Something I keep noticing: new models and features ship almost weekly, but most of us (myself included) are still using AI the same way we did a year ago, mostly as a smarter search box or a drafting assistant.
The people getting the most out of it seem to be doing things differently. They break big tasks into steps, give the model real context, check its output critically, and build repeatable workflows instead of one-off prompts.
A few questions I’d love your take on:
- What’s one way you use AI today that you didn’t a year ago?
- Where does it still let you down, whether that’s accuracy, reasoning, or trust?
- Do you think the bottleneck now is the models, or how we use them?