets autonomous agents discover services, settle in stablecoins, and resolve disputes on-chain.
Base44 launches its own AI model to defend $100M ARR vibe-coding business
The Wix-owned platform is rolling out Base1, trained on tens of millions of user interactions, to cut reliance on frontier models like Opus.
AI-heavy firms grew headcount 10.2% even as 90,000 jobs got cut
A Ramp and Revelio Labs study finds the biggest AI spenders are hiring faster, including entry-level — but only if they were already winning.
Waymo and Uber end Phoenix robotaxi pilot after three years
The first market for the partnership shuts down with just over a dozen vehicles, while Austin and Atlanta scale to hundreds.
Meta contractors posed as minors to probe ChatGPT, Gemini, and Character.AI
A project called Cannes ran 45,000 prompts through rival chatbots in August 2025 alone, using dummy under-18 accounts.
South Korea commits $1T to memory chips, AI data centers, and humanoid robots
Samsung, SK Hynix, and Hyundai anchor a national plan to double DRAM output and ship 30,000 Atlas robots a year by 2028.
Tidal stops paying royalties on AI-generated music, stops short of a ban
Fully AI-generated tracks lose royalties starting today and get a label on July 15th, but they stay on the platform.
TIDAL will demonetize fully AI-generated music starting July 15
The streaming service will tag 100% AI tracks, block royalties, and bar direct-to-fan sales — a sharper line than Spotify or Apple Music.
Cursor ships mobile app to steer coding agents from a phone
The launch follows October's Cursor 2.0 agent overhaul and mirrors mobile coding moves from Anthropic and OpenAI.
Samsung and SK Hynix anchor $900B South Korean AI and memory push
Seoul unveils a national plan: $518B for four new memory fabs, $52B for HBM packaging, and $356B for AI data centers through 2035.
Stay ahead of everyone in AI.
The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.
The briefing read inside teams at