Here’s what’s happening in tech..

Always inside: Free AI tools, simple crypto-security guides, data-protection steps, free courses, and the tools that help people work faster and smarter.

Let’s dive in

What was supposed to stay behind closed doors just became one of the biggest AI stories of the year. A configuration error inside Anthropic's content management system accidentally exposed nearly 3,000 internal files to the public internet, and security researchers found them before the company did.

Among those files was a draft blog post describing a new model called Claude Mythos, which Anthropic referred to internally as "by far the most powerful AI model we've ever developed." Anthropic has since confirmed the model is real.

Here is what the leaked documents revealed:

As a subscriber, you’ll always get a fresh list of premium AI tools we handpick for you in the Innorive+ Newsletter.

If you've been holding Bitcoin or USDC but don't have enough cash saved for a home down payment, a new partnership between Coinbase and mortgage lender Better Home Finance might change that.

Homebuyers can now pledge their crypto as collateral to cover a down payment without selling it.

Here is how it works:

Most people expect to be watched by security cameras in a store. Reading your facial expressions is a different category entirely.

Walmart filed a patent for facial recognition technology that scans customers in checkout lines and feeds footage into a computer to assess how each person is feeling. The stated purpose is customer service. The system links customers' facial expressions to their transaction data, meaning what they are buying and how much they are spending.

The patent has been on record since 2017 and there is no public evidence Walmart has deployed it in stores. But its existence raises a straightforward question about where retail surveillance is heading:

Each layer builds on the last.

Your data is worth a lot. Every piece of information about you, your location, what you buy, who you talk to, is being collected, repackaged, and sold. Companies profit billions, while you get nothing.

Step 1️⃣: Remove Past Data with Optery

Optery requests data removals from 1345+ data brokers and shows you verified proof with before-and-after screenshots. Instead of hoping companies delete you, you see it actually happen

✅Covers Google, Whitepages, TruthFinder, Instant Checkmate

✅30-day money-back guarantee

Step 2️⃣: Lock Down Future Data with Proton

Stop new data collection at the source. Proton encrypts your email, files, and browsing end-to-end starting with free plans. Data brokers have nothing new to collect from you.

Free plan available | Upgrade anytime

A new kind of worker is emerging. Not someone who does the work themselves, but someone who builds and manages a team of AI agents that does it for them, and brings that team to companies as a service.

Tools like Paperclip are making this more concrete. It lets users assign goals, coordinate agents across an org chart, and monitor work and costs from one dashboard.

The model looks something like this:

A solo founder profiled by Business Insider in February 2026 was already running 15 custom AI agents as a coordinated team, saving over 20 hours per week. That same setup could be contracted out to an entire department.

2025 was the year of the AI employee. 2026 is shaping up to be the year of the AI company.