- AI with Kyle
- Posts
- AI with Kyle Daily Update 125
AI with Kyle Daily Update 125
Today in AI: ChatGPT Ads are coming
What’s happening in the world of AI:
Highlights
📺 OpenAI Launches Advertising in ChatGPT
The moment we all knew was coming. OpenAI announced they're adding adverts to ChatGPT free and Go tiers in the coming weeks. If you're seeing people feign surprised outrage on Twitter and Reddit - they knew this was coming. Anyone paying attention knew.
The question now is how sensitively will it be done.
Kyle's take: Let's be honest about what Google, Meta, and now OpenAI actually are. We think of them as tech companies. They're not. Google makes 60-70% of revenue from advertising. Meta makes 99% from ads. They're advertising companies that happen to use technology.
OpenAI has 900 million users pushing towards a billion. They need cash for an IPO rumoured at $750-800 billion - one of the largest ever. They are BURNING cash right now and need to show a sustainable business model. They also have a lawsuit from Elon Musk. Advertising was always the plan and they are rolling it now because they need money.
Source: OpenAI announcement | Fidjissimo | OpenAI tweet
🎯 What OpenAI Promises (And What Will Change)
Their stated principles: responses won't be influenced by ads, conversations stay private from advertisers, ads will be clearly labeled, and Pro/Plus/Business/Enterprise tiers remain ad-free.
They claim "responses in ChatGPT will not be influenced by ads." What a load of rubbish! The whole point of advertising is influence. That's what adverts are for. Yes, the response text might not change, but placing an ad in the same visual space is specifically designed to influence you. It’s a crap ad if it doesn’t.
"Your conversations are private from advertisers." Tricky language. OpenAI will build profiles of who you are based on what you're asking, then serve ads based on that data. Sure the advertisers don't see your chats directly, but they get access to you based on your data. Subtle but important distinction.
I pulled up the evolution of Google ads over the last two decades. Early on, they got in trouble for not marking ads clearly (like Yahoo did before them - which killed Yahoo). So they added bright yellow "Sponsored" labels. Over time those became green, then green outline, then just tiny black text. The ads got bigger while the labels got harder to spot.

Evolution of Google ads over time
Whatever OpenAI promises now will not stay static. When they need more money, the balance shifts towards advertisers. So it’s to expected that their opening salvo is very careful and position so as to not trigger outrage.
💰 Why ChatGPT Ads Could Be Insanely Valuable
Here's what makes this interesting. Most advertising exists somewhere on the intent spectrum: informational → navigational → commercial → transactional. Social media ads catch you early in the funnel. Google search catches you when you're actively looking.

ChatGPT has something neither has: your deepest secrets. Your fears and desires.
People talk to ChatGPT about medical problems, relationship difficulties, career anxieties, financial worries. They use it as a journal. They offload thoughts they wouldn't type into Google. This creates an incredibly rich dataset about who we are and what we want.
That means ChatGPT can potentially go straight to commercial and transactional intent. They know exactly what you need because you told them in conversation.
Sigüll ran the back of the envelope numbers (recommend checking the full tweet). Meta made $58 per user (ARPU, average revenue per user) in 2025 from ads. If OpenAI hits 1 billion free users and monetises at just 9% of Meta's rate, that's $5 billion in incremental revenue. At full parity ~$60 billion per year from ads alone.
And potentially AI ads could drive MORE than social media ads. A higher average revenue per user. Because of the greater “intimacy” of information they have. We just don’t know yet - but we’ll find out soon what AI ads are worth to advertisers.
Source: Signal analysis
📱 The $10M App Rebuilding Playbook
Story time. I needed to copy a long URL onto my TV. Basic but annoying problem. Didn’t want to type it out manually with my remote control because it was incredibly long and convoluted. So went to find an app.
The remote control apps on the App Store all have predatory pricing - basic free versions, then $3/month for simple features like keyboard input. Thousands of reviews. Making money from this nonsense. Gross.
I got so annoyed I built my own in Claude Code. Rebuilt all the functionality in about 15 minutes. Ran it through Xcode, onto TestFlight, working within an hour.
Greg Eisenberg published a playbook for turning this basic idea into a business:
Find apps people pay for but hate. Use Sensor Tower to find apps making $50-100K/month in stable categories - productivity tools, utilities, niche professional apps. Look for sub-4-star ratings with strong retention. That's a signal people need it badly enough to keep paying despite hating it.
Look for the obvious missing features. Search for single-dev apps doing $100K+ per month with ancient UIs. Target apps charging for features that should be free. Find apps with steady revenue but no updates in six months.
Rebuild and undercut. Give Claude Code screenshots of the app, describe the functionality, tell it what platform you want. It'll rebuild everything. Charge a one-time fee instead of subscriptions. Or give it away free.
Also, handy timing! Manus just launched app publishing - build your app, click publish, test on your phone via TestFlight. No wrestling with certificates.
The building is easy now. Distribution and marketing become the differentiator. Still need to do that - but at least the build phase is sorted!
Source: Greg Isenberg playbook | Manus app publishing
Member Question from Greg: "Does Google winning the AI race financially offset the loss of search advertising?"
Kyle's response: This is the huge question. Google makes hundreds of billions from search - they need to actively kill their own cash cow by pushing AI overviews, not knowing if generative AI ads can replace that revenue. Nobody knows the answer yet - so Google will be watching OpenAI’s foreay into ads very closely!
We don’t know what AI ads are worth yet. This is all speculative. But I think Google don't have a choice. Either allow search to die (as it will over time) or kill it yourself so you control the narrative. And then hope to God you can monetise AI later!
Member Question from Dave: "I pay $30 a month for an app that uses one single feature. Probably 2% of the app. Could I recreate it?"
Kyle's response: Literally take screenshots of it. Go into the app, screenshot everything, then give them to Claude Code and say "rebuild this." If it has a sales page or feature list, give that too. Tell it what platform you want - iPhone, Android, web app. It'll figure out the languages and libraries.
I did exactly this to recreate a TV control app out of spite. For myself so I didn’t have to buy a predatory app. You could charge $10 one-off instead of $30/month subscription. People hate subscriptions. You'd have takers!
Quick Note For Business Owners: When the ChatGPT ad platform launches, experiment with it ASAP. Every new ad platform - Google, Meta, Instagram, TikTok - has a window where it's basically the easiest money you'll ever make. Low costs, less competition, high returns. That window closes within a year or two as prices rise and competition increases. Move fast and see if you can take advantage of underpriced ad inventory as they find their feet.
Want the full unfiltered discussion? Join me tomorrow for the daily AI news live stream where we dig into the stories and you can ask questions directly.
