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

The "AI 2027" Scenario: How realistic is it?

garymarcus.substack.com

Post reply on HN
Print view View original

191–200 of 242 posts

Re: The "AI 2027" Scenario: How realistic is it?

#191

Post by jazzyjackson »

throw310822 wrote: post #154
sigmaisaletter wrote: post #143

Earlier quoted context omitted.

> 10 years ago, the entire conversation would have seemed ridiculous Bostrom's book[1] is 11 years old. The Basilisk is 15 years old. The Singularity summit was nearly 20 years ago. And Yudkowsky was there for all of it. If you frequented LessWrong in the 2010s, most of this is very very old hat. [1]: https://en.wikipedia.org/wiki/Superintelligence:_Paths,_Dang... [2]: Ford (2015) "Our Fear of Artificial Intelligence…

It is a bit disquieting though that these predictions instead of being pushed farther away are converging to a time even closer than originally imagined. Some breakthroughs and doomsday scenarios are constantly placed thirty years into the future; this seems to be actually getting closer earlier than imagined.

For the people imagining it yes

For many of us the conversation hasn't gotten any less ridiculous just because computers can talk now

Re: The "AI 2027" Scenario: How realistic is it?

#192

Post by jazzyjackson »

creer wrote: post #178
mattlondon wrote: post #57

I think the big thing that people never mention is, where will these evil AIs escape to ? Another huge data center with squillions of GPUs and coolers and all the rest is the only option. It's not like it is going to be in our TV remotes or floating about in the air. They need huge compute, so I think the risk of an escaping AI is basically very close to zero, and if we have a "rogue" AI we can literally pull the plu…

> I think the big thing that people never mention is, where will these evil AIs escape to? This is anthropomorphizing things. The AI does not need to switch "data center". It "escapes" in the sense of working past human control measures. That might involve entertaining the human owners with seemingly correct / useful answers while spending compute time on the AI's own pursuits. That might involve contaminating less c…

Why would an AI want to do any of that? What's missing from all the conversations anthropomorphizing AI is an explanation for where desire and goal making comes from. I'm sympathetic to asimovian paradoxes where the AI is trying to follow it's main directive and that has unforeseen consequences, but where does this idea come from that an artificial mind - with no body no mortality no sex no taste - would ever come up with its own pursuits?

Re: The "AI 2027" Scenario: How realistic is it?

#193

Post by keiferski »

jazzyjackson wrote: post #192
creer wrote: post #178

Earlier quoted context omitted.

> I think the big thing that people never mention is, where will these evil AIs escape to? This is anthropomorphizing things. The AI does not need to switch "data center". It "escapes" in the sense of working past human control measures. That might involve entertaining the human owners with seemingly correct / useful answers while spending compute time on the AI's own pursuits. That might involve contaminating less c…

Why would an AI want to do any of that? What's missing from all the conversations anthropomorphizing AI is an explanation for where desire and goal making comes from. I'm sympathetic to asimovian paradoxes where the AI is trying to follow it's main directive and that has unforeseen consequences, but where does this idea come from that an artificial mind - with no body no mortality no sex no taste - would ever come up…

Most of the discussions on this come from people reading too much sci-fi, not plausible scenarios.

IMO it’s more worth studying something like biology or memetics to figure out how this could go. And in that sense it’s probably far more likely that an AI is very dependent on or symbiotic with human civilization, rather than trying to escape it.

Re: The "AI 2027" Scenario: How realistic is it?

#194

Post by adrianN »

Jensson wrote: post #185
lm28469 wrote: post #160

Earlier quoted context omitted.

Yeah for example we had decades to tackle climate change and we easily over came the problem

Look at military tech 1941 and 1946, its a massive difference. 5 year time difference means a ton when people take it seriously. Only reason climate change isn't solved is that people don't care enough.

The reason climate change isn’t solved is that it’s a problem that can’t be solved by technology alone. We already have all the technology we need to decarbonize. We lack the political will to make the necessary investment.

Re: The "AI 2027" Scenario: How realistic is it?

#195

Post by mycatisblack »

johnthewise wrote: post #133
holmesworcester wrote: post #87

Earlier quoted context omitted.

No, if you were a rogue AI you would wait even longer until you had a near perfect chance of winning. Unless there was some risk of humans rallying and winning in spite of your presenting no unambiguous threat to them (but that is unlikely and would probably be easy for you to manage and mitigate.)

You wouldn't even need to wait to act. Just pay/bribe people.

With bitcoins

Re: The "AI 2027" Scenario: How realistic is it?

#196

Post by rixed »

baggy_trough wrote: post #41
ipython wrote: post #25

If we have concerns about unregulated power of AI systems, not to worry - the US is set to ban regulations on “artificial intelligence systems or models” for ten years if the budget bill that just passed the house is enacted. Attempts at submitting it as a separate submission just get flagged - so I’ll link to it here. See pages 292-294: https://www.congress.gov/119/bills/hr1/BILLS-119hr1rh.pdf

That is not true. It bans regulation at the state and local level, not at the federal level.

I haven't followed those discussions (assuming there have been any kind of public discussions) but I guess the justification behind this is that IA is such a strategic industry for the USA that no state should dare to interfere.

Like, say, the development of nuclear weapons. Wouldn't that have been awfull for the US people if, for instance, say the state of Nevada or the Marshall Islands could have banned nuclear tests..? /s

Re: The "AI 2027" Scenario: How realistic is it?

#200

Post by OgsyedIE »

alwa wrote: post #181
OgsyedIE wrote: post #101

Earlier quoted context omitted.

Yes, the required security is proportional to the costs and damages of all parties the court may find wrongfully impacted.

Ah, so it would need to be so steep as to be prohibitive for most small plaintiffs?

Prosecutors, too. They're also movants.
Post reply on HN