Now that's a tremendous pointer, I'm going to have to try that.
Do you full on let GLM5 get stuff done on its own or is it more like a guided workflow? The former's what the point releases doubled down on and is also something that uses a lot of juice.
Ive been using openspec and let it do the whole spected out project until its done. I dont interact other than the initial proposal, then apply, archive steps.I often run several in parallel on the same project in opencode.
Now I see the issue clearly! But wait... now I have the full picture! But wait... Found it!
I gave up a few times because of it at first until I realized I just had to let GLM get on with it and what came out was great!
But once it was outright endearing- challenging bug, it said: I have been very thorough. Then it escalated where to look and aced it. Built in confucian values
If there’s one thing I’ve learned these past couple of days, it’s to resist the temptation to jab the escape button and start waving my arms! I wonder how much of this cyclical self doubt / self congratulating I go through in my own thoughts without even realising it. If you could verbalise or articulate all the half thoughts, snatches of ideas, feelings and ruminations the human mind goes through on some tasks it might be even more bizarre (or could just be me)
I'm like 90% sure the harnesses inject those tokens into the ai to make them check their work. Things like "but wait" and "but what if..." etc. Like literally inserting them artificially and then say "carry on from here" to the ai so it's working as though it itself output those words so the ai has an opportunity to turn around if its making a mistake. repeat a bunch of times and we get something useful.
I started noticing those in gh copilot right around when they turned off thinking traces end of last year
A command-line program for AI-assisted programming to get results as cheaply and reliably as possible. Especially interesting to use z.ai models, which will likely still work if US AI company prices rise too much.
It's also a pleasure to work with. Still needs polish and bugfixes, late alpha.