Saturday’s theme was the bill. Anthropic, per one report, made Claude Fable 5 permanent for Max and Team Premium subscribers at reduced limits, while Pro users get a one-time usage credit and then API pricing. France’s competition authority put a number on the market those prices live in: OpenAI, Google, and Anthropic hold 84% of the AI agent market, and the regulator wants data-access and interoperability rules before lock-in hardens. Investors keep paying in advance regardless: inference-chip startup Etched is running back-to-back rounds at $10 billion and $20 billion valuations, backed by a reported $1 billion in customer demand for a chip that hasn’t shipped. And McKinsey supplied the quiet stat of the day: as enterprises scale agents, nearly 60% of operating costs go to verification — paying people and systems to check the machines’ work.
A US judge declined to block Meta from laying off 26 workers who allege the company’s AI tools discriminated against employees with disabilities or on medical leave — the discrimination suit continues, but the layoffs proceed. The year’s broader tally, by one count: more than 165,000 people laid off across nearly 40 companies with AI adoption cited as the primary driver. A separate analysis describes career-ladder compression, with junior roles splitting into seniorized tracks and declining ones. And a workplace-consent gap got measured: only 35% of workers say they are always asked before AI transcribes their meetings.
Governance did not take the weekend off. Brazil’s electoral court added OpenAI, Anthropic, and ElevenLabs to its election disinformation pact ahead of the 2026 vote, extending deepfake rules from platforms to the model vendors themselves. One analysis reads the WAIC chair’s statement as a Chinese bid to write global AI governance rules. The scam economy moved faster than the policy: fraudsters are impersonating the White House Gold Eagle security initiative to harvest credentials from financial institutions, and a post-mortem examines the JadePuffer ransomware and its autonomy claims alongside the industry pushback it provoked. The EU’s Article 50 transparency duties still arrive on August 2, omnibus delay or not.
What People Are Asking
Is Claude Fable 5 staying available?
Per one report, yes for some tiers — Anthropic made Fable 5 permanent for Max and Team Premium subscribers at reduced limits, while Pro users get a one-time usage credit before being pointed to API pricing. Tech Times →
Did Meta win the AI discrimination lawsuit?
No — but a US judge declined to block Meta from laying off the 26 workers who brought it, so the layoffs proceed while the discrimination claims continue. The Hindu →
How concentrated is the AI agent market?
France's competition authority says OpenAI, Google, and Anthropic control 84% of it, and is urging regulatory action on data access and interoperability to prevent lock-in. PPC Land →
What is Etched worth now?
The inference-chip startup is pursuing two concurrent rounds at $10 billion and $20 billion valuations, backed by a reported $1 billion in customer demand for a product that has not yet shipped. MarketScale →
Is AI killing entry-level jobs?
2026 data shows career-ladder compression — junior roles splitting into seniorized and declining categories — and a separate count attributes layoffs of more than 165,000 people across nearly 40 companies primarily to AI adoption. Memeburn →
What is the JadePuffer ransomware?
A ransomware incident exploiting CVE-2025-3248, notable for claims about the malware's autonomy — an analysis examines those claims, the industry pushback, and the regulatory implications. Tech Insider →
Is the White House Gold Eagle program a scam?
The initiative itself is legitimate — but one analysis documents scammers impersonating it to run credential-harvesting attacks against financial institutions. Unbox Future →
What does Brazil's election AI pact require?
Brazil's Superior Electoral Court added OpenAI, Anthropic, and ElevenLabs to its election disinformation pact, enforcing strict rules against deepfakes and synthetic content ahead of the 2026 elections. The Rio Times →
Every Story From July 18
Jump to: AI labs and models (17) · AI infrastructure (12) · Policy and regulation (8) · Labor and workplace (6) · Security and agent risk (4) · Markets and macro (3) · Science and research (2) · Society, politics, and ethics (2) · Consumer tech and culture (1)
AI labs and models17 stories
A roundup of recent AI model releases including Inkling, Kim K3, GPT-5 Turbo, Command R+ 2.0, Gen-4 Turbo, and SD4 Turbo.
