AI with Kyle Daily Update 171 Today in AI: Agent Manager Kyle Balmer April 15, 2026 For the last two years, most of us have been using AI the same way: open a...
AI with Kyle Daily Update 170 Today in AI: Claude Mythos Kyle Balmer April 14, 2026 https://youtu.be/OuK8ihDVEk0 I’m back! I’ve been relocating from the UK to...
The skinny on what's happening in AI - straight from the previous live session:
One day after California pushed through a slew of AI child protection laws Sam Altman announced ChatGPT will include "erotica for verified adults" within weeks. The context: California's new laws require age verification, so OpenAI's response is essentially "fine, we'll age-gate AND get spicy."
Kyle's take: Didn’t see this one coming!
This feels a little like calling the lawmakers' bluff.
Monday: "We're protecting kids!"
Tuesday: Sam announces erotica.
Wild.
OpenAI and other AI companies are being forced to implement age-gating (like UK adult sites requiring ID), which increases friction and costs.
So their response? "Cool, inside that age gate we're going buck wild." Remember when OpenAI was prudish about this stuff? Now they're following Grok's lead. T
Source: Sam Altman Twitter
Walmart partnered with OpenAI to enable direct shopping in ChatGPT. Users can browse, compare, and buy products with a "buy button" right in the chat interface. This follows similar deals with Etsy, Shopify, and Booking.com.
Kyle's take: This is definitely part and parcel of the advertising endgame. Once you have transactions on your platform, ads are the logical next step. They're testing integrated payments in India (1.5 billion people - not exactly a small test market).
When you search for laptops, expect "sponsored" options mixed in. Google makes 70% of revenue from search ads - that money's about to shift to GenAI.
My bet: we see ChatGPT ads by end of year, definitely by 2026. Free tier gets ads, $20/month reduces them, $200/month removes them. It's the YouTube model. Once one AI company blinks and adds ads, they ALL will within months. But nobody wants to go first and lose trust.
Source: Bloomberg
AI researcher Eric Hartford's $200/month ChatGPT account was banned for alleged "CBRN-related safety testing" (Chemical, Biological, Radiological, Nuclear weapons). After 500K+ views on Twitter, OpenAI reversed it. He denies all wrongdoing.
My account was completely deleted.
My appeal was denied with no consideration and no process.
all my history is gone.
no chance to export and save my data.
I am completely innocent of any wrongdoing.
I have absolutely no idea why @sama would do this to me. @OpenAI— Eric Hartford (@QuixiAI)
9:22 PM • Oct 13, 2025
Kyle's take: Whether Hartford's innocent or not is beside the point. He runs something called "Dolphin Distributed Inference Network - antidote to centralized tyranny" so take all of this with a hefty pinch of salt.
But here's the real issue: OpenAI can nuke your account instantly. No warning. No recourse. If you're building a business on their API, if you have clients depending on your ChatGPT connection, you're always one algorithmic decision away from losing everything.
Hartford got reinstated because he got half a million views. That won’t be the case for everyone. This is the risk of building on closed systems.
Source: Eric Hartford Twitter / OpenAI response
Today’s build was creating a quick tool to supplement a teaching session - attendees can use it during the talk to reinforce what they're learning. Built with Cursor in about 15 minutes.
I’m giving a talk at the AI for the Rest of Us conference tomorrow in London. Usually I’d guide attendees through an exercise like this:
But tomorrow I’ll run them through the same exercise using a tool I vibe coded. Speeding up the process and getting them all to the learning outcome faster.
Any teacher, trainer, or presenter can now whip up custom tools for their audience in minutes. Giving a talk on budgeting? Build a calculator. Teaching marketing? Create a headline generator. Running a coding workshop? Make a syntax checker. Instead of just talking AT people, you can give them tools to immediately apply what they're learning.
These can be tools you deploy with the attendees live during the talk OR you give them tools to use afterwards.
Previously this required hours of development and technical skills. Now it's a 15-minute conversation with Cursor or Claude. You spec it out in plain English, the AI builds it, you deploy and then throw the QR code on your slides. Done.
Maybe add email capture for your list? The technical implementation doesn't need to wow anyone - it just needs to work. That's the real power of vibe coding - not building the next Facebook, but creating purpose-built micro-tools that solve immediate problems.
Here’s a video walkthrough of how to build something like this:
Member Question: "Would you still use ChatGPT with ads?"
Kyle's response: They'll likely add ads to free tier, reduce them at $20/month, eliminate at $200/month - the YouTube model. I pay, so it won't affect me. IF they use this model.
But here's the thing: once one AI adds ads, EVERYONE follows within months. There's no safe haven. It's "who blinks first." ChatGPT doesn't want to go first because users will flee to Gemini. But Gemini (Google) desperately needs to shift search ad revenue to AI. It's coming to all AIs unless you pay.
Subscribe to the Youtube channel to watch the whole show, or catch it live.
Kyle
Make us a Preferred Source on Google and catch more of our coverage in your feeds.
AI with Kyle Daily Update 169 Today in AI: Lead Capture Machine Kyle Balmer April 01, 2026 https://youtu.be/cRsjmxoo64E I want to show you the single most...
