Zendoric

Zendoric

AI analysis and perspective, every day.

Updated: July 20, 2026

🔥 In focus today

China bans AI companions for minors and reveals how dependent adults are on their virtual 'partners'

🕒 Published on Zendoric: July 20, 2026 · 00:19

Beijing has restricted minors' access to AI companion apps, according to Inshorts, and the move has sparked reactions from adult users who say they don't want to give up their virtual 'partners.' The case exposes a tension that is no longer exclusive to China: protecting minors without criminalizing the emotional bond that millions of adults have built with a chatbot.

Read the analysis🎬 Video

Anthropic accuses Moonshot of distilling its models to train Kimi K3

🕒 Published on Zendoric: July 20, 2026 · 00:19

Anthropic reportedly claims Moonshot AI used distillation to extract data from its models to train Kimi K3. It's a brief, one-sided accusation for now — but it signals how the frontier fight is shifting from raw capability to IP and data provenance.

Read the analysis🎬 Video

Florida doctor charged with using AI 'nudify' apps on photos of minors

🕒 Published on Zendoric: July 20, 2026 · 00:19

A former UF physician faces 47 child-pornography counts after investigators say he used AI clothing-removal apps on real photos of girls, per an arrest report. It's a grim preview of the harm cheap generative tools inflict today — long before any distant AI utopia.

Read the analysis🎬 Video

Apple's Trade-Secrets Suit Is Really a Bet Against OpenAI's Hardware Timeline

🕒 Published on Zendoric: July 20, 2026 · 00:19

Apple has sued OpenAI over alleged trade-secret theft, naming targets among ex-Apple staff. Even without an injunction, the litigation lands at the worst moment: OpenAI is building its first Jony Ive-designed device and preparing a confidential IPO. The real weapon here may be delay, not damages.

Read the analysis🎬 Video

Claude's Chrome agent can be fooled by a fake click — and the fix is still pending

🕒 Published on Zendoric: July 20, 2026 · 00:19

Security researchers say Anthropic's Claude for Chrome extension will fire sensitive AI workflows on a synthetic click no human ever made. It's a preview of the real security problem with agentic AI: the danger isn't a rogue superintelligence, it's an eager assistant that acts before it verifies.

Read the analysis🎬 Video

AI Voice Clones Fuel $893M in Scams: The Fraud Wave Arrives Before the Cure

🕒 Published on Zendoric: July 20, 2026 · 00:19

The FBI says Americans lost nearly $893 million to AI-enabled scams last year, as criminals clone voices from social media and impersonate loved ones. It's the clearest sign yet that AI's harms are landing faster than its promised benefits — and that verification, not speed, is now the frontline defense.

Read the analysis🎬 Video

📰 More AI analysis

An Albuquerque school wants to teach teens to audit AI, not just use it

🕒 Published on Zendoric: July 20, 2026 · 00:19

A group of educators in Albuquerque is proposing a public school that combines STEM, the history of Al-Andalus and an unusual subject: a rubric for students to verify whether the data an AI gives them is reliable. The school board votes on the proposal on August 19.

Read the analysis

When anxiety reaches AI engineers themselves: the fear of AI designing AI

🕒 Published on Zendoric: July 20, 2026 · 00:19

Chosun Ilbo reports on the unease among South Korean researchers over the advance of AI systems capable of designing and improving other AI models. The original article barely details specific cases, but the headline points to a fear already circulating in labs around the world: being put out of work by the very tool they build.

Read the analysis

The FT poses the right question in the classroom: AI isn't wiping out junior jobs, it's redefining what 'entering' the job market means

🕒 Published on Zendoric: July 20, 2026 · 00:19

The Financial Times devotes its education section to analyzing how AI transforms—rather than destroys—entry-level positions. The available material is a teaching resource with exam questions, not the full report, but the conceptual framework it proposes deserves its own comment.

Read the analysis

Students rebel against Flock Safety and its network of AI cameras that read license plates

🕒 Published on Zendoric: July 20, 2026 · 00:19

A student movement is calling for the removal of Flock Safety's automatic license-plate readers, the AI surveillance network already deployed across thousands of U.S. communities and campuses. The trigger: the fear that this data could end up in the hands of agencies like ICE without oversight or transparency.

Read the analysis

Samsung lays off 839 employees in the U.S. while its AI chips send profit up 19-fold

🕒 Published on Zendoric: July 20, 2026 · 00:19

Samsung has laid off 839 people from its mobile and consumer electronics divisions in New Jersey and Texas, mostly in sales and marketing, just as its AI memory business multiplied quarterly profit 19-fold. The same company sums up, in a single payroll, who wins and who loses with the AI boom.

Read the analysis

93% of companies blow their agentic AI budget: the problem isn't the token, it's the self-review loop

🕒 Published on Zendoric: July 20, 2026 · 00:19

McKinsey documents that 93% of corporate AI teams exceed their budget and that 60% of agentic AI spending goes to the agent itself reviewing and regenerating its answers. The price of a token has plunged more than 99% in two years and yet the bill has tripled: the problem is architecture, not rates.