Apple Intelligence receives regulatory approval in China through partnerships with Alibaba Qwen and Baidu, marking a key competitive shift.
WAIC 2026上，多家企业推出AI Agent产品，旨在将AI从个人提效工具升级为企业组织能力，竞争焦点转向上下文理解和任务完成率。
Anthropic makes Claude Fable 5 permanent for Max and Team Premium subscribers at reduced limits, offers Pro users only a one-time usage credit before requiring API pricing.
A detailed review of SpaceXAI's Grok AI assistant covering features, pricing, and model lineup in 2026.
A daily digest summarizing analysis on AI company structures, adoption phases, open-weight costs, proprietary data moats, agent infrastructure, regulation proposals, and physical AI stack mapping.
McKinsey report highlights that as enterprises scale agentic AI, financial sustainability and ROI are becoming the main challenge, with nearly 60% of operating costs spent on verification.
A roundup covering Apple's lawsuit against OpenAI for trade secret theft, EU ordering Google to open Android and Search, and OpenAI's new macro pad.
Microsoft invests $2.5B in embedded AI deployment, while survey reveals gap between software ownership and integration in manufacturing and B2B AI search visibility.
Billionaire Chamath Palihapitiya argues that US AI labs cannot complain about Chinese distillation when they themselves built models by scraping copyrighted data without permission.
OpenAI released multiple ChatGPT updates in July 2026, including a revamped desktop layout, increased custom instructions to 5,000 characters, unified search, WhatsApp return in EEA, and disabling new group chats.
A collection of news and analysis articles from The Guardian covering AI developments, including industry competition, ethical concerns, policy debates, and societal impacts.
LM Market Cap provides a live dashboard tracking recent AI model releases, provider activity, and pricing.
DemandSphere's AI Frontier Model Tracker provides benchmarks, pricing, and capabilities for major frontier AI models, updated regularly.
This documentation provides a detailed guide on running Alibaba's Qwen3.5 large language models locally using Unsloth, including hardware requirements, quantization updates, and usage instructions.
A detailed changelog for Claude Code, documenting version history, features, fixes, and breaking changes.
OpenAI published release notes covering bug fixes for Codex GPT-5.6 models and desktop app experience updates for ChatGPT business, enterprise, and consumer plans.
AI infrastructure12 stories
AI inference chip startup Etched is pursuing two concurrent funding rounds at $10B and $20B valuations, backed by $1B in customer demand for its yet-to-ship product.
Microsoft's September 2026 target for Teams Interactive Agents is contradicted by its own roadmap's admission of rollout inability, leaving IT administrators with uncertainty.
Bloom Energy's stock has fallen 39% from its peak as regulatory challenges, including rejection of Oracle's Project Jupiter and New York's data center moratorium, threaten its AI-driven growth story.
Architects are exploring designs where AI data centers can give back to host communities, e.g., using waste heat for public bathhouses, as opposition to new facilities grows.
The article examines how AI is transforming customer service in India by handling routine queries, while shifting human agents to more complex and empathetic tasks.
Feathery, an AI operating and decisioning system for financial services, raised $30M in total funding including a Series A.
Thira, an AI startup founded by Apptio co-founders, raised a $21M seed led by Madrona to develop AI agents for back-office tasks like IT support.
Indian startups raised $346.2 million in a week, led by AI company Emergent's $130M round, signaling a shift toward concentrated late-stage investments and AI-focused venture capital.
BenchLM publishes a profile of Nemotron 3 Ultra with 38 of 320 benchmarks tracked, showing mixed evidence and ranking unranked overall.
The 2026 World Internet Conference forum will explore new opportunities in digital and intelligent health development.
A detailed reference page listing the fee structure, admission dates, eligibility criteria, and scholarship options for the B.Ed programme at Amity University Noida.
This is a topic hub listing various AI and science-related articles from South China Morning Post.
Policy and regulation8 stories
France's competition authority warns that OpenAI, Google, and Anthropic control 84% of the AI agent market, urging regulatory action on data access and interoperability to prevent lock-in.
