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

Preparing for AI's economic impact: exploring policy responses

anthropic.com

Post reply on HN
Print view View original

51–60 of 85 posts

Re: Preparing for AI's economic impact: exploring policy responses

#51

Post by jononor »

mrshadowgoose wrote: post #13
blibble wrote: post #5

they seem to have omitted the scenarios where the newly unemployable electorate turn on them

I used to think that "AI operating in meatspace" was going to remain a tough problem for a long time, but seeing the dramatic developments in robotics over the last 2 years, it's pretty clear that's not going to be the case. As the masses fade into permanent unemployment, this will likely coincide with (and be partially caused by) a corresponding proliferation in intelligent humanoid robots. At a certain point, "turn…

What development have happened in robotics over the last 2 years, that changed your mind? I am familiar with general industry trends in electronics over the last two decades, as well as the 1 decade in machine learning.

Re: Preparing for AI's economic impact: exploring policy responses

#52

Post by intended »

If upskilling is your hope - you have no hope.

Without any modifications - MOOCs have single digit completion rates. This is high quality, free, publicly available educational material.

The vast majority of people do not simply have the time, money, or undivided attention - to get a new domain under their belt.

This is “help miners learn code” territory.

Re: Preparing for AI's economic impact: exploring policy responses

#54

Post by andy99 »

danpalmer wrote: post #37

If you write policy about AI you're doing it wrong. AI is an implementation, but policies must be written for outcomes. Discrimination by law enforcement, exclusion from loan approval, bad moderation on social networking, cheating on exams, creating fake news or media about people, swallowing up user data... all the negative social impact of AI can be achieved without it, and much of it is already illegal anyway. Leg…

Agreed, I think it’s more a lens (one of many) that helps show what’s possible with technology and what may require legal protections.

For example things like privacy and surveillance laws obviously need updating in the face of advances in networking, data collection at scale, etc. Same with copyright in the face of plentiful copying.

But good laws will as you say address what is now possible or dangerous, as opposed to any specific implementation or general purpose technology involved. The tech just sets the context for what protections are needed.

Re: Preparing for AI's economic impact: exploring policy responses

#55

Post by sateesh »

intended wrote: post #52

If upskilling is your hope - you have no hope. Without any modifications - MOOCs have single digit completion rates. This is high quality, free, publicly available educational material. The vast majority of people do not simply have the time, money, or undivided attention - to get a new domain under their belt. This is “help miners learn code” territory.

What other option you can propose. This article [1] says preferred suggestions by economists is: retraining, regulation, or social insurance and for most of the people surveyed "retraining" was the preferred approach.

Not sure MOOCs can be taken as an useful alibi to measure success of upskill. Most (employers) won't honor the MOCC certs, and people do MOOC while working. Taking a MOOC doesn't inherently ensure that the learner has mastered the course they took, hence there is less incentive in completing too.

1. https://www.foreignaffairs.com/united-states/coming-ai-backl...

Re: Preparing for AI's economic impact: exploring policy responses

#56

Post by intended »

sateesh wrote: post #55
intended wrote: post #52

If upskilling is your hope - you have no hope. Without any modifications - MOOCs have single digit completion rates. This is high quality, free, publicly available educational material. The vast majority of people do not simply have the time, money, or undivided attention - to get a new domain under their belt. This is “help miners learn code” territory.

What other option you can propose. This article [1] says preferred suggestions by economists is: retraining, regulation, or social insurance and for most of the people surveyed "retraining" was the preferred approach. Not sure MOOCs can be taken as an useful alibi to measure success of upskill. Most (employers) won't honor the MOCC certs, and people do MOOC while working. Taking a MOOC doesn't inherently ensure that…

Then There are NO options. We will have to live with that world and having no future.

I’m aware that this level of nihilism is difficult for many to stomach - but it’s only nihilism if you believe in fairy tales.

If we are talking about reality then we have to deal with the impossible challenges we are facing.

The fact is that retraining will not work - nor is this the first time it’s been named as a hope up in the past 25 years. (“teaching miners how to code” comes from the last time this was a big hope in America.)

If it helps you feel better - I said unmodified MOOCs.

With some changes, you can increase MOOC completion rates - but there isn’t enough lift even after that.

I used to be a champion for education initiatives to up skill workers - from a time before MOOCs.

The failure of MOOCs was the end of that hope because it showed there was a gape between the ideal and reality.

People simply can’t retrain like that.

Re: Preparing for AI's economic impact: exploring policy responses

#57

Post by danaris »

buu700 wrote: post #45
varispeed wrote: post #15

Earlier quoted context omitted.

