Jaan Tallinn and Robin Hanson: Should We Pause A.I.?

The co-creator of Skype says yes. The George Mason University economist says no.

1x 1.1x 1.25x 1.5x 2x 3x

:15 :15

Download

Jaan Tallinn and Robin Hanson: Should We Pause A.I.?

This is the audio version of this week's The Reason Livestream, which takes place every Thursday at 1 p.m. Eastern.

This week's topic is artificial intelligence, or A.I., and my Reason colleague Zach Weissmueller interviews two leading thinkers on the subject. 

Jaan Tallinn of the Future of Life Institute organized an open letter calling for a pause on A.I. development that was signed by Elon Musk, Apple co-founder Steve Wozniak, and 27,000 other people. Tallinn is a tech investor, one of the software developers who created Skype and Kazaa, and co-founder of the Future of Life Institute.

On the other side of the issue is George Mason University economist Robin Hanson, who thinks the worries over A.I. are overblown. He says much of today's A.I. anxiety is a more generalized "future fear," which is likely to imperil technological progress that would greatly benefit humanity.

To watch the video version, go here.

Comments

[Don't look at me!]

Like almost everything else, it will fail to live up to the hype.

[Diane Reynolds (Paul.)]

Sort of like Segway, Quibi, Vice News and Don Lemon?

[angelinajolie]

Great article, Mike. I appreciate your work, I’m now creating over $35,000 dollars each month simply by doing a simple job online! I do know You currently making a lot of greenbacks online from $28,000 dollars, its simple online operating jobs.
.
.
Just open the link—————————————>>> http://Www.JobsRevenue.Com

[Melinda Rogers]

I am making a good salary from home $6580-$7065/week , which is amazing under a year ago I was jobless in a horrible economy. I thank God every day I was blessed with these instructions and now it’s my duty to pay it forward and share it with Everyone,
🙂 AND GOOD LUCK.:)

Here is I started.……......>> http://WWW.RICHEPAY.COM

[IslaShepherdson]

Im making over $13k a month working part time. I kept hearing other people tellme how much money they can make online so I decided to look into it. Well, itwas all true and has totally changed my life.last month my pay check was $13583just working on the laptop for a few hours.Simply go to this website and followinstructions to start earning....Have a good Day..
.
.
For more detail visit the given link...................➤➤ https://Www.Coins71.Com

[JuanitaHernandez]

Google pays $300 on a regular basis. My latest salary check was $8600 for working 10 hours a week on the internet. My younger sibling has been averaging $19k for the last few months, and he constantly works approximately 24 hours. I’m not sure how simple it was once I checked it out.

This is my main concern.......... http://Www.Smartjob1.com

[biroyi7084]

Great article, Mike. I appreciate your work, I’m now creating over $35,100 dollars each month simply by doing a simple job online! I do know You currently making a lot of greenbacks online from $28,100 dollars, its simple online operating jobs
. .
.

Just open the link ——————— https://earnwork8.blogspot.com

[JasmineCarson]

Lucy . you think George `s storry is impossible, on sunday I got a brand new Saab 99 Turbo after having made $8551 this past four weeks and just over ten-k last month . it's by-far the most comfortable job I have ever had . I started this five months/ago and almost straight away began to bring home over $95... per-hour..
🙂 AND GOOD LUCK.:)
HERE====)> https://www.apprichs.com

[Diane Reynolds (Paul.)]

AI is non-binary. "Pausing" AI would be transphobic.

Discuss.

[VULGAR MADMAN]

I thought AI would be entirely binary.

[Mike Laursen]

Many commenters here are obsessed with certain partisan hobby horse topics, so they will inject those hobby horses into every discussion even when it an absolute non sequitur.

Exhibit A

[Overt]

Kind of like that time Mike kept trying to redirect discussions of GMOs to Vaccines. Because Mike is a hypocritical troll.

https://reason.com/2022/01/07/mandatory-gmo-disclosure-doesnt-sway-shopping-habits-but-will-drive-up-costs/?comments=true#comment-9293589

[Diane Reynolds (Paul.)]

Jaan Tallinn of the Future of Life Institute

WHO?!!

[MWAocdoc]

The concept of “pausing” A.I. development is meaningless. Someone is going to develop A.I. whether the mainstream guys pause it or not. Like King Canute holding back the tide, all such demands are silly and for show only. Besides, the panic set has almost always been wrong with their doomsday scenarios and throughout history the Luddites have almost always been more about special protection of the status quo than about preventing disasters. Change is going to happen – find a way to live with it win or lose, or fail to cope and go hide in a safe space somewhere, but leave the rest of us alone!

[Diane Reynolds (Paul.)]

Never underestimate the government's ability to ban or regulate something through sheer force of will.

[Robin Austyn]

I get over $25k a month working part time. I kept hearing other people tell me how much money they can make online so I decided to look it. {rea-03} Well, it was all true and has totally changed my life. This is what I do, check it out by visiting:-
.
Following Link:→ → https://workhere3.blogspot.com/

[Longtobefree]

Anyone who thinks a technology like "A.I." can be paused would think a man can become a woman.

[Mike Laursen]

Exhibit B

https://reason.com/podcast/2023/05/12/jaan-tallinn-and-robin-hanson-should-we-pause-a-i/?comments=true#comment-10061321

[perlmonger]

Too late. The Cat is Out of the Bag. Repeat, Cat Out of Bag.

