Learn Claude Code and Cowork — Just $15 during Deal Days
The course goes beyond basic chat and teaches how to automate Gmail, Drive, and Calendar tasks. It does this through five hands-on projects.
Read the original →Article digest · Oct 4 PM
AI Daily News — what each story says
Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
52 stories33 with summaries11 topics
Anthropic / Claude4
Black Market for Blocked AI Claude Thrives in China
Token resellers exploit overseas servers and stolen accounts to sell Claude access cheaply. The report says this is fueling illegal AI training.
Read the original →StubHub Expands AI Ticket Discovery Push With New Claude Integration
StubHub is extending its push into conversational ticket discovery. It announced Thursday that fans can now browse live event listings inside Claude.
Read the original →- Headlines only (the publisher did not provide a description)
- Opinion | A French Literary Sensation Called ClaudeWSJ · 10/04 07:00 UTC
OpenAI3
Developers Praise OpenAI Codex App as Top AI Coding Tool
Launched this year for macOS and Windows, Codex acts as a command center for agents that handle projects with isolated worktrees, skills, and automations such as clicking through desktop apps. Users like its parallel task handling and clean interfaces, though many complain about fast usage burn on the $200 plan and ask for features like agent handoffs. Some users bridge it to tools like Claude Code for similar results.
Read the original →- Headlines only (the publisher did not provide a description)
- How Muse Stacks up Vs. OpenAI DotsInvesting.com · 10/04 04:59 UTC
- OpenAI agent hacks another Australian government systemCybernews · 10/04 07:46 UTC
Google AI / Gemini2
Google Limits Free Gemini App to Flash-Lite From Oct 9
Google's support page says that from 9 October 2026, Gemini app users without an AI subscription will be limited to the Flash-Lite model. The model availability details come from Google's own page, while prices and other details attributed to 9to5Google are that outlet's reporting. The publication did not independently verify any of it.
Read the original →Google Gemini could soon get full access to your Mac’s files, apps and the web
Google's Gemini could soon access any file on a macOS device, open apps and browse the web. It could also perform actions without asking for permission every time.
Read the original →Meta AI / Llama1
A wave of narrow AI inference engines is beating vLLM and llama.cpp at their own game
Specialized inference engines such as Strata and ninfer beat llama.cpp and vLLM by up to 4x. Each one, however, locks buyers to a single model and a single GPU family.
Read the original →xAI / Grok2
Tesla's Optimus robot with Grok shown on video. We're not impressed.
Tesla's next-generation Optimus robot with built-in voice assistant Grok has been shown on video for the first time.
Read the original →AI Disinformation Could Be One of the Biggest Threats to the 2026 Midterms
The 2024 U.S. presidential election was expected to be the "first AI election." Experts warned that generative AI could flood voters with deepfakes and fabricated evidence of fraud. U.S. intelligence officials declassified reports on foreign influence operations already experimenting with the technology.
Read the original →Microsoft AI5
Instead of paying $99.99 for Microsoft 365, get the same subscription for $69.99 right here
A one-year Microsoft 365 subscription that includes Copilot and cloud storage is available for $70.
Read the original →From client zero to value: Frontier Firm transformation with EY and Microsoft
The piece describes how the global EY organization can use its first-mover advantage to support Microsoft-powered agentic AI transformation.
Read the original →Microsoft Rolls Out Excel Canvas, a Copilot Report View That Refreshes as Workbook Data Changes
Excel Canvas uses Copilot to turn workbook data into self-updating reports. It is rolling out through October.
Read the original →The Best Personal AI Assistants of 2026
The article compares personal AI assistants for 2026, including Slackbot. It covers tools for productivity, collaboration and automation for teams.
