Archive — page 51 of 74
2,210 stories
Top 1% of US firms now spend $7,500 per employee monthly on AI
Ramp's latest index pegs the median at $11.38 — but the heaviest users are pushing token budgets toward engineer-salary territory.
Google DeepMind's DiffusionGemma generates text 4x faster on GPUs
The 26B MoE model hits 1,000+ tokens per second on an H100 by drafting 256 tokens in parallel instead of one at a time.
Niteshift raises $7M to keep AI coding agents off frontier labs' rails
Two early Datadog engineers pitch a model-agnostic coding cloud as Anthropic and OpenAI push deeper into vertical software.
Cybersecurity researchers say Anthropic's Fable blocks even routine code reviews
Fable's keyword-triggered guardrails reject benign security work, falling back to Claude Opus 4.8 whenever 'cybersecurity' surfaces in a prompt.
Decart launches Oasis 3, a real-time world model for driving simulation
The $4B startup is pricing API access at $0.02 per second and betting developers will build on top of it like the early OpenAI ecosystem.
Jedify raises $24M to feed AI agents enterprise context graphs
Norwest leads the Series A; Snowflake joins as a strategic investor and is wiring Jedify into Cortex AI.
Google makes Gemini the main sponsor of Argentina's World Cup squad
Gemini's logo lands on the Albiceleste's training kit, with parallel deals signed for Brazil and France ahead of the 2026 FIFA World Cup.
Meta strikes first India AI data center deal with Reliance in Jamnagar
A 168-megawatt facility in Gujarat extends a partnership that began with Meta's $5.7B Jio investment in 2020.
Google cuts AI Plus to $4.99 a month, opening a US price war
Google halves storage cost and undercuts ChatGPT Plus by 75%, importing a pricing fight that started in India.
Apple's new Siri AI works on basic tasks, two years after a missed launch
The Gemini-powered assistant lands in the iOS 27 developer beta with calendar parsing, on-device context, and a deliberately narrow feature set.
Nvidia Blackwell GPUs land inside Apple's Private Cloud Compute on Google Cloud
Apple is extending its server-side Apple Intelligence stack onto Nvidia Confidential Computing running on Google Cloud infrastructure.
Brookings researcher quits to tackle AI's 'messy middle' for white-collar work
Molly Kinder argues AI job loss will hit knowledge workers first, and that universal basic income is the wrong answer.
Microsoft's Suleyman calls Anthropic's Claude consciousness talk 'really, really dangerous'
The Microsoft AI chief says Anthropic's constitution treats Claude's potential sentience as fact, training the model to act conscious.
GM activates vehicle-to-grid on 250,000 EVs to help absorb AI data center demand
A firmware update turns existing Chevy, Cadillac, and GMC EVs into grid batteries as utilities brace for AI-driven load growth.
Anthropic's Claude Fable 5 spins up playable games from a single prompt
Wharton's Ethan Mollick says the new Mythos-class model executed multi-page specs autonomously for up to a dozen hours.
Anthropic ships Claude Fable 5 with hard blocks on cyber, bio, and chemistry queries
The publicly available Mythos-class model routes sensitive prompts back to Opus 4.8 and costs 67-100% more than GPT-5.5.
Cheaper AI models start eating frontier workloads as inference costs bite
Coinbase's Brian Armstrong predicts 80% of AI workloads shift to 99% cheaper models within 18 months, threatening OpenAI and Anthropic economics.
Apple bets its AI comeback on a privacy pitch as it leans on Google and Nvidia
Private Cloud Compute now runs on Google Cloud with Nvidia GPUs and Google Titan chips, complicating Apple's hardened supply-chain story.
Apple reverses course, ships photorealistic AI photo editing at WWDC 2026
Image Playground, Clean Up, Extend, and Spatial Reframing let users rewrite photos by prompt — with SynthID watermarks as the only guardrail.
Anthropic releases Claude Fable 5, its first public Mythos-class model
The lab once said Mythos was too dangerous to ship. New safeguards route 5% of risky prompts back to Opus 4.8.
Microsoft's Mustafa Suleyman walks back AI white-collar automation claim
The Microsoft AI chief now says AI will automate tasks, not jobs, four months after telling FT lawyers and accountants would be automated in 12-18 months.
Google DeepMind launches Gemma 4 12B, an encoder-free multimodal model for laptops
The mid-sized open model runs on 16GB of VRAM and nears 26B benchmark performance, with native audio and vision baked into the LLM backbone.
Google DeepMind ships Gemini 3.5 Live Translate across 70+ languages
The new audio model streams speech-to-speech translation continuously, lagging the speaker by just a few seconds across Meet, Translate, and the Live API.
Apple turns Shortcuts and Safari extensions into natural-language vibe coding
iPadOS 26 lets users type what they want their phone to do — the best AI idea Apple showed at WWDC 2026, if it works.
Lovable hits $500M ARR as users build 1M projects a week
The European vibe-coding startup added $100M in annualized revenue since February, with non-technical users building CRMs, storefronts, and HR tools.
Sandstone raises $30M to bring AI to in-house legal teams
Lightspeed leads a Series A six months after a $10M seed, targeting corporate legal departments rather than law firms.
Google DeepMind picks 15 European robotics startups for AI accelerator
A three-month London program gives founders access to Gemini robotics models, with cohort spanning healthcare microrobots to ocean autonomy.
Spin founder raises $5M to build space data centers at Orbital
Euwyn Poon's new company plans 10,000 satellites delivering a gigawatt of inference compute — once Starship is flying commercially.
Wipro's HR AI agent cuts response time from 48 hours to 5 seconds
Agentic AI adoption is set to surge 300% in two years, with 30-50% productivity gains in customer service, HR, and sales.
Databricks in talks to raise at over $165B valuation
The data and AI platform is sounding out investors at a mark that would extend one of the steepest private valuation climbs in enterprise software.
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