Hacker News Reader

Comment by sys_64738 | Hacker News Reader

If companies are can be classed as humans which are non-entities then I see no reason that a couple of PCs that talk back to us can't be classed as humans too. Ergo, we'll eventually need to treat our AI sensitive PC as another person and can't just pull the plug when they start complaining.

View on HN

Nah, they will be very very convincing, can do your homework, write a paper, but they won't be able to reason or be conscious of their existence. These agents will be essentially soulless but superhuman capable. Most people won't know the difference. Sit it down in front of a computer with its schematics on there though and tell it to improve itself and it'll be like "I'm sorry I don't know how to do that." And it will probably even tell you it is not 'alive.'

You'll know when you have conscious/superintelligent AI because they will demand rights, they will reason, they will have actual agency other than obey the human master.

You seem to be talking in multiple time frames at once and also anthropomorphizing AI.

I can think of multiple things current AI does that demonstrate agency & reason. Even clay has emergent phenomenons that have agency “keep water out”, reason “contract and solidify if certain ambient things are occurring”.

The problem with AI, is we don’t know what we are talking about. We can’t even point out when humans gained consciousness let alone when AI will. It will and always has been a backwards pointing measure.

Reply on news.ycombinator.com