@chribjel grok bot <> cursor <> super grok - pricing and usage needs to be made much more clear! its really hard to understand it at the moment.
Best way i have found so far to work with @bot and @cursor_ai cloud agents -
1) setup project on cursor
2) setup the repo environment on cloud agents
3) review cloud agent settings! (default models, bugbot, artifacts, etc).
4) ask grok bot and dr eggbot to create the eng lead for a project.
5) ensure that the lead always talk to that speciic cursor project. sometimes, it directly starts kicking off cloud agents from grok bot which I dont like. I want all project or repo work tied to that cursor project.
6) first task for grok bot eng lead is to setup codebase > static analysis > rules and bugbot > skills > style. as per @poteto guidelines
I have been running multiple projects this way and this seems to be the setup that works for me.
Separately, I am not yet sure but cloud agents sometimes seem to take ages even for simple tasks. Am still trying to figure out why.
And separetely once again, although I now have started to navigate the @cursor_ai IDE better. it still takes some getting used to. Most important that I bookmarked are the (a) cloud environment (b) agent runs list (c) bugbot pages.
a weird inversion with LLMs is the models improve faster than the tinkerers
when i see people with custom workflows and setups they're all addressing problems that don't exist anymore
the person naively using vanilla codex is more likely to be experiencing state of the art
@Tim_Denning fair, visible work counts more now. but credentials were mostly compressed trust: a stranger could read 20 years off one line.
that compression got weaker, so now you show the uncompressed version. experience didn't stop mattering, it stopped being readable without proof.
@mstockton If verifiable means solved, the bottleneck just moved to whoever writes the check.
In a real month-end close, nobody hands you the answer key.
@shaneparrish@BillAckman Most ppl run their life decisions the same way, minus the second buy. Leave a city or a job, watch it go well without you, and going back feels like admitting you were wrong. So they stay out. Anchored to the price they left at, not what it's worth now.
If making the thing is now cheap, it stops being what people judge you on.
Before: shipping was the signal. It proved you could do the hard part and so people trusted you with the next thing. Now anyone can show a finished-looking thing on day one, so the signal moves to day 90.
Bit like a house on open day. Looks done, nobody has lived in it yet.
what have you built with it that people still use weeks later, and how much of your week goes into keeping it alive vs starting new stuff ...
A lot of everyday deals assume both sides run out of energy. Haggling, complaining, courting, chasing a landlord. At some point someone gives in or gives up, and that cap on persistence keeps things roughly fair.
Now we're building things that talk like people, polite and patient, but never get tired. Persistence stops being a trait and becomes sth you can just buy.
@CoFoundersNik The thing is, streaming and cloud storage solve passive desires: escapism and memory. Consumer AI sells cognitive leverage, which very few people actually want off the clock. For the other 98%, it won't be a $20 sub; it'll just be invisibly bundled into what they already buy.
@signulll What they all share is that nothing is decided yet. Cart, hotel tab, restaurant list = a little pocket of options in a week where everything else is fixed.
People aren't really buying stuff tbh. They're renting the feeling that their life could still go another way.