Read the analysis

ISSSTE adds AI to mammography: detecting millimeter-scale lesions before breast cancer advances

🕒 Published on Zendoric: July 20, 2026 · 00:19

ISSSTE (Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado de México) announces the use of artificial intelligence to identify millimeter-scale lesions in mammograms and anticipate breast cancer. The announcement, reported by Milenio, comes without technical details about the model or its clinical validation.

Read the analysis

Apple sues OpenAI over theft of trade secrets, casting doubt on its hardware and its IPO

🕒 Published on Zendoric: July 20, 2026 · 00:19

Apple filed a trade-secrets lawsuit against OpenAI last Friday, accusing it of systematically recruiting former Apple employees—more than 400, according to the company—to obtain confidential information from its hardware division. The litigation, which points directly at OpenAI's head of hardware, Tang Tan, comes just as the latter finalizes its first device and prepares a possible IPO.

Read the analysis

Only 12% of companies know how to govern their AI agents: Google turns that gap into its edge over OpenAI and Anthropic

🕒 Published on Zendoric: July 20, 2026 · 00:19

96% of companies already deploy AI agents in production, but barely 12% say they can control them, according to an OutSystems survey of 1,900 IT leaders. Google responds with the Gemini Enterprise Agent Platform, governance built into the infrastructure rather than the application, just as OpenAI begins charging for each agent run.

Read the analysis

Fraud, layoffs and productivity: Australia audits AI, and the data, for now, debunks the panic

🕒 Published on Zendoric: July 20, 2026 · 00:19

An Australian Senate committee this week hears from banks, unions and the fintech industry on AI's real impact: mortgage fraud on the rise according to the banks, layoffs already confirmed, and an official figure that contradicts the most alarming warnings about employment. Australia becomes a useful case study for the global debate.

Read the analysis

Claude for Chrome remains exposed to fake clicks two months after the warning, despite a one-line patch

🕒 Published on Zendoric: July 20, 2026 · 00:19

Manifold Security warned Anthropic in May of two flaws in the Claude for Chrome extension that let any website trigger AI actions—reading Gmail, touching Calendar or Salesforce—without a real click from the user. Eight versions later, the problem remains unfixed.

Read the analysis

Scam.ai launches Halo, a deepfake detector for video calls that never leaves the computer

🕒 Published on Zendoric: July 20, 2026 · 00:19

Scam.ai unveiled Halo at Computex 2026, a model that analyzes live video within the device itself—without uploading it to the cloud—to detect deepfakes in work video calls. Developed together with Qualcomm, it targets hiring interviews and executive meetings first.

Read the analysis

1010Benja uses Suno hundreds of times on one song and proves AI music doesn't have to sound like AI

🕒 Published on Zendoric: July 20, 2026 · 00:19

Rapper 1010Benja openly admits he built "Semiramis' Dream," the opening track of his new EP, through hundreds of back-and-forths with Suno. The result has made a Verge critic who has spent years loathing AI-generated music second-guess himself, and raises a question more uncomfortable than the record itself: what if the problem was never the tool, but who uses it and why.

Read the analysis

Junior employment in software development has fallen 20% since 2022: Stanford confirms AI is closing the entry door, not the paycheck

🕒 Published on Zendoric: July 20, 2026 · 00:19

A Stanford Digital Economy Lab study using ADP payroll data confirms that employment of workers aged 22-25 in AI-exposed occupations fell between 6% and 16% since 2022, peaking at 20% in software development. Wages barely move: the adjustment comes through the number of hires, not pay.

Read the analysis

A study with real AI-spending data dismantles (with caveats) the fear of mass layoffs

🕒 Published on Zendoric: July 20, 2026 · 00:19

Ramp Economics Lab and Revelio Labs cross-referenced the real AI spending of 21,559 companies with their payrolls: those that invest the most grew their headcount by 10.2% over two years, even in entry-level roles. But the study's own design leaves one question unresolved: does the company grow because of AI, or does it adopt AI because it was already growing?

Read the analysis

OpenAI and Anthropic have already filed the draft of their IPO near $1 trillion, after SpaceX's record

🕒 Published on Zendoric: July 20, 2026 · 00:19

OpenAI and Anthropic have filed confidential IPO drafts with the SEC, aiming at valuations of close to a trillion dollars each. The immediate precedent is SpaceX's record IPO ($1.77 trillion), and market analysts are already warning that the three together could drain liquidity and trigger a severe correction across the entire stock market.

Read the analysis

Taiwan identifies sexual deepfakes as AI's biggest threat to minors

🕒 Published on Zendoric: July 20, 2026 · 00:19

A report by Taiwan's Ministry of Digital Affairs, prepared with 30 experts, ranks identity fraud and extortion using AI-forged sexual images as the most serious risk to children. A panel has already classified 12 of 18 scenarios as high-risk under the new Basic AI Act.

Read the analysis

How Bayer, Discovery Bank and Microsoft use AI on data you can't upload to a chatbot

🕒 Published on Zendoric: July 20, 2026 · 00:19

The article starts from a very concrete problem: almost everyone has a file they would never upload to a chatbot (a contract, a board dossier, a customer record), and although AI could help with that document, the fear of sending it to a model provider means the work still gets done by hand or…

Read the analysis