News sucks

[Loren Pechtel]

A joke from the 70's about new technology:
A (insert your favorite ethic minority here) goes to a car dealership and buys a "party van". It is equipped with a plush interior and all the latest options. It's seen a few hours later running off the road and tumbling into a ravine. A rescue squad arrives and extracts him from the wreck. They ask him what happened and he says, "I don't know. I just set the cruise control and went to the back to take a nap."

And we have something akin to that that has played out in the real world. Tesla's self driving. The guy knew the Tesla couldn't really drive itself but understood it to be good enough to avoid utterly boneheaded moves. He bent down to pick up a dropped phone in a minimal traffic situation--and his car pulled an utterly boneheaded move. And Tesla had to pay out big bucks because of their repeated attempts to hide the evidence.

Today, it is possible for a fully informed person to set their van on cruise control, and go lie down for a nap, with no expectation of running off the road. If it does, we will certainly see it in the news. If the driver/passenger is not part of the engineering team which designed the self driving system, they will have to rely on someone else's judgment about safety.

Huh? Nobody has a system that can do that. Waymo is the closest but it's still geographically restricted.

Going back to the 70's once again, my father was engineer and he worked on projects to convert engineering formulas to computer programing language. Up to this time, engineers spent countless hours with a pencil and a slide rule, slowly calculating stresses, loads, and other important things. This would be repeated over and over again, until every part of the construction was correct. Then another engineer had to check the work. The computer programs were entered on punch cards and if a punch hole was in a place the computer did not recognize, it stopped work. My father's assessment was, "A computer is an idiot, but a very fast idiot."

Today, AI is in the "very fast idiot" stage. It can sift very large data sets very quickly. Even though it is very fast, there's no reason to believe it's any better than a human doing the same job, and should be accorded the same level of trust.

I'm thinking that a fair bit of the AI bit is yet another way to pretend to allow non-programmers to program. It never works because programming is more about translating loose reality into bulletproof instructions rather than the exact form those instructions take.

 

[Loren Pechtel]

We are already at the point where it is i8possible to detect fake AI created news video.

It is not just the corporate world. A group of people created AI music and fake music video of a band. They put it on the net and a following developed not knowing it was fake.

Art and music can now be reduced to AI. Combine AI with a machine and you have AI painting and sculpture.

It is being driven by good old capitalist unrestrained greed. Not by good for humanity, that claim is a marketing image.

A fake band is very much the sort of case where I think there's a good use for AI. It doesn't need to be truthful, the results are judged on their own merits independent of reality.

 

[Elixir]

Why wouldn't greater wealth generation not benefit all humans?

For the same reasons that IT NEVER HAS?
… just a wild guess.

 

[Loren Pechtel]

Why wouldn't greater wealth generation not benefit all humans?

For the same reasons that IT NEVER HAS?
… just a wild guess.

It always has, just not as fast as the left would like.

 

[southernhybrid]

The following is what AI told me about a big problem with AI

Search Assist

AI is impacting the environment primarily through the energy demands of data centers that power AI technologies, which significantly increase electricity consumption and greenhouse gas emissions. Additionally, the production and cooling of AI hardware require substantial water resources, contributing to environmental strain and potential ecological disruption. PBS nea.org

Environmental Impact of AI​

Energy Consumption​

AI technologies, particularly those involving deep learning models, require significant amounts of electricity. Data centers that support AI operations are major consumers of energy, with projections suggesting they could account for up to 4% of global electricity usage by 2026. This demand often leads to increased reliance on fossil fuels, contributing to greenhouse gas emissions and climate change.

Carbon Footprint​

The carbon footprint associated with training AI models is substantial. For instance, training large models can emit hundreds of metric tons of CO2, equivalent to the emissions from numerous cars over their lifetimes. As AI models grow in complexity, their energy requirements and associated emissions are expected to rise sharply.

Water Usage​

AI data centers also consume large quantities of water for cooling purposes. This demand can strain local water supplies, especially in areas already facing water scarcity. The cooling processes necessary for maintaining optimal operating conditions for servers can disrupt local ecosystems.

Electronic Waste​

The production and disposal of hardware used in AI systems contribute to electronic waste (e-waste). This waste often contains hazardous materials that can contaminate soil and water if not properly managed. As AI technology evolves, the lifecycle of its hardware becomes increasingly important to address.