Brazil's Superior Electoral Court has expanded its election disinformation pact to include AI vendors OpenAI, Anthropic, and ElevenLabs, alongside social platforms, enforcing strict AI rules against deepfakes and synthetic content ahead of the 2026 elections.
The 2026 WAIC chair's statement signals China's ambition to shape global AI governance rules, highlighting the growing geopolitical struggle over who defines acceptable AI behavior across borders.
Bar Council of India releases comprehensive circular regulating social media conduct, digital ethics, and professional responsibility for advocates, law students, and interns.
Despite the delay of high-risk AI obligations to 2027-2028, EU AI Act transparency duties under Article 50 and GPAI enforcement powers take effect on August 2, 2026, requiring immediate compliance for chatbots, deepfakes, and AI-generated content.
As attorneys face sanctions for AI misuse, judges are both setting guidelines and cautiously adopting AI tools, balancing innovation with preserving judicial integrity.
Commentary argues China is considering export controls on its advanced AI models as national security assets, highlighting risks of dependency on Chinese AI and recommending US countermeasures.
Tecnica della Scuola offers paid courses to help Italian schools meet AI Act and MIM guidelines for staff AI literacy training.
Labor and workplace6 stories
Microsoft's Version 2606 update makes SAEC devices Copilot-ready, shifting IT's focus from migration to licensing, data governance, and operational readiness.
Analysis of 2026 data reveals career ladder compression as AI disrupts entry-level roles, with junior jobs splitting into seniorised and declining categories.
US judge refuses to block Meta from laying off workers who sued over alleged AI-based discrimination in job cuts.
Also reported by: thehindu.com
Nearly 40 major companies have laid off over 165,000 employees in 2026, with AI adoption cited as the primary reason for workforce reductions across tech and other sectors.
Only 35% of workers always give consent before AI meeting transcription, highlighting legal and policy gaps for Teams admins.
As top creators evolve into studio-scale operations with employees and production infrastructure, brands must update their vetting and contracting processes to address new compliance risks.
Security and agent risk4 stories
An analysis of the JadePuffer ransomware incident exploiting CVE-2025-3248, discussing its autonomy claims, industry pushback, and regulatory implications.
Analysis of how the legitimate White House Gold Eagle AI cybersecurity initiative is being exploited by scammers targeting financial institutions with credential-harvesting attacks.
Guest commentary warns business leaders about the growing threat of AI-generated fake content that can rapidly damage reputations before the truth emerges.
A cybersecurity blog summarizes the week's major stories, including AI-discovered Linux vulnerability, Microsoft SharePoint flaws, and AI data center security challenges.
Markets and macro3 stories
The article analyzes how the AI chip stock sell-off signals a shift from 'infinite demand' to 'efficiency verification' in the semiconductor supply chain, affecting technology routes, competitive landscape, and regional investments.
European venture capital recovery is illusory; funding is rotating from speculative growth to strategically necessary industries and profitable companies.
The article summarizes a week of funding activities for Indian startups, including Udaan, Emergent, and others, with total amounts and investor participation.
Science and research2 stories
ScienceDaily aggregates the latest AI research news including energy-efficient AI, DNA robots, neuromorphic computing, and AI in healthcare.
A perspective article argues for using AI to shift chemical exposomics from mere detection to predicting which environmental exposures pose biological risks.
Society, politics, and ethics2 stories
Georgia Republican Gabriel Sterling expressed shock that Trump's election speech contained no new findings despite promises of shocking revelations.
Makale, Christopher Nolan'ın The Odyssey filmi üzerinden Hollywood'un yapay zeka karşısındaki dönüşümünü ve insan yaratıcılığının önemini analiz ediyor.
Consumer tech and culture1 story
The article argues that brands should invest in AR and XR ad experiences, which are becoming more accessible and offer higher engagement and conversion rates than traditional video ads.
About the RUDI Daily. Responsible Use of Digital Intelligence, daily. Compiled each day from same-day reporting across the web — every story links to its original publisher. How we build it →
Putting AI to work? RUDI helps organizations adopt AI responsibly — training, governance, and hands-on implementation. AI training · Consulting · Talk to us