Comments by "xybersurfer" (@xybersurfer) on "" video.

i'm a programmer. these Artificial Neural Networks are a black box, and the way they work internally is still being researched. the problem with models like ChatGPT is that they tend to make mistakes, and therefore need hand holding by someone that can recognize those mistakes. i think its success depends on the job it's replacing, and how many mistakes that job can tolerate and whether it doesn't take too much effort to fix its mistakes. maybe they will slowly creep into more important jobs as they get better, with more training or something else to prevent it from making mistakes or maybe the public has to wake up to the limitations. it's really hard to tell. the attention these models have been getting will probably accelerate research regardless
72
they probably have dreams that transcend keeping their day job (as do i)
7
@sidecharacter7165 i also feel uneasy about them tampering with these systems
2
@JamesJNothingIsTooSensitive why is it so wrong to speaking of limitations?
1
@JamesJNothingIsTooSensitive i get where you are coming from and i even agree. i also don't like people that keep speaking of limits as an excuse to avoid putting in the work. i only meant the limits of the current systems like ChatGPT or GPT-4. i don't think i would never claim that humanity can't surpass these limits with another system (that's not a world i want to live in). i think that recognizing limits in specific systems is the first step to surpassing them. i have doubts about the these LLMs and the way they are used, but i'm generally extremely optimistic about what can be achieved. the sky is the limit
1