> its interesting that the ai companies are already saying that. This is just cheap PR to launder legitimacy and urgency. To create false equivalence between AI agent and an employee. I think this is a sign of weakness, having seen AI rolled out in many companies where it already shows signs of being absolute disaster (like summaries changing meaning and losing important details - so tasks go in wrong direction and t…

Personally, if I were going to publish something like this as a leader of a major AI company today, I would actually try very hard to put together a good faith proposal that I genuinely believed to be in the best interests of the public. I can't speak to this particular proposal or the motivations behind it, but I think my approach is the smart play in the present circumstances. Why publish something brazenly self-se…

> Why publish something brazenly self-serving that will at best be forgotten two weeks later, or at worst be added to the list of reasons a bunch of people have to hate you, when you could instead earn some goodwill as a benevolent thought leader and maybe get some academics and politicians to come out of the woodwork backing your ideas?

For the same reason that the tech execs do all the other terrible things they do: because they want to own e v e r y t h i n g, and know that they can't do that by acting in good faith.

They want to be the new feudal overlords, and care much less about "goodwill" than they do about making it seem inevitable that they will be the gatekeepers of all thought and labor.

The more they can convince you, the people, and the policymakers that this "AI revolution" is real, and not just a bubble, the less likely everyone is to see through their exaggerations, misdirections, and outright lies to the fact that LLMs are not, and are never going to become, AGI. They are measurably not replacing any significant number of workers. They cannot do our jobs.

Re: Preparing for AI's economic impact: exploring policy responses

#58

Post by ndriscoll »

lm28469 wrote: post #44
danpalmer wrote: post #37

If you write policy about AI you're doing it wrong. AI is an implementation, but policies must be written for outcomes. Discrimination by law enforcement, exclusion from loan approval, bad moderation on social networking, cheating on exams, creating fake news or media about people, swallowing up user data... all the negative social impact of AI can be achieved without it, and much of it is already illegal anyway. Leg…

> all the negative social impact of AI can be achieved without it, With the big differences being massive automatisation, huge reduction of cost and no one to blame when things go wrong... It's like saying a nuke and a knife are the same because they both kill

Someone is to blame for approving the use (or unapproved use) of a tool/process that breaks the law, same as today. When I worked in a regulated industry, we kept records of all inputs and decisions made and an auditor would do random checks that the results matched our documented methodology (afaik that documentation was submitted to/approved by either an auditor or the regulator).

Re: Preparing for AI's economic impact: exploring policy responses

#59

Post by sateesh »

intended wrote: post #56
sateesh wrote: post #55

Earlier quoted context omitted.

What other option you can propose. This article [1] says preferred suggestions by economists is: retraining, regulation, or social insurance and for most of the people surveyed "retraining" was the preferred approach. Not sure MOOCs can be taken as an useful alibi to measure success of upskill. Most (employers) won't honor the MOCC certs, and people do MOOC while working. Taking a MOOC doesn't inherently ensure that…

Then There are NO options. We will have to live with that world and having no future. I’m aware that this level of nihilism is difficult for many to stomach - but it’s only nihilism if you believe in fairy tales. If we are talking about reality then we have to deal with the impossible challenges we are facing. The fact is that retraining will not work - nor is this the first time it’s been named as a hope up in the p…

Not that I completely disagree with you, yes indeed with excellent training not sure how many can be retrained to be good enough for new kind of job roles, skills. Also it isn't certain that there will be far more new jobs than what would get shed, in which case there won't be enough demand to absorb even the retrained (and skilled) labor.

Re: Preparing for AI's economic impact: exploring policy responses

#60

Post by cudgy »

danpalmer wrote: post #39
Cheer2171 wrote: post #38

Earlier quoted context omitted.

Oh, so we can't address any specific problems with any technology, because we should actually be fixing all of society at the root of all those problems. So while you wait for our broken political system to solve those root causes, enjoy feeling smug about not having implemented any imperfect, temporary bandaids to stop some bleeding. Are you working on fixing those root problems? Or after dismissing short term polic…

If there are problems that are specific to AI then sure we should legislate about it. For example defining what "fair use" is for AI training, that's a clearly new area. But most of the pushback I've seen to AI in policy is so over-fit to current AI that it would be trivial to work around it. You can argue that we'd be letting perfect be the enemy of good, but I think we'd be making policies that will be out of date…

“For example defining what "fair use" is for AI training, that's a clearly new area.”

I am not a patent attorney, but it seems like a clear violation of copyright. Based on your comment above regarding the breath and focus of laws and the fact that you feel the copyright law was not well specified for the AI situation. How could the current law have been written such that it would’ve handled the AI situation and avoided this mess that we’re in now?

My guess is none of it matters because now the AI is so important and so critical in the minds of many government leaders and business leaders that any violation of copyright will be excused, making the original law, meaningless in this situation, and undercutting this entire discussion.

Post reply on HN