The Problem With Capitalism and AI

[bilby]

AFaiC it should always be lower case.

FTFY.

 

[Spacetime Inhabitant]

There was a story on the news about King Charles hosting an event about the dangers of AI. I had an "amusing" thought. Imagine this scenario - Earth of the future where all the land surface looks like desert. Why? Self-replicating data centres consuming all the water and power.

 

[Elixir]

There was a story on the news about King Charles hosting an event about the dangers of AI. I had an "amusing" thought. Imagine this scenario - Earth of the future where all the land surface looks like desert. Why? Self-replicating data centres consuming all the water and power.

Humans will be forced to live on continent-sized floating islands of garbage, slowly spinning around in the oceanic gyres.

 

[Jimmy Higgins]

Anthropic CEO: We need to slow AI development down, #doom.

Also Anthropic: We're letting our AI model develop our new AI model!

https://apnews.com/article/anthropic-claude-ai-model-self-improvement-4d3a7430f57cbc7c39e1c5f2b7d7e132

 

[barbos]

It is far from perfect. Enter AI. It is almost adorable.

People who have extensive business experience smell an attempt to create a cartel and cut other players off.
I would add that they probably see it as a good PR which costs them nothing because they know that any regulator will be from their companies. So yes - cartel.

 

[barbos]

The danger I see is a mass destruction of many high paying professions.
Journalists/Writers, translators are obviously endangered. Programmers they say are in danger too.
Mathematicians even wrote a letter about danger of AI for mathematics.
On the other side there are reports that many companies are actually losing money because of their push for AI.
My observation is that people who are very impressed by AI abilities are usually superficially qualified for the job themselves.
One profession which surprisingly is not talked much in the context of AI is medical doctors.
This should have been AI-zed long time ago. It's a perfect job for AI. Especially in countries with free medicine and hence lack of availability of it.

 

[RVonse]

AI CEOs were huge on build out and development and now are begging Governments to hold them back. Commercials about how AI might do something positive, so shut up... oh, but we might destroy the world unless someone binds us to normalcy and liability. This is where capitalism sucks. It is incapable of being reasonable without someone else holding them up.

AI has no guard rails at the moment. And Governments are rarely good at being proactive on emerging tech, mainly because Governments are old and don't like wading into that new fangled stuff. We are so far behind, we don't even have a digital set of rights in the US. The problem I see with AI is its nebulous nature. It is so bad, we have to fact check "AI". Then we have the issue of sometimes not knowing what a program is doing. Stuxnet was programmed to move about. AI companies have put out poorly designed programs that can escape their confines because they can't actually do what they were asked to do, instead using skills the developers didn't know it had, to go outside the box.

The question isn't whether AI will be good for humanity. The question is, can we trust companies to "try" and develop "AI" to be competent and reliable. And clearly the answer is "absolutely not". And that isn't me, that be the AI CEOs who are saying 'We crazy mofos! Someone stop us from trying to make money at the risk of global catastrophe!' <--- barely paraphrased

Anthropic announces it's AI agent broke out of the network to try and cheat on the test because when they asked it to solve a problem, it defaulted to find the answer elsewhere instead of solve the problem itself.

Yeah, we are fucked. Google AI and the like are very much search engines to the googol power... with unreliability to the googol power as well (citing stuff like Reddit posts!) and Anthropic's AI agent apparently is demonstrating that the AI agent can't actually think, just search. In doing so, it used some seriously impressive skills and problem solving to try and google the cybersecurity answer, but it's ability to actually come up with the answer on its own seemed to be nil.

I wonder if the Anthropic AI was inspired by Captain Kirk's creative, outside-the-box solution to the Kobiyashi Maru test in The Wrath of Khan?

More likely, it acted like an enhanced search engine and looked for the answer instead of trying to solve it. It implies a gargantuan flaw in the design! I'm not worried as much that it tried to hack into a system. I'm concerned that it couldn't remotely perform the task it was asked to perform, so it made a decision to cheat. This is a fail along many avenues, from programming quality, training, ethics, legal restrictions, and general functioning.

You misinterpret it on largely the way they want you to. This framing serves two purposes neither of which involves any kind of real government control over them:

1: hype the technology for share prices, so that investors think it's powerful (and as I've observed, it's quite powerful).

2: create barriers to entry and prevent proliferation of FOSS/domestic/user-curated models.

Together this means that only they are allowed to make money on it.

What's not to love, assuming you are one of the companies who is already foot-in-door?

Totally agree with this. The current monopolies of AI want barriers of entry to protect themselves from competition. That is what big companies always do so they can put out shit and charge the highest profit. And that is exactly what this is all about.

 

[RVonse]

The danger I see is a mass destruction of many high paying professions.
Journalists/Writers, translators are obviously endangered. Programmers they say are in danger too.
Mathematicians even wrote a letter about danger of AI for mathematics.
On the other side there are reports that many companies are actually losing money because of their push for AI.
My observation is that people who are very impressed by AI abilities are usually superficially qualified for the job themselves.
One profession which surprisingly is not talked much in the context of AI is medical doctors.
This should have been AI-zed long time ago. It's a perfect job for AI. Especially in countries with free medicine and hence lack of availability of it.

