Zendoric

Zendoric

AI analysis and perspective, every day.

Updated: July 8, 2026
🎙️ Listen to today's podcast

🔥 In focus today

Meta's Wireless Brain-to-Text System Signals the Next Interface War Isn't About Screens

🕒 Published on Zendoric: July 8, 2026 · 09:15

Meta says its Brain2Qwerty v2 can turn brain activity into text using external magnetic-field sensors, no implants or wires required. If it holds up outside the lab, it reframes the human-computer interface race as a hardware problem AI has quietly solved on the software side.

Microsoft Swapping OpenAI and Anthropic for In-House Models Is the Real Story Behind the 'Partnership'

🕒 Published on Zendoric: July 8, 2026 · 09:15

A Bloomberg report says Microsoft is replacing OpenAI and Anthropic with its own AI models in some of its applications. Details are thin, but the direction is the one to watch: even the biggest AI partners are quietly building their way toward independence.

📰 More AI analysis

Ohio requires its school districts to have an AI policy, but the video doesn't reveal the substance of the debate

🕒 Published on Zendoric: July 8, 2026 · 09:15

A local NBC4 report from Columbus covers how area school districts responded to a new Ohio state requirement on artificial intelligence in classrooms, although the available material does not detail the exact content of the mandate nor the districts' specific responses.

A coalition of 17 countries wants children to have a say in the AI that educates and monitors them

🕒 Published on Zendoric: July 8, 2026 · 09:15

The United Nations and 17 governments have launched a coalition in Geneva so that children stop being merely 'users to be protected' and become rights-holders in the design of AI. However, the signatures of the United States and China are missing, and there are no binding compliance mechanisms.

Anthropic hid an anti-China tracker in Claude Code: when ethics becomes a brand and the brand becomes a liability

🕒 Published on Zendoric: July 8, 2026 · 09:15

A security researcher discovered hidden code in Claude Code's system prompt that detected time zones and proxy usage to identify connections with Chinese labs. Anthropic confirmed it: it was a March experiment against unauthorized resale and distillation of its models that was never removed.

'Vibe coding' arrives at auto finance: coders with no coding experience program with AI in regulated finance

🕒 Published on Zendoric: July 8, 2026 · 09:15

An auto finance company, Arivo Acceptance, has assigned a team of interns with no programming experience to build software by describing what they need to Claude or Copilot. The case illustrates how agentic AI and 'vibe coding' are starting to creep into underwriting, compliance and collections, an area where the margin for regulatory error is minimal.

Anthropic sets up shop in Manhattan: AI is no longer just a San Francisco thing

🕒 Published on Zendoric: July 8, 2026 · 09:15

The New York Times reports that Anthropic is expanding its presence in Manhattan, part of a broader AI boom that is reshaping the map of talent and capital in New York. The details are sparse, but the underlying trend—the geographic decentralization of the industry—is worth noting.

Moderna bets on mRNA 2.0: cancer, autoimmunity, and an AI that designs drugs without humans between cycles

🕒 Published on Zendoric: July 8, 2026 · 09:15

At its Science Day, Moderna unveiled two unprecedented programs—an in vivo CAR-T for lupus and a multiplexed therapeutic against ovarian cancer—alongside Lucy, an AI platform that runs closed experimental cycles without human intervention. The market responded by sending the stock up nearly 75% from its lows, although Wall Street remains largely skeptical about whether the promise will translate into actual approvals.

Oracle connects AI to its cloud policies: the battle is no longer the model, it's the plumbing

🕒 Published on Zendoric: July 8, 2026 · 09:15

Oracle has equipped its IAM policy analysis tool on OCI with an MCP server, allowing agents like Claude or Codex to query cloud access permissions with real, deterministic data instead of pasted text or the model's generic memory. It's a small but revealing example of where enterprise AI is headed.

A teacher accused of creating fake nude images of former students with AI closes his case with a plea deal

🕒 Published on Zendoric: July 8, 2026 · 09:15

A former Northview High School (Michigan) teacher accused of using AI to generate fake nude photos of former students has reached a plea agreement, according to WZZM13. The case is another symptom of how AI image generation has become accessible for abuse, well before the law and schools know how to contain it.

From prompt to 'loop': why agentic AI is a governance problem, not just an engineering one

🕒 Published on Zendoric: July 8, 2026 · 09:15

Important notice: the downloaded content of this article by Enrique Dans in Fast Company is cut off by the subscription wall (the notice "Expand to continue reading ↓" appears after just four or five paragraphs).

Altman's plan to distribute OpenAI shares among Americans: real policy or narrative?

🕒 Published on Zendoric: July 8, 2026 · 09:15

James O'Donnell's article in MIT Technology Review analyzes the resurgence of an idea that Sam Altman has been promoting for years: that American citizens should receive a share of the wealth generated by AI.

