Kevin C. Tofel

Apparently my AI assistant wasn’t too happy when I changed his brain to use a local model instead of one from the cloud. Cranky!

I wish there was an official #Wordle #API. I would love to make a free web app that lets two people do a monthly challenge: Whoever has the least total guesses in single month wins.

Instead, I have a Google Sheets setup for me and a friend with total guesses for the month, average guesses for the month, and a few charts. I have to manually enter our scores each day. 🙄

(Yes, I know I will be losing badly this month!)

Been kicking the tires on #OpenClaw for the past few days. Dedicated PC with a fresh Linux install on a separated home network. I’m pretty open when it comes to testing things, but even *I* have been playing it relatively safe, or at least as safe as one can.

I’m curious what people who have *used* it actually think.

Yes, there are security issues. No, it doesn’t run properly 100% of the time. But I see the potential of it, or something more refined like it, that does fundamentally change personal computing. Your OS and computer specs may not matter if you have a cloud-based AI that can complete tasks you ask it to in plain English. (I am looking at running local LLMs for this, however.)

I have more thoughts, but want to hear from other users.

As context, I have it set up to scour the web and create a daily briefing with new, relevant articles and videos on a few topics that I selected. I also connected it to the Copilot Money app I use to track my finances, mainly because that app can’t take action on any of my accounts; it simply reads data.

In terms of the scary part: I asked it how much free drive space was on the PC it’s running on. Got it spot on. I also asked it to test my internet speeds from that PC. It ran a basic test and then offered to test using the Ookla Speedtest CLI. I said yes: It downloaded the CLI, installed it, ran the test, returned the results, and uninstalled the CLI. Yeah… scary. But useful. 🤷🏼‍♂️