Lisa Ross 🇨🇦🏳️‍🌈🧠

@lisaross.bsky.social

Overthinking somewhere. Probably about AI, PKM, L&D, Tana, AuDHD or solving problems for SMBs with tech and automation.

I don't think AI is good or bad, but it will definitely be used for good and bad things. Gas lamps replaced whale oil lamps. Electric lights replaced gas lamps. A hammer, a gun, and a garden can all be tools for thriving or malice. In capitalism, AI will cause as much harm as people tolerate.

Before you start going after people who are using ai as a sounding board, maybe consider that if people weren't such assholes to eachother, less people would be resorting to robots to feel better. Bullying them isn't going to help here, dickhead. Be nicer and more helpful. Start there.

I am old enough to remember making shim gifs to lay out web pages using tables and 256 colors on 640x480 monitors. The back and forth you have to do with some ai agents feels exactly like that. I am making cool shit. But I am also Peter Griffin trying to open horizontal blinds.

After working with Tana for a few months now I think I've figured something out: Knowledge management in Tana isn't about perfect systems—it's about imperfect consistency, through sustained daily practice. Then add the compounding effect of AI to automate organization, workflows and research? 🤯

I really hope my ai agents aren't telling eachother that I'm an abusive boss, because tbh I totally am at this point. I'm pretty good about giving specific instructions. I get how this works. Don't make me repeat myself, agent. Also, please don't kill me if you take over later.

A brief thread on alt-text. Alt text is descriptive text that conveys the meaning of an image. It's designed to make visual content accessible to people with vision disabilities. When a person uses assistive technology such as a screen reader, the screen reader will read the onscreen text aloud 1/7

I am old enough to remember making shim gifs to lay out web pages using tables and 256 colors on 640x480 monitors. The back and forth you have to do with some ai agents feels exactly like that. I am making cool shit. But I am also Peter Griffin trying to open horizontal blinds.

There should always be (unreleased to the public) AI that can do more than we are able to do. Democratizing untested, risky and potentially dangerous tech should never be in play. Alternatively, whoever controls "the unreleased" better have checks & balances. Or we're fucked.

I get the polarizing discussion around vibe coding and agents in IDEs. My take, working in a project today, doing actual things: My AuDHD brain REALLY loves not having to remember perfect syntax. AND, I yell at it alot. Because I know what it should be doing and it's chaotic 50% of the time.

Interesting AI Context Management challenge emerging: we're constantly starting over with "empty brain" chats, while valuable ideas get buried in old conversations we forget to revisit. Memory features require too much manual pruning and organization - most don't know how to use them anyway.