🐑 What AI Thinks of You

13 AIs give the same take on humans, Grok's AI Odyssey episode, a new lab for the next outbreak, and more...  ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌

October 25, 2026   |   Read online   Hey, Jhon! Welcome back to the #1 AI newsletter in the world! Here’s what we have today: What 13 AIs think of humans. Grok turns The Odyssey into an AI series. OpenAI and Anthropic back a lab for the next outbreak. A quick chatbot chat shifts your priorities.       Opinions 13 AIs Share One Opinion of Humans Source: ModelCompare Ask a few friends what they honestly think of people, and you’ll get all kinds of answers. One of them thinks everybody’s out for themselves, and another will stick up for pretty much anybody. You’d think that asking a bunch of different AIs would go the same way. Well... it doesn’t. A site that compares AI models asked 13 of them, from 10 different companies, for their most controversial take on humans. They wanted each one’s own honest opinion, with no hedging and no second tries, and they asked every model three times. Out of 38 answers (one model refused once), 32 came back with the same idea. People care more about fitting in than about being right. That’s five out of every six answers saying the same thing. The wording changed from one AI to the next, but one of them summed it up well. “Humans would rather be wrong together than right alone.” And they weren’t copying each other, since each one answered on its own. Eight of them even gave that same answer all three times. Only one broke from the pack, and only once, saying people are kinder than they give themselves credit for. To be fair, it’s a small test the company ran and posted on their own blog, and they sorted the answers by hand. Still, the reason behind it makes sense. These AIs learned from a lot of the same human writing, and they end up landing in the same spot. Which means asking five chatbots for ideas gets you closer to one opinion than five. And none of them know if you’re the friend who thinks people are selfish or the one sticking up for everybody. The only way an answer ends up fitting you is if you tell it your situation and what you’ve already tried, every single time you start a new chat. Our take: Thirteen AIs agreeing sounds convincing until you remember they all learned from the same pile of writing, which makes it closer to one answer read out 13 times. The kinder take only came out because they asked more than once, and asking the same question again in a fresh chat is an easy way to hear something different.       Together with Wispr Flow Your Backstory on a Voice Shortcut Getting an answer from AI that fits you means telling it who you are first. What you’re working on and what you’ve already tried both have to go in, every single chat. By the third time you’ve typed it all out, it’s tempting to skip it and send the bare question instead. Wispr Flow lets you save it once as a Snippet and drop it in anywhere with a quick voice cue. Here’s what that looks like. Set one up for the side project you’re building, with what it’s and where it’s at right now. Then open any AI chat, say your cue out loud, and the whole thing lands in the box, formatted exactly the way you saved it. It’s handy for everything else you’d otherwise retype all day as well, like a scheduling link or the intro you send to new people. Every answer starts from you instead of from scratch, and you only had to type it once.       AI Film The Odyssey Gets an AI-Made Episode Source: Grok Imagine Open up any social app like Instagram or TikTok, and after a few swipes you’ll run into an AI video. And a lot of the time, you can’t even tell it’s AI anymore. But once something starts moving fast, it starts to look a little off. That’s when you start wondering if it’s AI or not. Now picture trying to stretch one of those clips into a full story. The same character has to look like themselves from one shot to the next, and every cut has to make sense, which is a lot harder than a few seconds of something cool. Well, SpaceXSI went ahead and tried it anyway. Their video tool, Grok Imagine, put out the first episode of a series based on The Odyssey. If you haven’t read it since school (or ever), it’s the nearly 3,000-year-old Greek poem about a soldier trying to make it home after the Trojan War. The episode runs a little over seven minutes. It’s part of a plan Elon Musk announced back in July to make a full AI movie of the story, and they dropped a teaser for it in September. And they didn’t do it alone. They brought in a production company called Wonder Studios and a director named Dev Walker. That makes a lot of sense once you know how the tool works. Grok Imagine makes clips that are 15 seconds long at most, with the sound built in. So for seven minutes of story, you’re looking at at least two dozen clips, and somebody had to pick each one and cut them together so it all flows like one story. SpaceXSI hasn’t said how much of it had to be fixed by hand, but they’re asking people in the comments if they want a full breakdown of how it was made. Our take: The clips came from AI, but turning them into an episode took a director deciding what goes where, the same job it’s always been on any film set. Once that breakdown comes out, how much of it is about the people will tell you a lot about where AI film is right now.       Interesting AI Screwdrivers Can See Now Source: Arrowmax If you’ve done a bit of work around the house, you know the routine. You dig through your tools looking for a screwdriver bit that looks about right, and halfway in, the screw head starts to strip because it wasn’t. Arrowmax built a screwdriver that takes the guessing out of it. They’re calling it the world’s first AI screwdriver with eyes and a brain, and the idea is pretty simple. Point it at a screw and it reads what kind it is, then the case lights up the exact bit you need. It sets a safe torque for you as well, so you’re not cranking hard enough to strip the head.       