AI & Machine Learning

WIRED and AD say the future home is analog, not smart

DOJ claims xAI's Grok is essential to US military ops

Estonian benchmark tests 60 AI models for Russian propaganda

Anthropic pauses Claude SDK billing overhaul amid OpenAI price war

OpenAI spent $34 billion in 2025, lost $39 billion

100+ cyber execs push Trump to lift Anthropic AI export ban

Anthropic shutdown forces Europe to confront AI dependency

Pokémon Go scans trained AI now powering military drones

Nvidia raises $20B in bonds as AI debt boom accelerates

DeepSeek raises $7.4B at $50B valuation in first funding round

Nadella warns AI could hollow out industries

US Demands Unhackable AI Models. Experts Say That's Impossible.

The Trump administration is accusing Anthropic of releasing Fable 5 despite security risks, but the demand for 'unhackable' AI models reveals a fundamental misunderstanding of how large language models work. Over 100 security experts have signed an open letter calling the export controls misguided.

Moonshot's Kimi K2.7 Code Costs 12x Less Than GPT-5.5

Chinese AI lab Moonshot AI released Kimi K2.7 Code, an open-weights model priced at $0.95 per million input tokens. While it trails GPT-5.5 and Claude Opus 4.8 on standard coding benchmarks, the model outperforms Claude on agent-focused tests and offers substantial cost savings for teams running heavy coding workloads.

Meta Employees Reject Zuckerberg's AI Hackathon After Layoffs

Meta's 70,000 remaining employees are pushing back against a companywide AI hackathon scheduled for July. Internal messages reveal frustration over workload increases following 8,000 layoffs, declining trust in management, and concerns that hackathon work won't count toward performance reviews.

Bezos' AI Startup Prometheus Raises $12B at $41B Valuation

Jeff Bezos' industrial AI company Prometheus has closed a $12 billion funding round, reaching a $41 billion valuation less than a year after launch. The startup aims to build an 'Artificial General Engineer' to automate physical product design and manufacturing.

Deezer's Free AI Music Detector Works on Spotify, Apple Music

Deezer has released a free tool that scans playlists on any major streaming platform for AI-generated tracks. The company says 44% of its daily uploads are now AI music, and 43% of users switching from competitors already have synthetic tracks in their libraries.

OpenAI IPO Slips to 2027 as Altman Cites Strategic Uncertainty

Sam Altman told OpenAI employees to expect a public offering 'within the next year,' pushing the timeline into 2027. The delay comes as rival Anthropic prepares to go public and OpenAI continues burning through cash at a projected $14 billion annually.

AI Cracks Security Patches in Hours, Not Weeks

Anthropic's security research team tested how fast large language models can reverse-engineer software patches into working exploits. The answer: hours, not the weeks defenders have long assumed. This finding upends traditional patch management assumptions across the industry.

Anthropic Launches Claude Fable 5 Tomorrow, Mythos Goes Public

Anthropic is set to release Claude Fable 5, a public version of its powerful Mythos AI model that was previously restricted due to cybersecurity concerns. The model found over 23,000 critical vulnerabilities during its gated research phase, including decades-old flaws in major software systems.

OpenAI Files Confidential IPO as AI Valuations Hit $1 Trillion

OpenAI has submitted confidential S-1 paperwork for an initial public offering, joining Anthropic and SpaceX in what could become a trio of trillion-dollar tech IPOs. The company raised $122 billion privately in March and now seeks public market capital to fund data center expansion and AI research.

Instagram AI Chatbot Bug Exposed 20,000 Accounts to Hackers

A flaw in Meta's AI-powered account recovery chatbot let attackers send password reset links to any email address without verification. The vulnerability went undetected for seven weeks, potentially compromising over 20,000 Instagram accounts.

DeepSeek Tops Ramp's June 2026 Vendor List as US Firms Cut AI Costs

Chinese AI company DeepSeek became the fastest-trending software vendor among US businesses in June 2026, according to Ramp data covering $100 billion in annual enterprise spend. The shift signals growing price sensitivity as companies seek alternatives to OpenAI and Anthropic, though security concerns persist.

OpenAI Declares 'Chat Is Dead,' Rebuilds ChatGPT as Agent App

OpenAI is overhauling ChatGPT from a question-answering chatbot into an autonomous agent capable of handling tasks across work and personal life. The shift, confirmed by Chief Product Officer Thibault Sottiaux, represents the biggest change to ChatGPT since its 2022 launch.

Anthropic Hires OpenAI's Second Chip Engineer Ahead of IPO

Clive Chan, the second hardware hire in OpenAI's custom chip program, has joined Anthropic. The move signals Anthropic's growing interest in building proprietary silicon as both companies prepare for public offerings.