AI Is Everywhere - Just Let It In

Welcome to the fourth edition of The Neural Net…

This issue is a little different - the main article is a story-driven deep dive into how AI can quietly shape our daily routines, often without us even noticing. Let us know what you think - should we keep mixing in long-form editions like this? And don’t worry, we still have the latest AI headlines from the week.

In this issue:

• AI Runs Ralph - Behind The Scenes

• AI Lawyers Are Coming

• Hedge Fund Shorts Nvidia

• IPO Likelihood Successfully Predicted by AI

• Oops, AI Did It Again - Featuring Microsoft’s Copilot

• Former OpenAI CTO Launches Startup

AI Is Everywhere - All You Need To Do Is Let It In

This is a story of AI meets human.

For most people, AI still feels like something out of a sci-fi movie—abstract, futuristic, maybe even a little intimidating. But here’s the thing: Lifestyle AI is already here, and if you let it, it can run the background of your entire life.

For example, meet Ralph.

He’s 45, a real estate agent, a father of two, and the designated family chef because his wife works late shifts. He doesn’t think of himself as some techie AI power user. He’s never asked ChatGPT to write an email, never used an AI voice assistant to transcribe notes, and wouldn’t know a neural network from a fishing net. But he has signed up for a bunch of AI products from his favorite companies after his wife sent him to the dog house for forgetting their anniversary.

Now AI is quietly working behind the scenes, making things easier, faster, and more efficient. And with everything happening behind the scenes, Ralph doesn’t even realize just how much his life has changed.

☀️ 1. Morning: AI Knows Ralph Better Than He Knows Himself

While Ralph was dead to the world, his personal AI entourage was already at work.

• 5:57 AM: Ralph set his alarm for 6:30, but his smartwatch is able to move that up (within reason) to coincide with the lightest point in his sleep cycle. Ralph just assumes he’s having a good sleep day.

• 6:00 AM: His phone shifts his calendar reminders to account for heavier traffic than usual on his morning commute. Ralph is still chasing butterflies and frolicking through his dreamscape.

• 6:15 AM: His coffee maker starts brewing early to be ready for the new earlier alarm, but determines this should be a single shot kind of day based on his sleep rhythms through the night.

• 6:23 AM: Ralph’s alarm goes off. He wakes up feeling suspiciously… okay. AI has already orchestrated his morning before he even opened his eyes.

🔹 AI at work: Apple Sleep AI, Google Maps AI, Alexa, Smart Coffee Machine

🏡 2. Work: AI Makes Ralph Look Like a Genius

AI is quietly making Ralph faster, sharper, and just a little bit luckier. He isn’t actively using ChatGPT or Claude, but he’s closing more deals, wasting less time on dead leads, and - somehow - managing to be a semi-responsible dad.

• 9:15 AM: His real estate platform auto-generates a listing description for a new client’s property based on his notes, transforming “cramped” into “cozy.” Ralph just assumes this year’s intern is an English major.

• 10:30 AM: Ralph’s listing tracking system flags a couple of buyers who have been obsessively clicking on the same listing at 2 AM, and suggests he reach out to them.

• 11:45 AM: A client calls and asks, "So, how’s the market looking?" Ralph pulls his weekly AI-generated report, loaded with the latest pricing trends. He puts his feet on his desk and nods like he already knew all of this.

• 12:30 PM: His video doorbell pings - not just with a motion alert, but with an AI-generated notification: “Package delivered: Tennis rackets.” That’s when he remembers: his daughter’s first practice is tonight!

🔹 AI at work: Zillow AI (built into his brokerage software), Automated Market Reports, Ring AI, Amazon AI (default notifications Ralph never disabled)

🍽️ 3. Evening: AI Saves Ralph from Himself (And Finds Him a New Best Friend)

Tennis practice was a success but went long, and his kids are dramatically claiming starvation. Unbeknownst to Ralph, AI is the only thing keeping the family’s evening from total chaos.

• 6:45 PM: A few months ago, Ralph bought a smart fridge because it was on sale. He thought it was just a fridge, until it asked him to sign into all his accounts to “optimize” his shopping. So when Ralph used the last of the milk this morning, it placed a grocery order which was just delivered to his front porch.

