can confirm this... it has its pros and cons, but I and my team have gen'd a ton of value by prototyping and doing a ton of PMF / demoing *so* much quicker, mock UX flows and etc -> then turning and minting prod-spec ourselves for consistent delivery

requires heavy alignment of biz<>eng vision tho!

heck yeah!! I was super touched by the game and its drop-dead divine styling & music when it first came out, I will be acquiring 😌🍻

oh my... I knew it was possible, but... 😳📥🙏

this is just how I normally work w/ my artist besties LOL, setting creative vision and letting the gang cook >>>> anything else

non-LLM practical AI/ML, like
- the use of deep learning on lidar and video sources to detect features (in a myriad of fields)
- convolutional nets in sparse spatial data, like intelligent mine planning
- ML in dense data for compute-intense physics and engineering
- the ML in CPU branch execution!

a PM (team)'s backlog(s) at this scale are so big that
- it'll be overlooked for eternity / not worth it
- "eh, another PM's problem"
- all the devs think workarounds are good enough (disconnect)
- distilled feedback for rev gen is louder than customer common gripes at huge scale
- and a few more

@gergely.pragmaticengineer.com great insights here ^^

they absolutely must've tore up a good amount of their observability infra, in the name of "cost savings" LOL