Good morning and thank you for joining us again!

Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated and edited, and published with the intention of making AI news and technology more accessible to everyone.

THE FRONT PAGE

AI + Pokemon Go? Pokémon Go Players Unknowingly Built a 30 Billion Image Database Now Training Delivery Robots

TLDR: Pokémon Go players spent years scanning real-world landmarks for in-game rewards, and those 30 billion images are now being used to help delivery robots navigate city streets with pinpoint accuracy.

The Story:

Niantic, the company behind Pokémon Go, spun off a new AI company called Niantic Spatial last year. That company just signed a deal with Coco Robotics to use player-submitted images for robot navigation. The 30 billion photos came from players who scanned statues, public art, and other landmarks through their phone cameras, often in exchange for in-game rewards. Niantic Spatial trained a Visual Positioning System on those images that can pinpoint a device's location within a few centimeters, far better than GPS. Coco Robotics now runs about 1,000 delivery robots across Los Angeles, Chicago, Miami, and other cities, completing over 500,000 deliveries so far.

Its Significance:

This is another example of regular people doing free work that ends up being worth a lot of money to a company. When 500 million people installed Pokémon Go in its first 60 days, nobody was thinking about delivery robots. But Niantic's terms of service give the company the right to use player-uploaded data however it wants, and to pass that right to other companies. Players can stop uploading new data, but they can't take back what's already been collected. It's a pattern we've seen before: Google's reCAPTCHA had people label training data for self-driving cars while they thought they were just proving they were human. If you've ever scanned a landmark in Pokémon Go, your photos might already be helping a robot deliver pizza across town.

QUICK TAKES

Amazon Is Building AR Glasses to Take on Meta

The story: Amazon is developing consumer augmented reality glasses, codenamed "Jayhawk," with a planned release in late 2026 or early 2027. The glasses will feature speakers, a microphone, a camera, and a full-color display in one lens. Amazon is also building a bulkier version for its delivery drivers.

Your takeaway: Meta currently controls 73% of the smart glasses market. Amazon's entry means the race for your face is heating up, and prices could drop as more companies fight for customers.

AI Facial Recognition Error Puts Innocent Grandmother in Jail for Nearly 6 Months

The story: Angela Lipps, a 50-year-old grandmother from Tennessee, spent nearly six months in jail after Fargo, North Dakota police used an AI facial recognition tool that wrongly identified her as a bank fraud suspect. She had never been to North Dakota. Bank records proved she was 1,200 miles away when the crimes happened, but police didn't interview her until she had already been locked up for over five months.

Your takeaway: Lipps lost her home, her car, and her dog while she sat in jail. No one from the police department has apologized. This is at least the third known case of AI facial recognition leading to a wrongful arrest in the past year, and there's still no federal law requiring police to verify AI matches before making an arrest.

Encyclopedia Britannica Sues OpenAI Over AI Training

The story: Encyclopedia Britannica and its subsidiary Merriam-Webster are suing OpenAI, claiming ChatGPT used nearly 100,000 of Britannica's online articles as training data without permission. The lawsuit says OpenAI's AI summaries take a "free ride" on Britannica's research, cutting into its advertising and subscription revenue.

Your takeaway: This adds to a growing pile of copyright lawsuits against AI companies. For anyone who uses AI tools for research, it's a reminder that the information these tools produce often came from someone else's paid work.

TOOLS ON OUR RADAR

✅ Super Productivity Free and Open Source: An advanced daily task manager and time tracker that completely respects your privacy by keeping all your data locally on your machine. (Alternative to Todoist)

🧱 Walling Freemium: A highly visual project management tool that lets you organize your daily ideas and creative research into beautiful interactive brick layouts.

🖍️ Glasp Freemium: An incredibly useful browser extension that lets you easily highlight important quotes on any website and instantly share your curated reading list with a community of lifelong learners.

📹 Guidde Freemium: An intuitive video documentation platform powered by artificial intelligence that instantly turns your software workflows into stunning visual guides for your entire team.

TRENDING

Netflix Pays Up to $600 Million for Ben Affleck's AI Filmmaking Startup - Netflix acquired InterPositive, an AI company co-founded by Ben Affleck that helps filmmakers handle tedious post-production tasks like removing stunt wires, relighting scenes, and fixing continuity issues. Director David Fincher has already used the tools on an upcoming project.

Google Scraps AI Search Feature That Gave Amateur Medical Advice - Google quietly removed its "What People Suggest" feature, which used AI to pull crowdsourced health tips from Reddit, Quora, and X. The company says it was part of a "broader simplification," not a safety concern, though it also recently pulled some AI Overviews from medical searches after investigations found inaccurate health info.

Meta Signs $27 Billion AI Infrastructure Deal with Nebius - Meta agreed to pay up to $27 billion over five years to cloud provider Nebius for AI computing power. The deal includes an early deployment of Nvidia's newest Vera Rubin chips. Meta plans up to $135 billion in AI capital spending this year alone.

Alibaba Plans Enterprise AI Agent Launch This Week - Alibaba is preparing to launch an AI agent for businesses, built on its Qwen language model and developed by the DingTalk team. The tool will gradually connect to Taobao and Alipay, letting companies automate tasks across computers, browsers, and cloud servers.

Writing Professors Push for the Right to Refuse AI in the Classroom - The Conference on College Composition and Communication, the largest professional organization of writing educators, passed a resolution affirming the right of students and faculty to opt out of using AI in writing courses. The group cited concerns about critical thinking, data privacy, and labor rights.

OneVest Launches AI-Powered Wealth Management Platform - Financial tech company OneVest rolled out an AI-native operating system that automates middle-office work for wealth management firms, handling tasks like account opening, fund transfers, and compliance workflows so advisors can spend more time with clients.

Purdue Researchers Build Privacy Shield for AI Photo Editing - Purdue University engineers created a system that masks your face before uploading photos to AI editing tools, then blends it back in after the edit is done. In testing, the system reduced AI's ability to identify traits like eye color and age by more than 80%.

TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)

🎯 Build a weekly goal tracker with day-by-day scheduling, progress bars, and category breakdowns

What this does:

A weekly planner where you set up to 3 goals, pick a category and priority for each, choose which days of the week you'll work on them, and check them off as you finish. A progress bar, goals-remaining counter, and category breakdown all update live as you complete things. You can filter between all, active, or done goals, and it's styled in a dark lime green theme.

What this looks like:

WHERE WE STAND(based on today’s news)

✅ AI Can Now: Guide delivery robots within a few centimeters of accuracy using 30 billion crowdsourced images from a mobile game

❌ Still Can't: Reliably tell one person's face from another's well enough to avoid putting innocent people in jail for months

✅ AI Can Now: Help filmmakers relight scenes, remove stunt wires, and fix continuity errors using their own footage

❌ Still Can't: Train on copyrighted content without getting sued, as Britannica's nearly 100,000-article lawsuit against OpenAI shows

FROM THE WEB

RECOMMENDED LISTENING/READING/WATCHING

The 12 Week Year by Brian Moran

Argues that annual goals fail because they give you too much runway. Compressing your goals into 12-week sprints creates the urgency of a deadline without waiting until December. Pairs perfectly with the weekly goal tracker prompt and you can utilize your favorite AI tool to help you do detailed planning for every day of the year if you wanted. They excel at this kind of planning/scheduling workflow.

Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!

-James

By the way, this is the link if you liked the content and want to share with a friend.

^{Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.}