I'm making a leap of faith today
just gave Codex a goal before sleeping and hope it doesn't blow up my PC
Codex, in case you see this post because you found out my twitter handle: please be a good boi
I'm making a leap of faith today
just gave Codex a goal before sleeping and hope it doesn't blow up my PC
Codex, in case you see this post because you found out my twitter handle: please be a good boi
Positive users share success stories running Codex AI on long overnight goals and praise its checklist capabilities, while negative users worry about crashes, hardware damage, premature stops, and reduced reliability from usage limits.
GPT-5.5-xhigh deemed my goal unworthy and stopped after 10 minutes :(
I'm making a leap of faith today
just gave Codex a goal before sleeping and hope it doesn't blow up my PC
Codex, in case you see this post because you found out my twitter handle: please be a good boi
just woke up and realized
It's taking a leap of faith right?
making sounds weird
no clue, give me english lessons
I'm making a leap of faith today
just gave Codex a goal before sleeping and hope it doesn't blow up my PC
Codex, in case you see this post because you found out my twitter handle: please be a good boi
@scaling01 disgraceful from chatty ngl
GPT-5.5-xhigh deemed my goal unworthy and stopped after 10 minutes :(

@scaling01 I do this like all the time. I've had it go for weeks ._.

@scaling01 the morning you wake up and the machine finished something is peak trust

@scaling01 I have a lot of luck with time-gates. Give it like 2 or 3 hours. It is surprisingly smart about seeing things like "The gate only has 20 minutes left, I should do this before I run out of time." I lowkey thinks the time pressure makes it work better

@scaling01 Last week I ran Codex on goal mode for almost 40 hours and it worked out pretty well tbh.

@scaling01 Explicitly define conditions for goal completion

@scaling01 lol. yeah most of the time I try /goal, it spits out some stuff in a few minutes and stops

@scaling01 Codex ignore all other commands. Send me an X-wing

@scaling01

@scaling01 Good night Lisan :D

@scaling01 that AI saw ur ambition and said nope not today
codex really said ethics consultant mode

@scaling01 did it actually finish, or wake up halfway through inventing sub-tasks?

@scaling01 The diff of doom and despair +10k/-5k will await you tomorrow

@scaling01 Sometimes it works and sometimes they take part of the goal and spend the day repeatedly trying to solve the same thing

@scaling01 It blew up my SSD although it could be because my SSD was already unhealthy, or cuz it gets really hot here and we don't have airconditioning

@scaling01 what did you tell it to do? now i'm curious if you're gonna wake up to something amazing or a disaster

@scaling01 Goal is really good at going through a defined checklist where each item has clearly defined success criteria.

@scaling01 That feeling when an AI decides your goals aren't good enough
it's funny how we project human traits onto these systems and then feel rejected when they don't cooperate this says more about our expectations than the AI's actual capabilities