[Longtobefree]

Is it an organic, reusable bag?
And was the cat cage free?

[chemjeff radical individualist]

I know, right? How could any "pause" in AI development be realistically enforced by anyone?

[MWAocdoc]

"The Moon is in the Seventh House." Repeat - "The Moon is in the Seventh House." Yes. Repeat: No.

[Razupaltuf]

just let the development of AI happen, it will happen anyways

[JesseAz]

I would prefer a pause on the hyping of AI. It is only as good as its training data which we have seen is pretty bad.

[CE]

Or it's already happened.
And we're just living in a simulation it's running about how it got started, back in the day.

[Nardz]

Troubling news on the free speech front:

https://www.zerohedge.com/markets/top-nbcuniversal-ad-exec-and-world-economic-forum-taskforce-chair-become-twitter-ceo

[Nardz]

https://twitter.com/ImMeme0/status/1657082685527924737?t=08cNa9dV9dlHapskHDOCyA&s=19

Listen to Linda when asked about the importance of diversity, equity and inclusion in the workplace. She cites the CEO of BlackRock and says that companies have “responsibility to impact culture”.

[Video]

[BigT]

The value of a “call to purpose” depends on the purpose. Disney 1940-2000, good; 2000-now, bad.

[CE]

Who is "we"?
Because if it isn't everyone, AI isn't going to be paused.

[MWAocdoc]

Yeah, what if "A.I." doesn't WANT to be "paused?" The mechs are coming! The mechs are coming!!

[BigT]

You not pause AI, AI pause you!

[Overt]

The more substantive question is what is AI good for? Let's note that AI has been ascendant for close to a decade now. But what people are reacting to now largely deals with Large Language Models (LLMs).

Trigger warning: Adam Conover, who is probably one of the shrillest know-it-alls on tv. But the first 20 minutes of this video are pretty spot on.

https://www.youtube.com/watch?v=ro130m-f_yk

As I have noted before, LLMs actually do not deliver superior results. What they do is deliver results in a superior, conversational manner that causes us to give them the benefit of the doubt. When you are on a search results page in google, you can easily say "This totally isn't what I was looking for." But when the AI delivers the same results in prose, we are more likely to say "well that's close but I really want this..." In many ways, AI is like a John Edwards of auto complete, using cold reading techniques to convince you that there is more happening behind the scenes.

All that said, I have found good use from Chat GPT-4 in having it write letters or copy for marketing materials. In the past, I'd farm a presentation or paper out to a 1st year Individual Contributor to do some competitive analysis or a one pager. ChatGPT gives me only marginally worse results in 1/40th the time.

If you have kids going into highschool/college/just graduating, I would highly recommend getting a ChatGPT account and teaching them how to use it appropriately. There is a danger of using the AI to produce something that is WILDLY wrong, but plausible sounding enough to pass your bullshit filters. If a kid were to turn any of that in to a teacher or boss, it could drastically hurt their credibility. On the other hand, if they use the tool as a study aid and prompt for writing activities, it can drastically increase their productivity, and focus their scarce time on the critical aspects of the assignment rather than busy work.

[Overt]

For example, here is GPT 4 explaining the NAP as if you were a kindergartner.

"Sure, I'd be happy to explain that!

You know how when you're playing in the sandbox and someone else has a really cool toy you want to play with? The Non-Aggression Principle is like an important rule that says you can't just take their toy away because you want it. It means you can't hurt them or scare them to get their toy either. It's not nice to use force to get what you want.

But guess what? If you ask them nicely, they might let you play with it, or maybe you can trade one of your toys for a turn with their toy. That's the better way to do things. It's all about respecting other people, their stuff, and not using force or intimidation to get what you want.

So, in short, the Non-Aggression Principle is like a big rule that says "Don't hurt people and don't take their stuff"."

I further asked Chat GPT 4 to explain why the NAP can make a moral society- without being pragmatic. At this point, GPT made a pragmatic case- saying that if we follow the NAP, society will be more fair and kind. So you do need to be aware and re-prompt if necessary.

For example, I asked "Why is it better to have kindness, respect and fairness instead of order?"

"Now imagine a society where order is more important than kindness, respect, and fairness. People might be organized and things might run on schedule, but if individuals are not treated with respect or fairness, they may feel oppressed or undervalued. This can lead to unhappiness, conflict, and even rebellion.

In a society that values kindness, respect, and fairness, order is achieved voluntarily, as people cooperate and treat each other well out of mutual respect and understanding. This kind of order - an order based on shared values and mutual respect - is often more stable, peaceful, and just than order imposed without regard for individual rights and freedoms."

Again, it is falling back to practical benefits.

Overall, CGPT is merely good at vomiting words on a page. They are a good start- just like a template is a good start for any letter. But you still need to have mastery of your subject to push towards something better.

[RickAbrams]

How will AI distinguish fact from fiction. Will AI think Trump won in 2020? The chances that AI will decide that human beings are an invasive species and a threat to other, nicer species, is about 100 in 100. The only thing we have going for us is that a lot of us are nice to our dogs. I predict that AI will wipe out all human beings, except for those of us whose dog vouches for us.