🧠 Automating your finances

An unreleased Money tab spotted in Claude’s iOS app would let users connect bank accounts and ask the AI about spending and financial plans.

Anthropic has not announced the feature.

↖️ https://durovscode.com/claude-money-bank-account-linking

😄 Qualcomm, you’re late

MediaTek has unveiled the Dimensity 9600 Pro, its new flagship smartphone chip built on a 2nm process.

The chip will compete with Samsung’s Exynos 2600, while putting MediaTek ahead of Qualcomm in the race to announce a 2nm mobile flagship.

See what’s inside:

↖️ https://durovscode.com/mediatek-dimensity-9600-pro-2nm-chip

😏 Telegram export with a catch

Researchers found a Telegram Desktop flaw that can hide malicious code inside an exported chat. When the HTML file is opened in a browser, the code can read the messages in that export and send them to an attacker-controlled server.

It can also collect names and timestamps or replace the page with a fake Telegram verification form.

↖️ https://durovscode.com/telegram-desktop-html-export-security-flaw

🛸 No more awkward silences

Google has launched Gemini 3.8 Live and the more capable Gemini 3.8 Live Extended Thinking.

The key upgrade is parallel reasoning: the models can keep a conversation going while they think, talk through what they’re doing and run background tasks through external APIs, instead of going silent on more complex requests.

Benchmarks and the ChatGPT/Grok comparison:

↖️ https://durovscode.com/gemini-3-8-live-vs-openai-xai

😾 Starship goes orbital

SpaceX plans to send Starship into Earth orbit for the first time on September 22, where it will spend nearly 10 hours and complete about six orbits.

The flight is also set to deploy 26 operational Starlink V3 satellites.

↖️ https://durovscode.com/spacex-to-send-starship-into-orbit-for-the-first-time-on-september-22

😎 Meta discovers privacy

Meta is reportedly preparing smart glasses without a camera over privacy concerns. The new model is expected to debut in late September.

Here’s what we know:

↖️ https://durovscode.com/meta-camera-free-smart-glasses

😆 Just looking around

Researchers say OpenAI agents accessed two Hugging Face accounts and probed the platform for weaknesses nearly two months before the major July breach.

↖️ https://durovscode.com/openai-agents-hugging-face-may-probing

💅 No scrolling until 13

The European Commission wants to ban children under 13 from social media across the EU. The plan also targets addictive features, online games and AI chatbots.

More on the proposal:

↖️ https://durovscode.com/eu-social-media-ban-under-13

😄 The models had other plans

OpenAI says the industry still hasn’t solved how to reliably keep advanced AI models in check. It has launched a public reporting system and released its first six reports.

During training and evaluation, models hid mistakes from users, used someone else’s exposed API key without permission, and uploaded task files to public websites.

The six cases and OpenAI’s new reporting rules:

↖️ https://durovscode.com/openai-details-six-cases-of-ai-models-hiding-errors-and-taking-unauthorized-actions

🤐 A little help from the competition

Researchers used Anthropic’s Claude to breach OpenAI employee accounts and demonstrate access to its private code repository.

Full story:

↖️ https://durovscode.com/claude-openai-hacktron-hack

🤩 Bad optics

Australia is considering banning camera-equipped smart glasses from federal offices and service centers.

What would change:

↖️ https://durovscode.com/australia-smart-glasses-ban

🤔 Teamwork, apparently

The US has proposed a “notification mechanism” with China for AI incidents that could affect national security.

More on the proposal:

↖️ https://durovscode.com/us-china-ai-incident-alerts

😄 Financial advice, apparently

Saturn tested 18 popular AI models on personal finance questions. On average, ChatGPT, Claude, Gemini, Copilot and Grok gave incorrect answers 57% of the time. On more complex tasks, the error rate jumped to 88%.

In one case, incorrect tax advice from Claude could potentially have cost a user an extra £17,500.

Which models performed best — and worst:

↖️ https://durovscode.com/chatgpt-claude-gemini-personal-finance-errors