AI and the Front End

Decoder’s latest episode is an interview with Microsoft CTO, Kevin Scott. It’s an interesting interview with a lot of information on AI but mostly about how AI and websites work together in relation to discover ability. I’d recommend giving it a listen.

There were 2 parts that stood out to me. The first was the host, Nilay Patel, speaking on the medium to host content.

“You know what’s interesting about that? I’ve asked a lot of people over the past several years, “Why would anybody start a website?” And the frame for me is when we started The Verge, the only thing we were ever going to start was a website. We were a bunch of people who wanted to talk about technology, so in 2011 we were going to start a website…Now in 2025, I think, Okay, if I had 11 friends who wanted to start a technology product with me, we would start a TikTok. There’s no chance we would be like, we have to set up a giant website and have all these dependencies.” 

This treats like where you host your content as some type of binary, so I understand the point he’s getting at. However I think some folks see this as an absolute truth. “Well we can have a YouTube channel, or we can have a website”. The benefits to a website, especially for a content business, is it’s the hub for all your content! The tradeoff to hosting content elsewhere is discoverability, obviously, but check out any instagram influencer’s page. There is a Linktree link. That’s their “website”. What happens when a billionaire loses his marbles and buys the website you’re hosting content on? It doesn’t have to be the main thing but owning your own space online comes with a lot of benefits and to say “YOLO just going to try and game an algorithm” leads to a lot of issues.

  • Your discoverability platform…is the same as everyone as your competing with
  • You share X% of ad revenue with the service
  • Perhaps those ads are ones you don’t want associated with the brand
  • You don’t own your content

I’m not saying YouTube channels are bad, I just think it’s a good idea to have a URL with those same videos.

The other piece I found interesting came from Kevin Scott.

“I think the thing that’s hard right now is I feel as if we’ve got this capability overhang with models where the models are actually capable of a lot more than what they’re being used for. And so, even inside of Microsoft, I maybe overestimated how quickly people were going to just lean all the way into the platform capabilities of the basic AI models.“

Recently someone asked me if AI had reached the point where it could take over a lot of jobs and if it was getting close. My answer was no. AI can output text, code, ok-ish images but you still have to make the decision to use it. And I think a lot of people’s interaction with AI is the “Staring in front of the ChatGPT/Copilot screen”. Maybe some people use it to brainstorm, as a search engine, or even help with code but the basic functionality is entering text into a box. Hey that’s how I’m writing this blog post.

These AI models are capable of so much but the front ends for them are limited. They are limited in the sense that the UI gives a bit of a “blank page” problem. That’s no one’s fault. That’s basically every new technology. Folks didn’t use email for marketing in the early days of email. We weren’t writing web apps when Web 2.0 began. Influencers weren’t a thing in the early days of social media (well not as we know them now). I think AI use cases would benefit from front ends that leverage AI in ways that are helpful to the user. Meeting notes and summaries are a good example. But I think if we want more user inputs from people using AI, we have to give them more useful applications than the blank box, and run with it. The possibilities are endless but so is the Netflix library.

I certainly don’t want to come across as negative. I think the points here are valid, but it made me thing about why things are the way they are, and how we can improve them.

Want to share this post on Bluesky? I made a button for that :).