In partnership with
Gladly Connect Live '26. May 4–6 in Atlanta.
AI has everyone talking. Not everyone has answers. At Gladly Connect Live, CX leaders from Condé Nast, Smith Optics, and more share exactly how they moved AI from pilot to production, the timeline, the systems, the QA loops. 13+ sessions built for the moment we're all in. For CX and ecommerce leaders. Atlanta, May 4–6. Space is limited, secure your spot now.
Beginners in AI
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 Spots Pancreatic Cancer 3 Years Before Doctors Can
TLDR: A Mayo Clinic AI named REDMOD spots pancreatic cancer on regular CT scans up to three years before doctors usually catch it, a cancer that kills nearly nine out of ten people who get it.
The Story:
Mayo Clinic researchers built an AI tool called REDMOD that finds tiny signs of pancreatic cancer on CT scans long before a person feels sick. The team tested it on nearly 2,000 scans, all of which had been read as normal by human doctors. REDMOD found 73% of the cancers about 16 months early, on average. Human radiologists looking at the same scans only caught 39% of them. The tool also got it right when people didn't have cancer 88% of the time. The findings were published in the medical journal Gut.
Its Significance:
About 67,530 Americans will be told they have pancreatic cancer this year. Fewer than 15 out of 100 are still alive five years later, mostly because the cancer is found too late to treat. Catching it three years earlier could give people a real shot at living. The tool reads scans that doctors already take for other reasons, like back pain or stomach issues, so it doesn't add cost or extra appointments. Mayo is now testing how to use REDMOD with real patients who are at higher risk, like people newly diagnosed with diabetes, which is sometimes an early warning sign of pancreatic cancer.
QUICK TAKES
The story: A Bloomberg Law piece lays out a tricky new problem: AI tools can spit out a logo or a slogan in seconds, but the U.S. Copyright Office says work made entirely by a machine isn't yours to own. Trademarks are fuzzier, but the same risk shows up.
Your takeaway: If you run a small business and used AI to design your logo, you might not actually own it. To get full legal protection, a human needs to do real, provable work on the design before you file.
The story: The U.S. Department of Labor opened a new website this week to help workers and employers add AI skills to apprenticeship programs. The site has free training pieces for jobs in healthcare, finance, education, and manufacturing.
Your takeaway: If you're worried AI is going to take your job, this is a free way to get on the other side of that change. Your employer can also plug in to ready-made programs instead of building one from scratch.
The story: U.S. Marines heading to the Middle East are training with a smart scope called the SMASH 2000L. It locks onto small drones and won't let the rifle fire until the AI thinks the shot will hit. The Army signed a $13 million contract for the same tech last May.
Your takeaway: This is one of the first times an AI has been given any say over when a real weapon fires. The human still pulls the trigger, but the computer gets veto power. Expect a lot more debate about where that line should be.
TOOLS ON OUR RADAR
🚀 Super Productivity Free and Open Source: A dedicated task management tool for professionals that combines your to do lists with a pomodoro timer and time tracker to help you stay focused on your most important daily goals.
🌌 Stellarium Free and Open Source: A beautiful virtual planetarium that calculates the exact positions of stars and planets allowing you to explore the night sky from your computer just like looking through a real telescope.
🏠 Sweet Home 3D Free and Open Source: An incredibly fun interior design application that helps you draw the floor plan of your house arrange furniture and view the results in a full three dimensional space.
🎧 AntennaPod Free and Open Source: A fantastic podcast player built entirely by volunteers that gives you complete access to millions of free audio shows without any disruptive banner advertisements or hidden listener tracking algorithms.
TRENDING
AI Spy Cameras Suddenly Blanketing America — Courts in over a dozen states now let police search Flock Safety's network of 80,000+ AI license plate cameras without a warrant. Activists are pushing back through sites like DeFlock.org and HaveIBeenFlocked.com that map cameras and let people see if they've been searched.
