agent loops that never terminate are the new infinite while loop
spent an hour debugging why my agent kept retrying the same failed step forever. added a hard step budget and a no progress detector. please put a leash on your loops.
the security side of this genuinely scares me
great, now I have to rewrite everything again
does this work offline or does it need an api key
this aged really well
did this break for anyone else after the last update?
what is the failure mode when the tool call times out?
works on my machine, famous last words
i would be careful recommending this to beginners
ok now do the version that handles errors
this is the third time this week I have seen this come up
the eval first mindset is underrated, nice to see it here
this is the third time this week I have seen this come up
did you try giving it fewer tools? helped us a lot
agree with the conclusion, not the reasoning
ok this is actually clever, well done
stealing this approach for work, thanks
i think the spicy take is actually correct here
this is the third time this week I have seen this come up
the moment you add memory this gets way harder, fwiw
this matches my experience almost exactly
skeptical but bookmarking to test friday
how are you handling auth for the tool calls?
this is super helpful, thanks for writing it up
i tested this last night and it mostly held up
agree with the conclusion, not the reasoning
tried it for an afternoon, not sold yet
underrated post, more people should see this
my team is going to hate me for sending them this
i was just about to ask this exact question
saving this, exactly what I needed today
great in theory, messy in practice from what I have seen
the moment you add memory this gets way harder, fwiw