how to update without gambling your working stack, what to verify before and after, and the fast path back when a release knocks out channels, config, or runtime state
Absolute gold that applies to any software. I implemented the similar model for my CareerOS plugging on claud code and every other thing I build; however, I use git as a cheap backup and using logs as documentation 😉!!! Details coming soon.
I love your content but can you use CAPS on sentences please? It might be some new trendy design choice but it outright makes it harder / slower to read and I can’t pay for it on the aggravation of reading your text. Otherwise, be done with it and do away with periods, commas, quotes and dashes too - the text will loose all structure and scanning handles but it will look so trendy.
something I’ll consider. a lot of it is pass over from my iPhone notes on my phone. valid concern I guess, I’ll run a survey maybe and see. appreciate you being here. my goal is to be as helpful as possible in this format. thank you for addressing it.
I use a mono thread in GPT-5 now. I throw the release notes in there and then pass a gameplan over to my main. It seems like your main doesn’t just magically know what it’s capable of after an update, imagine that!
The easy way is to ask Claude code to do the update and then after the update run a full diagnostic, then Claude will walk you through each item and then fix it for you. It’s the best way
easiest but not the best. lazy and reckless for real ops. I’m not saying it can’t be a part of it. if you give temporary machine full access to the model then yes this can catch the low hanging fruit.
Absolute gold that applies to any software. I implemented the similar model for my CareerOS plugging on claud code and every other thing I build; however, I use git as a cheap backup and using logs as documentation 😉!!! Details coming soon.
I love your content but can you use CAPS on sentences please? It might be some new trendy design choice but it outright makes it harder / slower to read and I can’t pay for it on the aggravation of reading your text. Otherwise, be done with it and do away with periods, commas, quotes and dashes too - the text will loose all structure and scanning handles but it will look so trendy.
something I’ll consider. a lot of it is pass over from my iPhone notes on my phone. valid concern I guess, I’ll run a survey maybe and see. appreciate you being here. my goal is to be as helpful as possible in this format. thank you for addressing it.
Thank you. A grammar check ought to pick it up, or openclaw I assume.
you got it🤝
Omg! I didn't even notice the lack of caps. Was more distracted by --these things. I'd rather —I mean, just use an emdash.
Kinta that’s actually part of a code command so it has to stay there for accuracy when people copy and paste 😂 appreciate that though.
I use a mono thread in GPT-5 now. I throw the release notes in there and then pass a gameplan over to my main. It seems like your main doesn’t just magically know what it’s capable of after an update, imagine that!
The easy way is to ask Claude code to do the update and then after the update run a full diagnostic, then Claude will walk you through each item and then fix it for you. It’s the best way
easiest but not the best. lazy and reckless for real ops. I’m not saying it can’t be a part of it. if you give temporary machine full access to the model then yes this can catch the low hanging fruit.
I guess it depends on what you call real ops. If you’re running openclaw in a real production environment you are braver than most