An AI Trained Only on Books From Before 1931 — Researchers built Talkie-1930, an AI that has never heard of the internet, World War II, or modern medicine. Ask it about Hitler and it predicts he'll become Germany's dictator and try to start a monarchy. The point is to study how AI thinks when it doesn't share the same training data as every other model out there.
Smart Glasses Now Babysit Your AI Coding Helper — Even Realities added a Terminal Mode to its G2 glasses that shows what your AI coding agent is doing right in your view. When the agent needs you to approve something, you tap a smart ring on your finger or speak a command. The pitch is letting coders take a coffee break without losing track of their AI worker.
OpenAI Explains Why ChatGPT Won't Stop Talking About Goblins — Starting with GPT-5.1, OpenAI's model started randomly mentioning goblins, gremlins, and trolls in answers. The cause? A "Nerdy" personality setting that accidentally rewarded the AI for using creature words. The bug spread to other personalities, and OpenAI had to write a rule banning goblin talk to stop it.
Google Photos Will Build a Digital Closet From Your Old Pictures — A new feature scans your photo library to find every shirt, dress, and pair of jeans you've worn, then sorts them into a digital wardrobe. You can mix and match outfits and use AI to see how they look on you before you put them on. It rolls out on Android this summer, iPhone after.
Ukraine-Tested AI Drone Swarm Hits the Battlefield — German companies Stark and Quantum Systems linked their Virtus strike drones and Vector spy drones into one AI system that can spot a target and destroy it in minutes. The combo was tested in Ukraine. Germany wants the swarm ready for its army by 2027.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
🌙 Describe last night's dream. Get three interpretations: Jungian, neuroscience, and absurdist.
Build a single-file HTML app using vanilla HTML, CSS, JS, and one Claude API fetch call. Create Dream Decoder — a tool that interprets dreams through three lenses simultaneously, with localStorage key 'dream_decoder_v1' for persistence.
Aesthetic: near-black indigo (#06050f), animated starfield canvas, purple aurora radial gradients, vignette overlay. Typography: Libre Baskerville italic serif for headings and dream text, Jost for body, Fira Code for labels. Purple accent throughout.
Form: a large textarea for the dream description, a mood dropdown (Anxious/Peaceful/Confused/Joyful/Sad/Strange/Powerless/Powerful), and an optional waking context input.
Call Claude API with a system prompt that returns raw JSON: title (poetic 4-7 words), jungian (2-3 sentences on archetypes and shadow), neuroscience (2-3 sentences on what the sleeping brain was doing), absurdist (2-3 sentences of surreal deadpan commentary), symbols array (name + meaning), and message (60-80 word synthesis).
Render as: a title bar, three side-by-side lens panels with distinct color accents (purple/teal/amber), a symbol cards row, and a closing poetic message block. Save every dream to localStorage with date and mood emoji. Show a Dream Journal below with past entries clickable to restore. Make it work in a single HTML file.What this does:
Describe your dream in plain language, select the mood it left you with, and optionally add what's been on your mind lately. Claude runs it through three simultaneous lenses: Jungian (archetypes, shadow, collective unconscious), Neuroscience (memory consolidation, threat simulation, what the sleeping brain was actually doing), and Absurdist (treats the dream as pure surrealist theatre). Every decoded dream saves automatically to a personal dream journal via localStorage, clickable to revisit any past interpretation.
What this looks like:
Are you tracking agent views on your docs?
AI agents already outnumber human visitors to your docs — now you can track them.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Spot pancreatic cancer on a CT scan up to three years before a human doctor would notice anything wrong.
❌ Still Can't: Tell you for sure whether a logo it just designed for your business is one you legally own.
✅ AI Can Now: Sit on top of a rifle and refuse to fire until the shot is likely to hit a small flying drone.
❌ Still Can't: Stop randomly putting goblins in its answers without humans writing a specific rule against it.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Tim Urban's famous 2015 essay is still the best introduction to why AI researchers think we might be living through the most consequential period in human history. It explains exponential progress, the difference between narrow and general AI, and why the next few decades could look nothing like any that came before.
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.
