Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Won't be surprised to see all these hand-picked results and extreme expectations to collapse under scenarios involving highly safety critical and complex demanding tasks requiring a definite focus on detail with lots of awareness, which what they haven't shown yet.

So let's not jump straight into conclusions with these hand-picked scenarios marketed to us and be very skeptical.

Not quite there yet with being able to replace truck drivers and pilots for self-autonomous navigation in transportation, aerospace or even mechanical engineering tasks, but it certainly has the capability in replacing both typical junior and senior software engineers in a world considering to do more with less software engineers needed.

But yet, the race to zero will surely bankrupt millions of startups along the way. Even if the monthly cost of this AI can easily be as much as a Bloomberg terminal to offset the hundreds of billions of dollars thrown into training it and costing the entire earth.



My concern with AI always has been it will outrun the juniors and taper off before replacing folks with 10, 20 years of experience

And as they retire there's no economic incentive to train juniors up, so when the AI starts fucking up the important things there will be no one who actually knows how it works

I've heard this already from amtrak workers, track allocation was automated a long time ago, but there used to be people who could recognize when the computer made a mistake, now there's no one who has done the job manually enough to correct it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: