Anthropic, together with Amazon, Microsoft, and Google, proposed the CJS scale (CJS-0 to CJS-4) — the first cross-industry standard for rating AI jailbreak severity across four axes: capability gain, breadth, weaponizability, and discoverability.
Sustainability reports from Google and Amazon revealed Google's carbon emissions rose 25% year-over-year and Amazon's 16%, as AI infrastructure expansion has made previously pledged carbon neutrality goals significantly harder to achieve.
The UN Independent International Scientific Panel on AI released a preliminary report warning that AI capabilities are advancing faster than regulatory frameworks. The Global AI Governance Dialogue begins July 6 in Geneva.
Crusoe, the AI data center company with compute contracts for Meta and Oracle, is in talks to raise $3 billion in a round that may triple its valuation from $10 billion to $30 billion, as it contracts 4.9 GW of computing power with 40+ GW in pipeline.
Kling AI, a spinoff from Chinese short-video giant Kuaishou Technology, raised an initial $2 billion in venture capital to expand its AI video generation operations.
The righttointelligence.org campaign asserts that people should have the legal right to download, own, run, study, and modify open AI models locally without licensing requirements, warning that new state laws could put local AI behind a permission wall.
Manufact (YC S25) launched a cloud platform for MCP servers — enabling deployment, testing, monitoring, and app store distribution for AI agents like Claude, ChatGPT, and Cursor. Hit 97 points on HN on launch day.
A practical methodology for working with AI coding agents without losing code quality or control: no YOLO mode, per-subtask commits, mandatory dual AI+human PR review, and full author self-review — directly addressing the spike in production incidents from AI-assisted coding.
GitHub Copilot Vision became generally available on July 1, 2026 for all Copilot plans including Free. Developers can attach images and PDFs directly to chat prompts so Copilot can reason about visual context alongside code.
Copilot CLI from July 1 automatically selects the best model per task (Auto mode, 10% discount). Enterprise customers now get real-time streaming of Copilot agent session data into external tools via a streaming endpoint or REST API.
Anthropic on July 22 updated the Claude Managed Agents API: configurable effort levels on model configuration, webhooks for environment and memory store lifecycle events, session seeding with initial events, and event deltas for subagent thread streaming.
xAI shipped a rapid series of Grok Build patches (0.2.66–0.2.73) over three days: sandbox profiles with kernel-deny lists, no-restart MCP server management, new agent dashboard with model visibility, and a fix for stdio hangs on Windows.
Claude Opus 4.8 and Haiku 4.5 are now GA in Microsoft Azure Foundry with native Azure authentication, consolidated billing through Microsoft Enterprise Agreements, and optional US data zone for compliance.
Simon Willison released llm-coding-agent 0.1a0, an open-source Python library implementing a Claude Code-style agentic coding loop with tools for file reading/editing and shell command execution, demonstrated creating a SwiftUI CLI app for ASCII art time display.
Anthropic released Claude Apps Gateway — a self-hosted control plane for enterprise Claude Code on Amazon Bedrock and Google Cloud, with SSO (OIDC), central policy enforcement, per-user spend limits, and OTLP telemetry.
xAI added 21 multilingual voices to Grok Voice on July 6 (25+ languages) and launched a no-code Voice Agent Builder for production voice agents with built-in telephony, MCP connectors, and observability.
Willison responds to Geoffrey Litt's AIE conference talk: for developers to remain active creative participants when working with AI agents, they must maintain sufficient code understanding — otherwise cognitive debt accumulates.
Willison used DSPy to systematically evaluate system prompts for Datasette Agent and found a concrete bug: schema listings showing only table names (without column names) force the model to guess columns and enter retry loops.
The US Department of Commerce lifted an 18-day export-control restriction on Anthropic's flagship Claude Fable 5 and Mythos 5 models, restoring full global commercial access as of July 1, 2026.
Sam Altman is in early talks with the Trump administration to grant the US government a 5% stake in OpenAI, valued at roughly $42.6 billion, as a way to share AI upside with the public.
Anthropic launched Claude Sonnet 5, a new frontier-tier model for coding, reasoning, and agentic tasks at introductory rates of $2/$10 per million tokens through August 31, 2026 — roughly 60% cheaper than its top Fable model.
Nvidia is granting token compute credits to AI startups in exchange for a share of future sales, aiming to expand access to large-scale GPU infrastructure beyond well-funded companies.
Meta Platforms is building a cloud infrastructure business to monetize its vast AI compute surplus, competing directly with AWS, Microsoft Azure, and Google Cloud. Meta shares surged more than 9% on the news.
ByteDance has chosen Brazil's Ceará state for a $39 billion, 1 GW data center campus — its largest outside China — to serve TikTok's international operations, with first halls operational in late 2027.
New Bloomberg data shows accelerating payroll declines in information and financial-activities sectors averaging 28,000 jobs lost per month in 2026 — the clearest documented evidence yet of AI-driven white-collar displacement.
SoftBank has reopened talks for a $10 billion margin loan against its OpenAI stake, adding a self-repayment guarantee after lenders rejected an equity-only collateral structure for a private company.
Google expanded its Gemini Spark agentic AI assistant to macOS, adding real-time topic monitoring and Google Tasks and Keep integration alongside several new capabilities.
The July 19, 2026 Claude Code update changes /verify and /code-review skill behavior — Claude no longer triggers them autonomously, only on direct invocation. Increases agentic pipeline predictability.
Anthropic launched Claude Science, a specialized AI workbench integrating 60+ scientific databases and toolkits for genomics, proteomics, and cheminformatics, with auditable artifacts and a $30K grant program.
Square launched integrations with ChatGPT and Claude allowing consumers to discover restaurants and place orders directly inside AI chat interfaces, with Square processing transactions at low fees and no marketplace commission.
Google released Gemini Omni Flash for conversational video creation/editing and Nano Banana 2 Lite (Gemini 3.1 Flash-Lite Image) for enterprise-grade 4-second image generation at $0.034 per 1,000 images, available via Gemini API and Google AI Studio.
Palantir CEO Alex Karp publicly criticized the token-pricing model used by OpenAI and Anthropic as fundamentally broken, saying enterprise CEOs are frustrated by skyrocketing AI costs and demand for closed frontier models could erode.
Snorkel AI launched Senior SWE-Bench with 156 real pull request tasks requiring agents to diagnose bugs from runtime logs and match codebase conventions — best model Claude Opus 4.8 scores just 24%.
Alibaba Cloud's HappyHorse 1.1 AI video generation model climbed to the No. 2 position in global benchmarks, overtaking both OpenAI's Sora and ByteDance's Seedance with full API access via Alibaba Cloud Model Studio.
Weave Robotics unveiled Isaac 1, a fully mobile $7,999 home robot that autonomously tidies rooms, picks up dirty laundry, and handles loaded hampers, with California deliveries planned for fall 2026.
Privacy-focused AI platform Venice AI raised a $65 million Series A at a $1 billion valuation — its first external fundraise — after organically growing to over 3 million active users offering access to 200+ AI models with zero data retention.
GitHub Copilot added Kimi K2.7 Code from Moonshot AI as the first open-weight model in the model picker, rolling out to Pro, Pro+, and Max plans with Business and Enterprise expansion planned.
UK utility National Grid is acquiring a 35% stake in Joulent LLC, developing a 2.67 GW natural gas plant in West Texas with Chevron to supply a Microsoft data center under a 20-year power purchase agreement.
Cursor updated CursorBench to v3.1 focused on codebase understanding, bug-finding, and code review across 36 models. Fable 5 Max leads at 72.9%, while Composer 2.5 offers the best value at $0.55 per task.
SpaceX showed investors a prototype for a handset-like AI device slimmer than an iPhone, running a proprietary OS, using a Qualcomm Snapdragon chip, and integrating xAI technology — at an early stage with no commercialization guarantee.
OpenAI opened a public beta for ChatGPT Sites — a feature that allows creating and publishing websites directly from a ChatGPT conversation.
Copilot Agent is now available inside JetBrains AI Assistant, expanding agentic coding to millions of IntelliJ-based IDE users. Enterprise admins can now set per-user AI credit budgets for cost centers. Claude Sonnet 5 is also GA in Copilot.
Willison analyzed Claude Sonnet 5 and flagged that the new tokenizer produces ~30% more tokens for English text, significantly reducing the advertised pricing advantage over older models.
Willison released shot-scraper 1.10 with a `video` command letting AI coding agents automatically record video demonstrations of their own work via Playwright YAML storyboard — the feature was written entirely by GPT-5.5 via Codex Desktop.
The Bank for International Settlements named an AI bust among the three biggest threats to global prosperity, warning that hyperscalers' >$1T capex in 2025-2026 exceeds earnings and free cash flow, with 'circular financing' deals amplifying systemic risk.
CoreWeave launched ARIA, an autonomous AI research agent built into Weights & Biases that can analyze experiments, form hypotheses and run continuous model improvement loops with minimal human oversight.
Bloomberg finds animators using generative AI to cut animated film production costs by up to 90%. Jeffrey Katzenberg says a world-class animated movie that once needed 500 artists over five years 'won't take 10% of that' going forward.
Proception, founded by a former Tesla Optimus technical lead, settled a year-old trade secret lawsuit with Tesla and simultaneously announced an $11 million seed round led by First Round Capital.
Bloomberg's Catherine Thorbecke argues India is the world's #2 market for ChatGPT and Claude, but with ~24 official languages and 100+ dialects, AI risks becoming yet another technology dividing the English-speaking elite from the rest.
A class action filed in the Northern District of California accuses the three biggest memory makers of secretly coordinating to cut DRAM production and shift capacity to HBM for AI accelerators, driving consumer memory prices up roughly 700% in four years.
China, India and Hong Kong are the only major markets where top firms now hold a smaller share of total market cap than a year ago, while Korea (65%) and Taiwan (56%) saw concentration surge thanks to AI champions.
Chinese super-app Meituan published technical details and weights for LongCat-2.0, a 1.6-trillion-parameter mixture-of-experts model with about 48B activated per token and a 1-million-token context, trained almost entirely on domestic Chinese chips.
China's State Council issued a five-year blueprint making AI a core capability for every student, integrating it across all schooling levels as part of Xi's push to dominate advanced technologies.
Anthropic is running a live broadcast showcasing how Claude is being deployed across pharma, biotech and academic research, with product demos and customer case studies from leading life-sciences institutions.
Cursor released its native iOS app for iPhone and iPad into public beta for all paid plans, letting users spawn and manage cloud agents, remote-control a desktop session, and code by voice from mobile.
The vLLM project published a technical post arguing that small open models, bundled into bounded "micro-agent" runtimes inside the serving layer, can match or beat frontier closed models on real tasks at a fraction of the cost.
Anthropic raised Claude API rate limits across the board — Sonnet and Haiku now match Opus at every tier — and collapsed usage tiers into three (Start, Build, Scale), powered by a new SpaceX compute deal.
Google has restructured its AI coding strike team, adding a new "midtraining" phase and pulling Sergey Brin and DeepMind CTO Koray Kavukcuoglu into direct oversight after losing six senior researchers in five months — most recently Denny Zhou to Meta.
A viral analysis on Hacker News argues that Iceland, with abundant cheap geothermal and hydro power and a cold climate, should be Europe's natural home for AI data centers — but bandwidth, latency, talent and policy frictions keep hyperscalers building in Ireland, Germany and Spain instead.
GitHub Copilot now offers Claude Opus 4.8 in a new fast-mode preview that delivers significantly faster output token speeds while keeping the same intelligence as standard Opus. The variant is rolling out to Copilot users in preview, aimed at speeding up agentic and long-form coding tasks.
