It sounds like most of the work went into getting the slideshow to work, setting up a feedback loop so Astra could work out / blunder into a solution with the desired characteristics.
It sounds like most of the work went into getting the slideshow to work, setting up a feedback loop so Astra could work out / blunder into a solution with the desired characteristics.
> I first tried GPT-5.6 Sol with reasoning effort set to Extra High. It wrote code to connect to the Kindle but couldn't resolve an error that prevented it from storing data on the device. Then I tried Fable 5.1 and Opus 5.5, but Anthropic's cybersecurity safeguards blocked each attempt after two prompts. Finally, I tried GPT-6 Astra with reasoning effort set to High. We were able to jailbreak the Kindle.
"We were able to".
It takes a modern AI human and 4 LLMs to do something that could be done by googling "kindle jailbreak" and following instructions - they have to be performed on Kindle anyway.
I sometimes catch myself saying "we" when prompting AI so it really doesn't bother me to see this language being used. But I concede it can sound jarring or hubristic.
I ended up installing LanguageBreak but it took many attempts and many many reboots of the kindle into demo mode which takes forever. Even now searching how to jailbreak an unregistered kindle Paperwhite of my generation (can't remember off the top of my head) top results point to using LanguageBreak.
Once I got WinterBreak2 installed it was pretty easy. The site mentions a lot of stuff about package managers that seem to not exist at all which was confusing.
Now I'm happily booting straight into KOReader using an init script that I wrote that is gated on the presence of a file in the USB storage area.
But just kinda randomly trying at various reasoning efforts reminds me a lot of hearing people just randomly replacing various libraries or code snippets with stuff found on stack overflow. No real theory of the system.
There's not even much of an excuse, cuz you can "just" sit there and type "why?" or "why not?" over and over into an agentic chat session and probably get things working quite quickly. Principled debugging is state space elimination for the most part!