Some things that excite me about AI for knowledge work:
Tools like Google Antigravity, ChatGPT Codex, and Claude Cowork are giving Markdown and Text Files new lease on life. Clean, structural, hierarchical text is the perfect interface for both human readability and LLM ingestion.
Taking important context from across the web and making it useful (anyone still using Goodreads?)
It’ll be interesting to see how companies balance openness with the desire to own customer data. If I’m being (perhaps naively) optimistic, perhaps a token-based economic model and bring-your-own-data functionality can re-center power back with the user. A special shoutout to the folks at Are.na for being a shining example here.
I’ve geeked out about productivity systems and methodologies (Zettelkasten, GTD, 43 Folders, 2nd Brain, etc) for most of my career, and it feels like AI helps get us closer to the ultimate vision of human productivity than we’ve ever been.
“Soul files” as a way to help one understand and interrogate their own beliefs. I’m betting we will see the commercialization of digital identity frameworks within the next 2 years.
It has reduced SO much of the mindless and low-value work that knowledge workers spend a lot of time doing - summarizing meetings, cleaning up Google Sheets, cross referencing multiple versions of the same material, copy-and-pasting content from one place to another, etc.
Interestingly, it has more neatly cleaved a lot of my work in 2: AI Work and Human Work. In the context of work, if I know that an AI can do a pretty good job of transcribing a meeting and writing a summary, I can feel more present in a conversation with another person.
If you haven’t already been thinking about how to “make the implicit, explicit” for the purposes of your individual or team’s AI tools, now’s a great time to start.













