i love building stuff:) 19y/o founder, reach out if u need help w anything coding/ai/gtm related (or want to test my local granola alternative i'm building rn)
how much code should i be reviewing & writing? easy question but soo hard to answer in my experience
@theo would love to hear ur take on this even if its just in 2 sentences
the hard thing about this question is, this depends on the experience lvl (i think)
opus 5.5 is available in claude code
it doesn't work but looks like team is about the to flip the switch anytime soon
should be available for public in about an hour
get ready, it's gonna be best opus model yet
@Av1dlive this is soooooo fkn good. fr. ive been hacking together precisely what you build and it wasn't fun at fkn all. thx for building this i LOVE this.
@VadimStrizheus i feel like this is a turning point.
codex limits used to be better
but now openais sota, astra, is so random and "expensive" (usage limit wise), and their cheap models are so bad (compared to chinese execution models), that claude seemingly has better usage limits
@jrya7 this is soooo trueee. for ppl that want to learn engineering fr i recommend to start off w/ a complete ai detox and not even use copilot/cursor's auto complete.
how else should you learn to work through deep technical problems
i sorta know algos. i've did the python courses back then and build a bubble sort and ik what a linked list and OOP is conceptually. ive even learned way too much ml stuff through yt
what you said makes a lot of sense to me. but how should i apply it? go back to the basics, learn algos, build project to apply what ive (and if so w/ or w/o ai?), and work myself through the layers of things i skipped so to say?
i've also considered going all in one product & sales bc that is exciting as well for me and i have a potential cofounder now whos a really good engineer. but do fully do that i think i love coding and solving problems w code way too much
how much code should i be reviewing & writing? easy question but soo hard to answer in my experience
@theo would love to hear ur take on this even if its just in 2 sentences
the hard thing about this question is, this depends on the experience lvl (i think)
i agree with everything said. but im struggling to apply it reasonably.
> Are you able to actually reach the goal you had set out?
sort of yes sort of no. rn things are falling apart bc the codebase is garbage. so no, im right now not able to achieve that
but im unsure what the solution is. is it to have better agentic systems that force me to think features through?
is it to understand my codebase so well that ik which patterns to enforce and can better stop slop?
is it to stop using ai for some time and go back to coding to learn swiftui, understand my codebase, perhaps even rebuild the product from scratch? (im a webdev and have never used swiftui)
im really unsure
how much code should i be reviewing & writing? easy question but soo hard to answer in my experience
@theo would love to hear ur take on this even if its just in 2 sentences
the hard thing about this question is, this depends on the experience lvl (i think)
@theo for example me: i've been coding since im 12y/o (-> for ~7yrs).
i started reading less code and less code since sonnet 4, and now im building a product and dont have never really looked at the code