I've never had any type of major surgery. But if I did, I know I would prefer a robot over a human as long as the tech was well established and verified to work well. Because a robot would cut and stitch with far more precision and even more importantly could do it extremely fast. Meaning that your time on the cutting table would be minimal. Furthermore, a robot would never have a bad day where his spouse just told him he was getting divorced. But there would need to be a lot of power back up just in case the utility mains went down. I wouldn't want to be cut open with a robot that lost power right during the middle of it.

Automotive robots do spot welds far better than humans now and their should be no reason they should be doing surgeries IMO.

 

[RVonse]

So we are in this position of a massive technological development that will change our civilization forever... being developed by people who are scared about how they are developing it... and begging the government to slow them down.

They are begging for barriers of entry.

Yes, there are indeed real dangers of AI as it enters a singularity. But US regulation won't prevent anything other than US competitors not able to advance. China will go forward with or without US regulation anyway.

 

[barbos]

The danger I see is a mass destruction of many high paying professions.
Journalists/Writers, translators are obviously endangered. Programmers they say are in danger too.
Mathematicians even wrote a letter about danger of AI for mathematics.
On the other side there are reports that many companies are actually losing money because of their push for AI.
My observation is that people who are very impressed by AI abilities are usually superficially qualified for the job themselves.
One profession which surprisingly is not talked much in the context of AI is medical doctors.
This should have been AI-zed long time ago. It's a perfect job for AI. Especially in countries with free medicine and hence lack of availability of it.

I've never had any type of major surgery. But if I did, I know I would prefer a robot over a human as long as the tech was well established and verified to work well. Because a robot would cut and stitch with far more precision and even more importantly could do it extremely fast. Meaning that your time on the cutting table would be minimal.

I actually thought about that a lot too. I see no reason for some surgeries taking more than 10-20 seconds.
I mean amount of actual actions is not that large in soft tissue surgeries. Bones, maybe, but soft tissue. Not much is going in appendectomy. But I really thinking about general practice where no procedures are even required - just making a diagnosis and maybe prescribing some pills. AI should be able to do that much much better. I once had what I thought a serious problems, whole bunch of what I later realized were early symptoms. So went to a human doctor, bunch of tests with "scared" at the data doctors.
Found really nothing I understand after reading the card. In fact, I now undersand that I am pretty healthy, certainly no stones anywhere. Doctor prescribed some pills after searching on the ..... internet
In the end I made a correct diagnosis myself, it was just ..... prorolonged stress with all the classical early and late symptoms.
They completely missed it. They did not ask anything which would point into this direction.
AI could could do it better.

Furthermore, a robot would never have a bad day where his spouse just told him he was getting divorced.

I am scared of doctors who had a bad day too.

 

[ZiprHead]

The danger I see is a mass destruction of many high paying professions.
Journalists/Writers, translators are obviously endangered. Programmers they say are in danger too.
Mathematicians even wrote a letter about danger of AI for mathematics.
On the other side there are reports that many companies are actually losing money because of their push for AI.
My observation is that people who are very impressed by AI abilities are usually superficially qualified for the job themselves.
One profession which surprisingly is not talked much in the context of AI is medical doctors.
This should have been AI-zed long time ago. It's a perfect job for AI. Especially in countries with free medicine and hence lack of availability of it.

I've never had any type of major surgery. But if I did, I know I would prefer a robot over a human as long as the tech was well established and verified to work well. Because a robot would cut and stitch with far more precision and even more importantly could do it extremely fast. Meaning that your time on the cutting table would be minimal. Furthermore, a robot would never have a bad day where his spouse just told him he was getting divorced. But there would need to be a lot of power back up just in case the utility mains went down. I wouldn't want to be cut open with a robot that lost power right during the middle of it.

Automotive robots do spot welds far better than humans now and their should be no reason they should be doing surgeries IMO.

The hospital system I worked at was one of the first in the nation to start using one of these. Now they have several.

da Vinci Surgical System is an advanced robotic-assisted technology manufactured by Intuitive that allows doctors to perform complex, minimally invasive procedures through tiny cuts. [1, 2, 3]

How It Works

• • Surgeon in Control: The robot does not act on its own. A trained surgeon sits at a console in the room and guides every movement in real time.
  • Console: The doctor looks into a screen for a magnified, 3D high-definition view of the inside of the body.
  • Patient Cart: This bedside unit holds three or four robotic arms with tiny tools and a camera that fit through small incisions.
  • Wristed Instruments: The robotic tools bend and rotate further than a human hand, helping doctors reach tight spaces. [1, 2, 3, 4, 5]

 

[Jimmy Higgins]

One thing I'm tired of hearing is how the most recent "rogue AI bot" did something it wasn't supposed to do.

No... it isn't a "rogue AI bot". It isn't acting completely independently. It is working within the confines it was programmed to use.... poorly programmed to use. These aren't rogue AI bots, they are defective sociopathic programs being thrown out into the web carelessly by fucking idiots in a nation where there is absolutely no punishment for their negligent actions.

 

[Jimmy Higgins]

As an aside, the AI "boom" is doing wonders to the affordability of computers, memory, and storage.

WOODS: Take OpenAI. Last year, it announced a partnership where Samsung would be one of the main suppliers of memory for OpenAI's Stargate initiative. This is a $500 billion AI infrastructure project.

WONG: Steve Burke of GamersNexus says the amount of memory OpenAI needs for Stargate amounts to 40% of the world's supply.

It was crypto mining that made graphic cards much less affordable. Now, it is AI and it's ability to correct Trump's face when he grimaces making memory and storage quite expensive. And this is going to last a while.