👯 Asked AI for Ideas Lately?

Why asking AI the same way can get you the same idea, a WWII U-boat code gets cracked, and more...  ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌

October 27, 2026   |   Read online   Hey, Jhon! Welcome back to the #1 AI newsletter in the world! Here’s what we have today: Why asking AI the same way gets you the same idea. One founder and Claude take on a rare disease. AI helps crack a WWII U-boat message. Birdwatchers’ notes become data for science.       Brainstorming The Toy Idea Strangers Shared Source: Wharton A desk fan and a brick don’t sound like the start of a great toy. But in a study from Wharton researchers on how people come up with ideas with and without AI, that’s what everyone got to work with. The challenge was to invent a toy using only those two things. Some people brainstormed on their own while others had an AI chatbot helping them out. The ones working alone came back with all kinds of different toys. With AI helping, only 6% of the ideas were one of a kind, and several people working separately gave their toy the exact same name (Build-a-Breeze Castle, if you’re wondering). The researchers were quick to point out the ideas themselves were great. What tripped people up was how they asked. A lot of them typed close to the same request, and AI goes for the most likely answer to whatever you give it. Ask the way everyone else does, and there’s a good chance you’ll end up with the same idea everyone else got. That’s a little awkward when it’s a gift idea, and a lot worse when it’s the side project you’ve been planning to spend your weekends on. It was one toy challenge, not a verdict on every idea AI has ever handed out. The researchers had a few suggestions. Start with a couple of rough ideas of your own before you bring AI in. Put the ask in your own words instead of typing the obvious request. Break it into smaller steps rather than asking for one idea all at once. Run the same ask through a few different AIs and mix what comes back. One of them put it pretty bluntly, saying changing up your prompts costs next to nothing and it’d be foolish not to. Our take: The real lesson here is that AI hands back what you put in, which means the person asking still decides how original the idea ends up. Getting a fresh idea is only half of it. Before you put your weekends into one, you’d want to know whether it holds up, and that takes a lot more than another round of brainstorming.       Together with Rocket Does Your AI Idea Hold Up? An app that splits grocery bills between roommates sounds like a sure thing at 1 am. By morning, you’re wondering whether anyone would use it, and whether it’s worth the time you’d sink into building it. That’s the kind of question a quick brainstorm can’t settle. Rocket’s Solve takes that question and comes back with findings, evidence, and a straight recommendation on whether your idea holds up. You type it the way you’d ask a friend, and before it goes looking for sources, it works out what matters for your situation, like how crowded the space is and what could trip you up. When the idea’s still fuzzy, it checks in with a quick question or two. What comes back reads like a short report with the sources right there, and you can save it as a PDF or slides to look over later. If the answer’s yes, the plan carries straight into Rocket’s builder, so you’re not starting from scratch when it’s time to make the thing.       Drug Discovery Claude Helped Run a Drug Startup Source: Anthropic Getting a new medicine made has always been a team sport, and a big, expensive one at that. It’s a big reason so many rare diseases go years without anyone working on them. When only a small group of people has a condition, it’s hard to justify paying a team to chase it. Well, one person with Claude made a really good case that it doesn’t have to work that way. He ran a tiny drug company on his own, working with 11 outside labs, and handed Claude the contracts and the experiments with each lab, plus a lot of the day-to-day running of the place. In under 18 months, a possible treatment for one form of muscular dystrophy started as an idea and showed promising results in mice. It went well enough that the company got acquired, and the research is still going, with more programs on the way. If you’ve ever shelved an idea because it needed a team to pull off, this is the same thing playing out in medicine. None of it has reached a patient yet, though. It’s early lab work, and the hope is to start human trials as soon as late 2028 if the results keep holding up. Our take: Claude didn’t come up with the science here, and 11 labs still did all the testing. What it took off one person’s plate was the running-the-company part, which is often what stops a good idea before it ever starts. If that holds up, the diseases that were never worth a big team’s time might not need one anymore.       