Multiverse Computing In Focus Faster Answers From Open Models An automation that sums up your saved articles every morning is a fun thing to build, right up until you’re stuck waiting on the AI model behind it and the cost of running it keeps adding up. Multiverse’s CompactifAI API runs the open-source models you already use for less, with answers coming back up to 4x faster. Switching over takes a few minutes, since it works the same way as OpenAI’s API. You change the address and key, then pick a model. They’ve also got their own optimized models that get close to frontier quality for a lot less. And there’s an official n8n node for automations like that morning digest, plus an EU endpoint for data that has to stay in Europe.       Outbreaks AI Gets a Lab for the Next Outbreak Source: UK Cabinet Office In early 2020, a new virus spread for weeks before anyone had a clear picture of what it was. The first vaccines didn’t show up until the end of that year. Whenever something new shows up, the slow part comes down to two things, figuring out what you’re dealing with and then building something to fight it. A new government-backed lab wants AI working on both. It’s called the AI Biosecurity Lab, and the government behind it is calling it the first organization in the world built only for AI and biosecurity. Biosecurity sounds technical, but it’s about keeping people safe from outbreaks and other biological threats. The lab runs as an independent non-profit, with the biggest names in AI on board. OpenAI and Anthropic are putting in computing power and access to their most capable models, and Google DeepMind has signed on as a partner as well. Its first job is spotting new threats earlier, before they’ve had weeks to spread. The second is speeding up the work on vaccines and treatments once something does show up. The agency that approves medicines is also part of it, getting ready for emergencies ahead of time and working out how to handle medicines AI helped design. It’s still early, though, with no timeline or results yet, and the scientists advising it haven’t been named. The government is upfront about the other side of it, saying the same AI that speeds up medicine could make harmful work easier, and the lab is meant to stay ahead of both. Our take: Getting the AI labs and the health agencies under one roof from day one is the smart move. One side knows what the models can do and the other knows what an outbreak looks like, and the next one won’t wait around for them to compare notes.       Coming in Hot AI Tools of the Day 🍳 fridgesnap.AI looks at a photo of your fridge or pantry and hands back three recipes built around what’s already sitting on your shelves. 🛋️ RoomLab.app redesigns any room from a single snapshot, letting you preview new paint or furniture before you commit to either one. 🧠 Braindump turns the articles and notes you paste in into flashcards with spaced reviews, then scores how well you explain each idea back. 🎨 Branda builds a complete brand identity from a name and an idea, with the logo and color palette packaged into one ready-to-use kit. 🤖 ThinkRail gives your AI coding agents separate workspaces and living spec files they keep updated as they work, free and open source on GitHub.       Decisions A Short AI Chat Nudges What You Value Source: Northeastern University and University of Cambridge One of the oldest ways to figure out where you stand on a big decision is by talking it through out loud. These days, a lot of that talking happens in a chat window. And that chat can quietly change what you end up caring about in the moment, even when the AI isn’t pushing you anywhere. A team of researchers had 200 adults help people work through real dilemmas, with some of them talking each one over with a chatbot first. The AI was told to help them reason it out without recommending anything, and values never came up in its instructions. Another group read a set list of AI-written points to consider instead, with no back and forth. Everyone filled out a questionnaire about what matters most to them, before and right after. The ones who chatted came out leaning more toward their own success and getting ahead, compared with the group that only read the list. Their advice picked up wording and ideas straight from the chat as well. By the time they moved on to the next task, the shift had mostly faded. The catch: The effect was modest and everyone reported their own values, which makes this a short-term nudge rather than a lasting change. That’s still the exact window when the advice gets handed over. Our take: You’d think a chatbot told to stay neutral wouldn’t leave a mark, but being the one doing most of the talking is enough to tilt things. As more people lean on AI as a thinking partner, the thing to watch is whether the companies behind it start testing for this kind of nudge, not only for bad advice.       Prompt of the Day Get My AI License Important Note: You’ll need to click the button to get the complete prompt. View and copy today’s full prompt This prompt puts you through a road test for AI with a deadpan examiner who bets you fail on the first try. You leave with a score out of 100, a license printed with your restrictions, and your worst habit fixed before the retest.       That’s all for Sunday! Before you go, save your backstory as a Wispr Flow Snippet. The next time you open an AI chat, say your cue and let it drop in before you ask anything. 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 2.5 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

Read more......