Deeply resonating - I've seen better outcome lately by clearly stating the overarching objective and asking open ended questions rather than prescribing specific solutions
Hi, I'm Felix. I'm an eng lead working on Claude. I love watching people take flight and do things they never thought possible.
We often describe that as "being more ambitious" when using AI, which is tricky advice to follow. I wrote a tutorial:
https://t.co/vNPPs5FWm8
#Muse just saved me $900+, I’ve been procrastinating to make a call to ask Dyson about this lid issue after only a couple of months in, Muse wrote the email, asked me to take some pictures and videos as proof, and got me a replacement the next day.
The time it takes for me to describe the issue, take the pictures, and review the email drafts, is probably similar to what I spent with Muse, but I haven’t done it for months, maybe the perception of AI agent saving time is just as real as the reality.
Hot take: AI is a bigger unlock for solopreneurs than for big tech employees. Yes, the AI boosts in productivity is real, but the real disproportionate winners are one-person companies.
Here’s what I’ve learned vibe coding lately: asking the right question is most of the work. And when you already hold all the context, the right question is obvious. You still debug, but prompt → generate → debug → ship gets super fast.
Good AI tools should not need training, they are intuitive enough that all you need to do is to just get started. Better products are almost always like this, you don’t need a user manual to use Claude Code, Tesla, or iPhone.
Went to an internal AI workshop, it was very informational and well prepared with specific instructions to do certain tasks. but I can’t help thinking it’s missing the point. The point of AI is to unleash creativity and try to test the boundaries of what it can do.