Interesting AI Type Changes Into a Live World Source: Google DeepMind Picture a guy playing a saxophone on your screen like a scene from a movie, except none of it was ever filmed. That’s Genie 4. It’s Google’s AI that builds an entire world as you watch. It even adds sound to match the scene. And you can change it while it’s running. Type something like “add a whale’s tail behind him” or “make it snow,” and the scene shifts right there without starting over. Some of the people in it will even talk back. The bigger reason it exists is robots. In one test, another AI broke a task into steps and Genie 4 played out each one first, a kind of dress rehearsal before a real robot ever tries it.       Coming in Hot AI Tools of the Day 🪞 Elmo tracks what ChatGPT and other AI assistants tell people about your brand and which competitors they name instead, free and open source on GitHub. 🎤 Colloqa plays the other side of a job interview or salary talk out loud, pushes back like a real person, then coaches you on what to fix. 🧭 Ayven chases down the one teammate who holds each answer inside Slack or Teams, then files tickets only once you approve them. 📸 AIPhotoEdit.ai swaps the outfit or background in any photo from a single typed sentence and leaves your face exactly as it was. 🤖 Orca runs Claude Code and Codex at once on your existing subscription and pings your phone the moment a job finishes.       Notable AIs Notable AI Tools 🎬 Clueso is an AI agent that turns a screen recording, slide deck, doc or just an idea into a polished, on-brand video in minutes.       Codebreaking U-Boat Code Cracked With AI’s Help Source: Crypto Cellar Research During World War II, cracking Enigma took rooms full of machines and a small army of codebreakers working around the clock. AI has already cracked a couple of old army messages on its own. The navy’s were a different story. Their version had an extra wheel inside that made each message that much harder to read, and a few of them have sat on a list of unsolved codes for years, waiting for someone to crack them from scratch. One of them finally gave, and the person behind it had AI build most of his toolkit, including a working Enigma machine in software and the code-cracking tools to go with it. The real legwork was his. He dug through old wartime records until he found the decoded version codebreakers filed back then, and used it to build a long list of guesses at what the message might say. The guess that cracked it was number 123. Then he checked it twice. A second guess landed on the same key, and a separate program that shares no code with the first decoded it all over again. After all that, the message turned out to be a schedule for radio beacons over three days in March 1945. Not exactly a war secret, but there are only two navy messages left on that list now. Our take: Each of these breaks has looked a little different, with AI running the whole show on one and building the tools for a person on this one. Either way, the pile of messages still waiting to be cracked keeps getting smaller, and at this pace, it won’t be a pile much longer.       Nature Birdwatchers Helped Science Without Knowing Source: University of Michigan Somebody at their kitchen window watches a blue jay chase a cardinal off the feeder, types a quick note into their bird app, and goes back to their coffee. Multiply that by millions of people, and you end up with a huge pile of notes about how animals and plants get along with each other. It’s the kind of thing scientists have a hard time collecting. The trouble was reading them all, which is why the job was handed to AI. It went through more than 18 million birdwatching records from one state and pulled out the moments where one species was dealing with another. They tried the same thing on a butterfly and turned up about 20 plants people had spotted it on that a major research database had never listed for it. The AI got around four out of five right, and it still needed people to step in on the odd stuff, like a note about “bohemians” that was really about a bird called the Bohemian Waxwing. Our take: Old diaries turned out to be some of the best history we have, even though the people writing them were only jotting down their day. Birdwatching notes look headed the same way, and AI is what finally makes it possible to read them all instead of letting them sit.       Prompt of the Day Forecast My Week Important Note: You’ll need to click the button to get the complete prompt. View and copy today’s full prompt This prompt turns your next seven days into a weather forecast, with pressure systems, fronts and a storm warning on the day most likely to go sideways. You move one item out of the storm’s path before the chat ends, then grade the forecast when the week is over.       That’s all for Tuesday! Before you go, run your idea past Rocket’s Solve. Type it the way you’d say it out loud, and you’ll have a straight answer before you build a thing. How did we do today? Rate this issue, then tell us what you thought. We value your input and will use it to make TAAFT better for you. ⭐⭐⭐⭐⭐ Absolutely Loved It ⭐⭐⭐⭐ Better Than Usual ⭐⭐⭐ Meh, It Was Okay ⭐⭐ Could Be Better Thanks for reading, — There’s An AI For That     Reach out to over 1.8 million AI enthusiasts today. Become a sponsor Update your email preferences or unsubscribe here © 2026 There's An AI For That® SRL 228 Park Ave S, #29976, New York, New York 10003, United States