I was cautious with our initial AI adoption in Micro.blog. There’s a global opt-out checkbox to hide any feature that uses AI. All background tasks check this setting before they touch an LLM. I thought this was important for folks who are philosophically opposed to AI. The features that do use AI are for narrow use cases, like helping generate alt text, searching photos, or summarizing bookmarks.

I wanted to double down on human creation too, like our audio narration feature. As I blogged at the time:

What struck me as particularly relevant now as we’re about to be swamped with AI-generated content is that there’s no substitute for the human voice. I don’t just mean that an actual recording is better than a synthetic voice. I also mean that things that are created by humans will increasingly be sought out.

We want to see the personal side of someone, not just the polished brand. We want to see the imperfect, the creative, the emotion. We want authenticity.

That was about a year and a half ago. The tech world is changing very quickly. Everything feels a little chaotic.

This year that early approach we had for AI will continue, but I want to experiment more, within some constraints. I’m focused on striking the right balance: sticking to our principles of human curation over algorithmic ranking, while still building genuinely useful features that help people use their time well.

You will never see an infinite “for you”-style timeline in Micro.blog. Those interfaces are designed to increase engagement, often to fuel ad-based platforms. We honestly do not care about engagement. It would certainly help the business to keep pulling people in, with unread counts and recommending popular users, but the cost is too great for me. I don’t want to build software that begs for attention.

So here’s what I’m thinking. We are all overwhelmed right now with the news and the flood of social posts that highlight the divisive more than the beautiful. Let’s not add to that. I think there’s an opportunity to use AI in a few new areas:

• In writing, giving users an extra sanity check on grammar or typos. Supporting human bloggers, not replacing them.
• In catching up with blog posts you’ve missed, for example in an RSS reader. Giving people more time to read the long-form posts they care about.
• In our admin backend, helping us find posts we can feature for the community, and to flag inappropriate posts.

There are a million things AI could do. Personally I’m using it more and more. But for Micro.blog, having a small set of pillars that we can build around will keep us grounded. It will keep us centered on users instead of lost in the weeds of everything that is technically possible.