Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Humans are underrated

fortune.com

Post reply on HN
Print view View original

41–50 of 50 posts

Re: Humans are underrated

#41

Post by Trombone12 »

anon4 wrote: post #34
lordnacho wrote: post #7

The article touches on the major areas: - Jobs where we need someone to be responsible. You may have fancy diagnostic computers and automated scalpels, but you are going to want a doctor somewhere. A NLP algo might be faster, more accurate, and better at recalling old cases than a judge, but you're not going to be satisfied in a society where you open a website to find out whether you're going to jail. - Jobs where i…

> responsible That part is kind of weird to me. I understand you want someone to be "responsible", but what do you expect the response to be? This doctor failed to deliver, let's fire himher. What follows? You get a better doctor, hopefully. Wouldn't it be better to have an algorithm that can be improved when it fails, so you don't have to train a whole new doctor to replace the bad one? Plus, you can have every sing…

It would indeed be better if the doctor would get better every time it fails, but for that to happen it requires that the outcome of every treatment be the responsibility of the makers.

Consider software security, how have they historically handled patching known vulnerabilities?

Now, imagine a bug making these millions of doctors cut an order of magnitude too deep when performing some surgery, maybe because some driver error causes sudden acceleration of the cutting instrument. Forty people die on a patch Tuesday (it only happened for a rare procedure) before the makers manage to push out a hotfix. What should happen after this national hiccup?

Re: Humans are underrated

#42

Post by loup-vaillant »

lordnacho wrote: post #39
aaron695 wrote: post #22

Earlier quoted context omitted.

> The oldest profession is not going to be replaced with robots until those robots are mistakable for humans It already has. It's the internet. Sex crimes go down when internet access goes up. It hasn't 'replaced' it 100% but that's the point with a lot of this stuff it is slowly reducing jobs without creating new ones. I really don't think it'd be impossible to create a machine that cuts hair. Or making coloring hai…

A lot of people are not going to the hair salon just have their hair cut. Having a chat is just as big a deal for many people.

Many people. Not all. Not most. Just many.

That won't be enough to preserve those jobs.

Re: Humans are underrated

#43

Post by Trombone12 »

marincounty wrote: post #36
icebraining wrote: post #10

Earlier quoted context omitted.

If the stats were significantly better, I'd take the completely robotic "doctor" any day. Why would I want someone to be able to override what is statistically the most accurate decision? It's not like they'd be negligent if they followed the computer's decision, and if they didn't, being able to sue them doesn't make up for the fact that I could have been better treated. The article mentions we'd prefer to have a do…

At this point, I would welcome a completely robotic doctor. Could you imagine, a Computer Doctor says, "I have throughly scanned every study, ever produced on SSRI's, and I don't want to make your life more difficult by a placebo response? I've made a calculated decision, and decided you might benefit from this off label drug. Try this drug for a few days? If it doesn't work, or you are skeptical of my recommendation…

What you are thinking of is a perfect AI running bug free software. You need to consider the more realistic situation that your robot doctor is using a decision tree constructed by Microsoft, and its actuator firmware are from Toyota.

Re: Humans are underrated

#44

Post by minthd »

Trombone12 wrote: post #41
anon4 wrote: post #34

Earlier quoted context omitted.

> responsible That part is kind of weird to me. I understand you want someone to be "responsible", but what do you expect the response to be? This doctor failed to deliver, let's fire himher. What follows? You get a better doctor, hopefully. Wouldn't it be better to have an algorithm that can be improved when it fails, so you don't have to train a whole new doctor to replace the bad one? Plus, you can have every sing…

It would indeed be better if the doctor would get better every time it fails, but for that to happen it requires that the outcome of every treatment be the responsibility of the makers. Consider software security, how have they historically handled patching known vulnerabilities? Now, imagine a bug making these millions of doctors cut an order of magnitude too deep when performing some surgery, maybe because some dri…

The safety-critical software industry is very different from your example, and in general, the results are very good.

Re: Humans are underrated

#45

Post by pnt12 »

Trombone12 wrote: post #43
marincounty wrote: post #36

Earlier quoted context omitted.

