AI will be like spreadsheets in the future, we’ll all use them and not talk about them, but right now we’re like, dude check out what’s in my spreadsheet
@Mark_Goldberg_ been doing this since the new year! you hook up a couple api’s to a cron job to make a report with claude and execs are like holy shit 🤯 most fun i’ve had at work in years!
i structure it as a 1 hour per week meeting for $3k per month and those who want it don’t even blink
@Shpigford i installed enso first, got it running, and then told it i wanted it to import all the skills, jobs, and memories from my openclaw agent directory
@teddy_riker@hnshah the irony with salesforce is that the experience is so convoluted for humans that pivoting to agent-first might actually make people use it, since the UX will improve dramatically if its just claude you’re interacting with
yea, I’ve run a single agent overnight a few times, and it only works if I have a list of github issues clearly outlined on a 0-1 software project, and use a tool like harbour to cron their wake-up every few minutes
the output has been like 70% satisfactory, as in it usually does a few tickets poorly, and i always think, “i’ll need to redesign that later.” but it does feel like magic to wake up to a bunch of finished issues.
i know @geekforbrains is running small agent “teams” for various companies for customer support, bug fixing, ad spend optimization, and software iteration. he’d be worth chatting with if you want to hear no bs approaches to this stuff