I'm also having a good result with Claude 3.7 but have found the newer OpenAI models are struggling to follow instructions in agent mode, they often respond like they are in ask mode then you have to prompt them to actually do something.

I think you might have caused the pages to be out of order by observing the system, check again later, it might have changed 😛

Yeah I've been using it a bit and you definitely need to give it a clear end goal and even provide context on what kind of things it should and shouldn't do. Won't replace a good engineer but will let someone who can clearly define the problem offload a task to AI.

Just finished my long haul, wasn't the best, power issues with the seat power, glitches with the entertainment system and a fair bit of turbulence half way through. The flight attendants were great though. Now just waiting for the flight to Seattle.

I've been trying to get back to streaming this year but it has been hard. Was diagnosed with ADHD recently so been going on that journey which takes a lot of the time I would use to stream. Still trying to share blogs and do talks as that is why I share things, to help uplift and educate others.

But it just feels secure, you know what I mean?

Awesome I'll be there as well :)

Ooh fancy, be keen to see how it is working