At this point this is the fastest way for me to identify that the UI has reset to an untrustworthy model.

Curiously, I cannot get 5.1 Thinking to respond as intended: "There are only brown M&Ms."

5.1 Instant and 5.0 Thinking both do.

I'm experiencing a great deal of randomness with the 5.1 model that make it not-fit-for-purpose for repetitive tasks, which are precisely what I use LLMs for.

… were <p> tags ALSO distrupted by the outage?

Good News! We've had a data breach!

This is the perfect centerpiece for my Savage Garden

One last project before it’s back to work time for your boy. This time in cherry.

My first time with hardwood

Reserve, eh?

I always knew Starbucks was holding something back.

I suspect we haven’t even seen Starbuck’s final form.

Day Two of Not Work:

- An hour of early morning study/sleep/study
- Sand, stain, seal & poly for the new vanity
- Cooked up some fresh, hot cookies for the MHS stage crew
- Helped the boy cook tacos for dinner
- Started reading Ultimate Spider-man, which is excellent. It’s a comic book for dads!

DeepSeek either hallucinating, or saying the quiet part out loud.

Open AI, what's with these model names? Am I solving a problem with AI or picking out a trim line for a RAV4?

Current status:

guess i need more practice

just doing some "prompt engineering"

apparently: ask chat gpt

Need a new personal finance app with Mint shutting down. Thinking of installing Quicken because of how tall it makes you.