Generated by Codex with GPT 5.6 Sol High
offer at Robinhood got rescinded yesterday (Blind)
- jaсob (Snap): Incoming layoffs.
- Glinda: Hopefully, you did not give notice to your current employer, if you are working.
Was the req cancelled? - dhjelle (Roku): That’s muck’d up
Uber’s culture compared to meta (Blind)
- JHqQ30 (Meta): Any place will be better than Meta
- LWPu88 (Uber): More stable and less chaotic than Meta. Very top to bottom down culture. Refreshers depends on how much your manager likes you.
- tUvdgggg (Uber): Believe me uber is 2-3x Amazon. Don’t go to that place
Is SWE really dying? (Blind)
- kDRv27 (NVIDIA): whoever says software engineering is dying (because of LLMs) doesn’t really understand how things work (the populace including incompetent engineers). either that or they want you to think that way (marketing/manipulation masters like Anthropic/OAI or their investors). the vast majority making noise out there is in those 2 camps. the job is not being automated. the tasks are. it may be almost plausible to say ‘coding is automated’, but that’s a small portion of SWE. as you realized yourself, the SWE job is far far far away from being automated. we just no longer need silly people that take all day to refactor a function that will be obsolete 2 months later.
engineers are moving up the food chain and getting larger scopes of responsibilities. of course it doesn’t automatically mean they’ll thrive… you’ll need more product, social, business skills (all the things many ICs didn’t really care about) to succeed.
if you like the craft and you care about it, cut out the noise and keep going. couple of years from now you’ll have extremely powerful tools and your judgement. - TheFrom (Apple): China AI bubble just burst. People are reducing spend. It turns out normal people aren’t using AI as much as their biggest customer (software engineers).
AI still programs in a very human way in the sense that they’ll program in the literal way you want it to. The ambiguity from business requirements is still there. The edge cases you can think of or the common pitfalls that you know how to avoid are still valuable. It’ll just implement the design not tell you it hacked around just to get to it, are you willing to sign off on that? - gotPIP (Amazon): I think the concern is that it just keeps getting better and better bro
Why are people are acting like the AI bubble has popped, after like 2 bad days (Blind)
- DeGenAI-1 (NortonLifeLock): There is no evidence to prove AI has increased productivity
- qupwpphxv (ByteDance): This is just beginning
- menhhshwjw (Salesforce): When you know you’re part of a Ponzi scheme people who sell first, win. No one wants to be left holding the ai ponzi bag. Most likely this all ends in a taxpayer subsidized bailout
Did your code base improve after AI? (Blind)
- F*ai (Cohesity): Problem is we are pushing code way more than we can understand
- knowshits (Amazon): Nobody cares. As long as it works.
- tUDw24 (Meta): Much better. All of the “bugs” I’ve seen is people not pre-empting edge cases via prompting. I don’t think I’ve seen AI genuinely write flawed logic or hallucinated code in 4-6 months at this point and before then it still wasn’t a common occurrence past around sep 2025.