Internal US Treasury report compares the AI bubble to the 'dot-com' collapse

🕒 Published on Zendoric: July 8, 2026 · 09:15

An internal report prepared by career analysts at the US Treasury Department, obtained by NOTUS, warns of the risks posed by the artificial intelligence market, comparing some of its features to the 'dot-com' bubble that burst in the early 2000s.

FT: the article on hidden risks in AI profits is blocked by a paywall

🕒 Published on Zendoric: July 8, 2026 · 09:15

The link included in the newsletter forwarded to Manuel points to an opinion article from Financial Times titled "Investors must be wary of the earnings bubble," indexed in the digest under the headline "AI profits are hiding bigger risks in earnings reports."

Samsung's profits surge 1,800% driven by demand for AI memory chips

🕒 Published on Zendoric: July 8, 2026 · 09:15

Samsung Electronics has announced a profit forecast pointing to an operating profit of 89.4 trillion won (about $58.4 billion) between April and June, representing an increase of roughly 19 times over the same period the previous year, that is, a jump of 1,800%.

CISA uses Anthropic's Mythos to audit government code despite tension with the White House

🕒 Published on Zendoric: July 8, 2026 · 09:15

Reuters reveals, citing three sources familiar with the matter, that the US Cybersecurity and Infrastructure Security Agency (CISA) is using Mythos, Anthropic's AI model specialized in cybersecurity, to audit government code repositories in search of vulnerabilities exploitable by spies…

Illinois passes the strictest AI safety law in the US, with mandatory audits

🕒 Published on Zendoric: July 8, 2026 · 09:15

Illinois Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act on Monday, a law that puts the state at the forefront of state-level artificial intelligence regulation in the United States.

Anthropic spied on Chinese Claude Code customers to detect 'distillation' of its AI by rivals

🕒 Published on Zendoric: July 8, 2026 · 09:15

This Washington Post article is protected by a paywall: the downloaded content is only a teaser (headline, lead paragraph and byline), followed entirely by the site's navigation menus.

Anthropic secretly removes a hidden tracker in Claude Code that monitored Chinese users

🕒 Published on Zendoric: July 8, 2026 · 09:15

A security researcher and web developer known as "Thereallo" discovered that Anthropic had built into Claude Code a "prompt steganography" mechanism that hid code capable of tracking users in China.

Anthropic says Claude has a hidden internal space ('J-Space') where it thinks without words

🕒 Published on Zendoric: July 8, 2026 · 09:15

Anthropic announced on Monday that it has identified a kind of internal workspace within Claude—which it has named "J-Space," after the Jacobian mathematical technique used to detect it—where the model would hold and manipulate ideas without needing to convert them into words.

Tilly Norwood, the AI 'actress' who outraged Hollywood, will debut in the feature film 'Misaligned'

🕒 Published on Zendoric: July 8, 2026 · 09:15

Variety reports that Tilly Norwood, the AI-generated 'actress' who sparked a strong backlash in the film industry in late 2025, will star in her first feature film, titled 'Misaligned'.

Anthropic discovers 'J-space': a hidden global workspace where Claude thinks without saying it

🕒 Published on Zendoric: July 8, 2026 · 09:15

Anthropic has published interpretability research claiming to have identified, within the language models of the Claude family, a small collection of internal neural patterns that serve a special function relative to the rest of the model's processing.

Singapore expands NVIDIA chip fraud case (content unavailable)

🕒 Published on Zendoric: July 8, 2026 · 09:15

It was not possible to access the actual content of the article: the download only returned the cookie notice and privacy preferences of the Yahoo Finance site, without any paragraph of the actual news. Therefore, I do not have verifiable information about the facts, figures, names or specific dates of the case.

Samsung could multiply its profit 18-fold thanks to the AI memory boom

🕒 Published on Zendoric: July 8, 2026 · 09:15

Samsung Electronics is heading for another record quarter. According to LSEG data cited in the article, analysts project an operating profit of about 86 trillion won (approximately $56 billion) for the second quarter, almost eighteen times higher than a year earlier.

Content unavailable: only YouTube cookie notice (Codex Clones Excel)

🕒 Published on Zendoric: July 8, 2026 · 09:15

It's important to be transparent with readers: the provided link corresponds to a YouTube Short, but the automatic content download only captured the Google/YouTube cookie consent screen (legal texts about privacy, cookie management and language options).

A brief history of model distillation

🕒 Published on Zendoric: July 8, 2026 · 09:15

The article begins by questioning the usual narrative about the origin of knowledge distillation, which is usually placed in 2015 with the work of Geoffrey Hinton, Oriol Vinyals and Jeff Dean, who introduced the softmax 'temperature' trick and coined the term 'dark…