AI with Kyle Daily Update 017
Vogue Features First AI Model, The Human Turned Chatbot, How Safe Is Vibe Coding?
Watch this video on our dedicated watch page
Better viewing experience with related videos and full-screen player
The skinny on what's happening in AI - straight from the previous live session:
Highlights
📰 First AI Model Hits Vogue (Sort Of)
GUESS used an AI-generated model in a Vogue advert - first time artificial intelligence has appeared in the fashion bible. But the AI generation is absolutely terrible, looking like something from a year ago, not the photorealistic stuff we can do now. Weird.
Kyle's take: This is getting coverage because it's Vogue, but honestly, this is going to become the norm very quickly. The economics stack up - why hire photographers, models, and studios when you can generate product shots for pennies? The scary bit? The bread-and-butter work for models, fashion photography and crew isn't high-end runways, it's stuff e-commerce product photography. That's getting wiped out first, the bread and butter pay-the-bills work.
Source: BBC News coverage
🤖 Meet Geraldine: The Human Turned Chatbot
Peterborough City Council turned their go-to staff member Geraldine into a chatbot called "Hey Geraldine." After 35 years of answering everyone's questions, they've digitised her knowledge so staff can get instant answers in her "chatty and direct manner."
Kyle's take: This is brilliant for Geraldine personally - she hasn't lost her job and gets to be digitally immortal. Cool. But here's the darker side: once we can extract institutional knowledge from employees automatically (from emails, meeting notes, etc.), we won't need those people anymore. This manual process is just the beginning. Right now it’s a cute puff-piece about Geraldine but this is a much bigger story.
Source: BBC News coverage
Member Question from Has One: "How safe and secure is Lovable?”
Kyle's response: Lovable itself is secure, but the problem is vibe coding lets people without security knowledge build apps. If you don't know how to instruct Lovable about security, you'll create vulnerable apps. Use Supabase integration for the backend - that’ll help handle most security issues with user details etc. But if you're going to production with real (paying!) users, pay a security expert to audit it properly. Don't cheap out when you're handling customer data.
This question was discussed at [11:52] during today's live session.
Member Question from JJ: "How do we learn to vibe code?"
Kyle's response: Just start! Go to Lovable.dev, make a free account, and type something you want to build. Don't watch tutorials yet. There’s no need really - build a simple calculator first. When you hit problems, take screenshots to ChatGPT and ask for help like you're talking to customer service. That's literally how you learn - by hitting friction and working through it. Work? Yes for sure. But way less work than learning proper coding, trust me.
BTW: not sponsored by Lovable! It just happened that we had lots of questions about it today!
This question was discussed at [21:00] during today's live session.
Member Question from Grace: "If only elite labs can afford reasoning models, global south and open communities are left behind."
Kyle's response: Spot on, and it's not just the global south - normal people are getting left behind too. We're seeing the most powerful technology in human history concentrated in the hands of five or six billionaires. That's dangerous.
What can we do? If you use AI, you've got an obligation to teach others. We can't let a handful of people control this entirely. We need to keep up.
This question was discussed at [1:12:00] during today's live session.
Streaming on YouTube and TikTok (follow and turn on notifications for Live alerts).Want to submit a question? Drop it below this video and I'll cover it in a future live.
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.