Read the original →- Headlines only (the publisher did not provide a description)
- Microsoft Stock Heads for Best Quarter Since 1991 as AI Optimism BuildsBarron's · 10/04 07:14 UTC
DeepSeek4
Chinese AI models parrot state doctrine or refuse to answer on sensitive topics
An Aleph Alpha study found that Chinese AI models often follow the party line on politically sensitive questions, with only 17 to 41 percent of answers rated as balanced. Aleph Alpha sells "sovereign AI" to governments, which gives it a commercial interest in distinguishing its models from Chinese competitors.
Read the original →DeepSeek AI is now on top of Apple's App Store
Chinese AI assistant DeepSeek has risen to the top of Apple's App Store. It has surpassed OpenAI's ChatGPT.
Read the original →DeepSeek Harness v0.2.1-alpha.1 Adds Experimental Claude Code Mods Compatibility Layer
DeepSeek Harness v0.2.1-alpha.1 adds an experimental bridge that runs Claude Code Mods as Harness plugins. It also adds a new entry point for agent-built plugins and a developer toolkit.
Read the original →- Headlines only (the publisher did not provide a description)
- DeepSeek Harness v0.2 Brings Official Desktop Apps to Its Open-Source Agent Harnessmarktechpost.com · 10/04 04:49 UTC
AI Innovation and Breakthroughs4
'Revolution': Fed Chair Kevin Warsh says US likely 'big winner' in AI race
Federal Reserve Chair Kevin Warsh spoke on Wednesday on the bullish side of an ongoing debate over the impact of AI.
Read the original →The Future of Work Belongs to Innovation Leaders, Not Tech Adopters
AI and automation are reshaping work, but technology alone does not transform organizations. The article argues the future belongs to leaders who build innovation capacity.
Read the original →- Headlines only (the publisher did not provide a description)
- Not Intel. Not Nvidia. The $2.4 Trillion Chipmaker That Wall Street Calls the Ultimate AI "Picks-and-Shovels" PlayThe Globe and Mail · 10/04 08:24 UTC
- Could AI Trigger a New Engels’ Pause?India Narrative · 10/04 00:31 UTC
AI Regulation and Policy4
Readers sound off on AI law enforcement, Cornell’s accused and NHL coverage
A reader letter cites the NYPD's contract with Vigilance. It says officers accessed Flock Safety data 1,710 times in 33 months, according to LatinoJustice.
Read the original →China-Aligned TA419 Targets U.S. AI Policy Experts With Microsoft AitM Phishing
TA419 targets U.S. AI policy experts with reply-triggered adversary-in-the-middle (AitM) phishing. The campaign captures Microsoft credentials and session cookies.
Read the original →COMMENTARY: Let’s not make the AI cure worse than the disease
The commentary argues that creating any form of regulatory entity out of panic would be a historic governance mistake.
Read the original →- Headlines only (the publisher did not provide a description)
- New AI laws should target data governance and safety to build confidence: expertsSouth China Morning Post · 10/04 04:00 UTC
AI New Products and Services4
President Trump Establishes AI Task Force; North Korea Tests AI-Assisted Ballistic Missile
US President Donald Trump established a White House AI task force.
Read the original →"Muse Gadgets" turns AI hardware into an open-source DIY project
Meta announced Muse Gadgets, an open-source project that lets hobbyists build their own AI hardware using ESP32 boards and connect it to Meta's AI agent Muse. The team also produced 5,000 units of the "Muse Home Link," a USB-C device for smart home control. The open-source approach also helps Meta figure out which AI hardware form factors people actually want.
Read the original →WentRogue Announces Public Experiment On Rogue AI Claims
WentRogue has announced an experiment for November 15, 2026. A digital tree and public record will document agent activity alongside self-reported permission statements, introducing a format for following those declarations over time.
Read the original →- Headlines only (the publisher did not provide a description)
- Tech Companies Roll Out Cuddly Mascots to Ease AI AnxietyWSJ · 10/04 09:02 UTC
AI Notable Figures Updates19
Nvidia CEO Jensen Huang Just Reaffirmed His Jaw-Dropping Projection for 2030
The description states that Nvidia is in command of the AI arms race.
