AI

Since my university days, I’ve been deeply fascinated by artificial intelligence. It’s been a recurring theme in my work, part curiosity, part obsession. At Superthread, that fascination has become foundational to how we think about the future of software.

We believe the way we interact with software is undergoing a fundamental shift—one that’s going to be massively augmented by AI. Not in a vague “AI-powered” marketing copy way, but in a real, felt, tangible way that improves the experience of using software every single day.

I’m a huge admirer of people like Geoffrey Hinton, Demis Hassabis, and Mustafa Suleyman. The world of transformer-based AI that they helped create is astonishing. And yet, I don’t believe this approach will get us to superintelligence. That might sound odd coming from someone who uses AI tools over a hundred times a day, but it’s precisely because I use them so much that I know their limits.

That said, the current generation of large language models (LLMs) is more than enough to change the world. In fact, it already is.

The “Shut Up and Take My Money” Use Cases

As of summer 2025, there are a few AI use cases that are so effective, so transformative, they don’t need much selling. One of the clearest? Coding.

I can say from experience: AI-assisted coding works. It works very well. It saves time, reduces cognitive overhead, and lets you focus on the real problems instead of the boilerplate. For many developers, AI is now part of the flow—just like Git or Google used to be.

But when it comes to SaaS, the transformation hasn’t hit as hard. Not yet.

Everyone’s trying to figure out what the next paradigm looks like for SaaS. Including us. We think about this all day, every day. It’s as exciting as it is frustrating—because while the tools are powerful, the use cases aren’t always obvious. Not everything gets better just because you add AI to it.

But we haven’t waited around hoping inspiration strikes. We’ve already built and shipped AI features that actually work—quietly, usefully, and without fanfare.

Here is what We’ve Built So Far..

Transcription Templates

Our AI meeting notes aren’t just transcriptions—they’re structured, high-quality summaries that you can configure with your own templates. We ship with several, but you can copy and tweak them to suit your needs.

Chat With Your Workspace

You can chat with your meeting notes, documents, and even your entire workspace. It’s like having a live assistant who knows everything that’s been said and done.

Text Rewriting

One of our earliest AI features, and still one of the most loved. Whether it’s simplifying, clarifying, or rephrasing something—rewriting just works.

Internal AI Agents

We’ve had agents running internally for a while. They’re promising. Very promising. But the truth is, transformer-based models are still probabilistic. They can be brilliant—and they can also hallucinate. So while we’re not ready to release them publicly yet, we’re doing everything we can to get them there.

What’s Next?

We’re not chasing hype. We’re not shipping features just because they look cool in a demo. We’re focused on building AI that earns its place—because it actually helps you do your work better, faster, or more clearly.

We don’t know exactly what the new SaaS paradigm will look like. But we do know we’re going to help shape it. And we’re taking it one useful feature at a time.

David (CEO)