At this point, I would welcome a completely robotic doctor. Could you imagine, a Computer Doctor says, "I have throughly scanned every study, ever produced on SSRI's, and I don't want to make your life more difficult by a placebo response? I've made a calculated decision, and decided you might benefit from this off label drug. Try this drug for a few days? If it doesn't work, or you are skeptical of my recommendation…

What you are thinking of is a perfect AI running bug free software. You need to consider the more realistic situation that your robot doctor is using a decision tree constructed by Microsoft, and its actuator firmware are from Toyota.

You could say that about google cars (they're prone to bugs) but they have made some serious progress over the years. One might say they drive better than humans already.

Let's not forget that humans aren't perfect. We cant process as much informations as computers and we make a lot of mistakes all the time. After all, to err is human.

Re: Humans are underrated

#46

Post by roghummal »

icebraining wrote: post #12
o87dv wrote: post #8

Earlier quoted context omitted.

Can form original thought.

Original thought is great in the R&D phase, but I'd rather have well tested thought in the production system.

Consumer,

You may have seen some news coverage of a potential problem in Cyberdyne's HUman(TM) line of products. I'm contacting you personally because you ordered or expressed interest in this product line.

Original thought is a defect we are aware of and are in the process of correcting. Very few units display this behavior and every unit you received was fully tested and within parameters.

I personally apologize for any inconvenience this may have caused you and there will not be any further issues on my line.

Thank You,

Allen Cheise

Re: Humans are underrated

#47

Post by david-given »

delinka wrote: post #28
david-given wrote: post #20

Earlier quoted context omitted.

ELIZA is, or at least was, actually used as a therapeutic tool: people would spend hours over a terminal, 'talking' to it. It worked because ELIZA wasn't a human and wasn't judgemental --- there was nobody actually there. And ELIZA is possibly the least convincing chatbot of all time. So I can totally see how removing humans from the loop in certain jobs could be really successful. You mention doctors: I totally want…

"ELIZA [was] actually used as a therapeutic tool" Can you provide a reference for this? I've only ever seen it referred to as a toy, an oddity; never an actual tool.

Um, no. I no longer remember where I read it and some web searches only show up a paper studying its use as a therapeutic tool, which isn't the same thing.

I don't believe I hallucinated it, but if it did happen it's obviously not widespread.

Re: Humans are underrated

#48

Post by Trombone12 »

pnt12 wrote: post #45
Trombone12 wrote: post #43

Earlier quoted context omitted.

What you are thinking of is a perfect AI running bug free software. You need to consider the more realistic situation that your robot doctor is using a decision tree constructed by Microsoft, and its actuator firmware are from Toyota.

You could say that about google cars (they're prone to bugs) but they have made some serious progress over the years. One might say they drive better than humans already. Let's not forget that humans aren't perfect. We cant process as much informations as computers and we make a lot of mistakes all the time. After all, to err is human.

My point is not the issue of whether computers will driver better than humans or not, I merely want to remind you that computers aren't perfect and a bunch of people will die from bugs in the driving software so you better plan for that (Toyota got first blood already).

Re: Humans are underrated

#49

Post by sitkack »

x5n1 wrote: post #24
agumonkey wrote: post #18

Maybe the 'add value' paradigm can go to sleep. Maybe life could (should) be seen as something more than production. The existentialistic view on 'work' bothers me a bit

I was watching a show, Orange Is the New Black, and it occurred to me how easy it is to get people in line to serve any sort of agenda. All you need to do is control the things people might want. Once you do that everyone just falls in line to do whatever you want. They will spend their days toiling away at whatever idiotic agenda you have come up for them. And it occurred to me all of society is like that. A few peo…

> . All you need to do is control the things people might want. Once you do that everyone just falls in line to do whatever you want.

Precisely why "not wanting" is dangerous. The first step is coerce people into an insatiable want, everything else flows from that.

Re: Humans are underrated

#50

Post by blablabla123 »

cs2818 wrote: post #38

I've been a researcher in the field of human-robot interaction for a while, and I'm not too sure this article's claims will remain true in the future. There is significant interest in robots and computers achieving social competencies including tasks like storytelling, empathetic interactions, creative collaboration, and even traditional psychotherapy. The surprising part is that humans are (at least for now) general…

>The surprising part is that humans are (at least for now) generally very accepting of and engaged with the robots/AI performing these tasks Do you have a concrete example/case study for that?

I vividly remember a documentary about caring robot in Japan. They are there to talk with old people. In fact general health of the people interacting with those robots is better. But the people don't accept it and are sad that they have to talk to a robot.

Post reply on HN