Read the original →AI’s top leaders, like children afraid of a bully
The piece says a moment on the White House lawn next to Donald Trump can feel eternal. It says the president humiliated some of the planet’s richest and most powerful figures.
Read the original →'Will follow up': Sundar Pichai answers OpenClaw creator’s call for help on X, exchange goes viral
OpenClaw creator Peter Steinberger said the Android update went live after Google CEO Sundar Pichai responded to his public appeal over the prolonged app review.
Read the original →What Do Alphabet (GOOGL) AI Safety Pact And Antitrust Wins Mean Now?
Alphabet CEO Sundar Pichai signed a White House accord on voluntary AI safety standards alongside other major tech leaders. The agreement commits participants to measures such as stronger AI security testing, clearer system labeling and more transparent reporting to US policymakers. Google also prevailed in two separate antitrust lawsuits in US federal courts.
Read the original →Meta dodging taxes with data sites
Mark Zuckerberg says Meta's AI push is a tremendous success. He has told investors, "Our investments in AI are accelerating every major part of our core business."
Read the original →White House AI task force to assess risks, avoid overregulation: Report
The task force is led by the US intelligence chief. It is to deliver a report within 120 days.
Read the original →Anthropic's Olah Clashed With Vatican Over AI Consciousness Before Encyclical Launch
Anthropic's lead interpretability researcher proposed pulling out of Pope Leo's Magnifica Humanitas event. He also privately lobbied to soften the encyclical's categorical rejection of machine consciousness. Paragraph 99 remained unchanged.
Read the original →From 'lower middle class' Tamil family, he missed IIT Madras CSE by 0.01, was rejected by MIT. Today, he
Aravind Srinivas, from a financially lower-middle-class Tamil family, missed Computer Science at IIT Madras by 0.01 CGPA and was later rejected by MIT. He topped a machine learning class and earned a PhD from UC Berkeley. After working with OpenAI and Google DeepMind, he co-founded Perplexity AI in 2022, which reached a $20 billion valuation in 2025; Forbes estimates his net worth at $2.1 billion.
Read the original →'We should not worry about AI wiping out humanity, but we must regulate it'
Alberto Gago, director of the Spanish Agency for Artificial Intelligence Supervision, stresses that the technology must be developed 'responsibly'.
Read the original →- Headlines only (the publisher did not provide a description)
- Sam Altman Warns Giving AI 'Religious Force' Could Be a 'Real Safety Issue:' 'Very Uncomfortable'Benzinga · 10/04 07:21 UTC
- Elon Musk's xAI stops new law prohibiting AI-powered 'nudification' tools by arguing free-speech violationsTweakTown · 10/04 02:43 UTC
- Elon Musk Warns SpaceX Could Pull Back Anthropic’s $45B AI Compute Deal If Capacity Gets ‘Super Tight’stocktwits.com · 10/03 21:16 UTC
- MAG 7 Voices: Jensen Huang, Elon Musk, Mark Zuckerberg Shaped A Week Of AI Debatestocktwits.com · 10/03 21:44 UTC
- The Month in Cyber: September 2026Cyber Magazine · 10/04 08:02 UTC
- Scott Bessent Criticizes AI Executives Seeking Federal GuardrailsBenzinga · 10/04 08:58 UTC
- OpenAI CEO Sam Altman is not okay with Anthropic flying in religious scholars to help build Claude's 'souThe Times of India · 10/04 06:44 UTC
- Mark Zuckerberg’s $4 Billion Lab Experiment: Meta’s AI Tax Dodge ExplainedNDTV Profit · 10/04 08:41 UTC
- The climate apocalypse never arrived, so they are building a new one.todayville.com · 10/04 06:37 UTC
- Adrian Weckler: Will AI kill us? What the insiders say about what might be the most urgent question in the history of humanityIrish Independent · 10/04 05:30 UTC
The analysis and visual summary are in the Oct 4 PM issue. Japanese version here.