• 7:00 PM: By the time Ralph gets home, the oven has preheated itself, and his smart speaker is playing “LoFi Beats For Evening Vibes.” After putting away the groceries, he can relax and quickly whip up dinner.

• 8:00 PM: Ralph collapses onto the couch, opens Netflix, and before he can even scroll, AI has already queued up a “highly recommended” show. Ralph watches the trailer and is thrilled to have another banger to watch once his wife Rose gets back.

• 8:15 PM: The front door opens as Rose comes home. But she’s not alone. In her arms is the cutest, tiniest poodle puppy Ralph has ever seen. What Ralph doesn’t realize? AI set this up. For weeks, Petfinder had been slipping “Recommended for You” puppies into her social feeds, subtly reinforcing the idea until, today, as she drove past the shelter, she just had to stop.

🔹 AI at work: Samsung Family Hub, Instacart AI, Google Nest, Spotify AI, Netflix AI, Petfinder AI

Ralph never set out to “use AI.” He isn’t somebody who’s always at the cutting edge, he’s just somebody that decided to let AI into his daily life.

AI is already working for the world’s biggest companies.

The only question is - Is it working for you?

Heard in the Server Room 
Quick hits roundup of AI news

• Hopeless To Handy: AI Lawyers Are Coming

  AI tools are getting better at answering tough legal questions, but they’re still no match for a real lawyer, at least that’s what top UK law firm Linklaters found. Their latest tests show OpenAI’s newest model, o1, performed the best and made big leaps over past versions, but still makes mistakes and needs human oversight. But hey, at least it won’t steal clients or angle for the corner office.

• Elliott Management Calls AI ‘Overhyped’, Shorts Nvidia

  Hedge fund giant Elliott Management is betting big against Nvidia, calling the AI-fueled stock surge a “bubble.” The firm, led by billionaire Paul Singer, has taken a massive short position, wagering that the hype around AI (and Nvidia’s dominance in it) might not last. Talk about swimming upstream.

• Can AI Predict the Next Big IPO? Crunchbase Thinks So

  Crunchbase is using AI to predict when startups will raise funds, get acquired, or go public, leveraging 17 years of data and user signals. The company’s new AI prediction engine has shown impressive accuracy, such as predicting Anthropic’s $2 billion raise and Coda’s acquisition by Grammarly. Ironically, when the model was asked to predict Crunchbase’s own future, it determined it was ripe for acquisition.

• DeepSeek Just Got Deep-Sixed in South Korea

  South Korea just slammed the brakes on DeepSeek, China’s AI chatbot, after discovering it was leaking data straight to ByteDance—aka TikTok’s parent company and a name that instantly raises red flags. Officials found that DeepSeek, which was supposed to be China’s answer to ChatGPT, was found to be “harvesting inexplicable amounts of data about its users.” China’s AI isn’t exactly giving the world the warm and fuzzies—more like the cold and surveilled.

Oops, AI Did It Again
AI fails to keep you laughing for days

Katharine, a good friend and fellow Neural Net subscriber, sent over a hilarious AI fail that we just had to feature. We spend a lot of time celebrating AI wins, but let’s be honest - for every success, there’s also a spectacular misinterpretation of instructions that reminds us AI still has a ways to go.

Today's star fail comes courtesy of Microsoft Copilot - the AI tool baked into Windows. It’s designed to supercharge productivity, helping with everything from Excel to slide creation. Theoretically.

Here’s what Katharine sent, she was trying to create a table for work:

The Copilot output provided a table, complete with coffee cups and five chairs. It also gave her a helpful title: Coherent Actions Table.

Microsoft may already be aware of issues like this. They have been pushing hard on security and privacy, even recently upping their bug bounty for Copilot - offering up to $5,000 if you find even a moderate vulnerability, such as a security or a reliability threat. But what about hilariously wrong (and completely unhelpful) results like this? Shouldn’t there be a prize for that too?

Have any AI fails of your own? Send them in - we’ll feature the best ones right here.