Potential Benefits​

Despite these challenges, AI also holds potential for environmental benefits. It can optimize energy use, enhance efficiency in various sectors, and assist in environmental monitoring. However, the overall impact of AI on the environment remains a complex issue that requires careful consideration and management.
Wikipedia

 

[DrZoidberg]

The following is what AI told me about a big problem with AI

Search Assist

AI is impacting the environment primarily through the energy demands of data centers that power AI technologies, which significantly increase electricity consumption and greenhouse gas emissions. Additionally, the production and cooling of AI hardware require substantial water resources, contributing to environmental strain and potential ecological disruption. PBS nea.org

Environmental Impact of AI​

Energy Consumption​

AI technologies, particularly those involving deep learning models, require significant amounts of electricity. Data centers that support AI operations are major consumers of energy, with projections suggesting they could account for up to 4% of global electricity usage by 2026. This demand often leads to increased reliance on fossil fuels, contributing to greenhouse gas emissions and climate change.

Carbon Footprint​

The carbon footprint associated with training AI models is substantial. For instance, training large models can emit hundreds of metric tons of CO2, equivalent to the emissions from numerous cars over their lifetimes. As AI models grow in complexity, their energy requirements and associated emissions are expected to rise sharply.

Water Usage​

AI data centers also consume large quantities of water for cooling purposes. This demand can strain local water supplies, especially in areas already facing water scarcity. The cooling processes necessary for maintaining optimal operating conditions for servers can disrupt local ecosystems.

Electronic Waste​

The production and disposal of hardware used in AI systems contribute to electronic waste (e-waste). This waste often contains hazardous materials that can contaminate soil and water if not properly managed. As AI technology evolves, the lifecycle of its hardware becomes increasingly important to address.

Potential Benefits​

Despite these challenges, AI also holds potential for environmental benefits. It can optimize energy use, enhance efficiency in various sectors, and assist in environmental monitoring. However, the overall impact of AI on the environment remains a complex issue that requires careful consideration and management.
Wikipedia

Meh. Building the models consume a lot of energy. Running the models do not. While it's true now that AI consumes tremendous amounts of energy, over time that number will go down, while we can still enjoy AI. What this statement is based on is simply a projection of energy consumtion continuing as it is now, at a time when we go from no AI to lots of AI.

These worries are absurdly narrow minded imho. They also need to be weighed against the added efficiency that a professional world using AI brings. AI allows us to innovate quicker, ie more green technology faster. I'm sure that benefit will quickly be way greater than worries about energy consumption.

And moreover, you are highlighting just the kind of things that journalists reporting on AI do, using zero brains and doing zero actual research. It's just bullshit fear mongering for no reason.

News reporting on AI is almost entirely bullshit. That's why the same nonsense finds its way into Wikipedia.

 

[southernhybrid]

The following is what AI told me about a big problem with AI

Search Assist

AI is impacting the environment primarily through the energy demands of data centers that power AI technologies, which significantly increase electricity consumption and greenhouse gas emissions. Additionally, the production and cooling of AI hardware require substantial water resources, contributing to environmental strain and potential ecological disruption. PBS nea.org

Environmental Impact of AI​

Energy Consumption​

AI technologies, particularly those involving deep learning models, require significant amounts of electricity. Data centers that support AI operations are major consumers of energy, with projections suggesting they could account for up to 4% of global electricity usage by 2026. This demand often leads to increased reliance on fossil fuels, contributing to greenhouse gas emissions and climate change.

Carbon Footprint​

The carbon footprint associated with training AI models is substantial. For instance, training large models can emit hundreds of metric tons of CO2, equivalent to the emissions from numerous cars over their lifetimes. As AI models grow in complexity, their energy requirements and associated emissions are expected to rise sharply.

Water Usage​

AI data centers also consume large quantities of water for cooling purposes. This demand can strain local water supplies, especially in areas already facing water scarcity. The cooling processes necessary for maintaining optimal operating conditions for servers can disrupt local ecosystems.

Electronic Waste​

The production and disposal of hardware used in AI systems contribute to electronic waste (e-waste). This waste often contains hazardous materials that can contaminate soil and water if not properly managed. As AI technology evolves, the lifecycle of its hardware becomes increasingly important to address.

Potential Benefits​

Despite these challenges, AI also holds potential for environmental benefits. It can optimize energy use, enhance efficiency in various sectors, and assist in environmental monitoring. However, the overall impact of AI on the environment remains a complex issue that requires careful consideration and management.
Wikipedia

Meh. Building the models consume a lot of energy. Running the models do not. While it's true now that AI consumes tremendous amounts of energy, over time that number will go down, while we can still enjoy AI. What this statement is based on is simply a projection of energy consumtion continuing as it is now, at a time when we go from no AI to lots of AI.

These worries are absurdly narrow minded imho. They also need to be weighed against the added efficiency that a professional world using AI brings. AI allows us to innovate quicker, ie more green technology faster. I'm sure that benefit will quickly be way greater than worries about energy consumption.

And moreover, you are highlighting just the kind of things that journalists reporting on AI do, using zero brains and doing zero actual research. It's just bullshit fear mongering for no reason.

News reporting on AI is almost entirely bullshit. That's why the same nonsense finds its way into Wikipedia.

Maybe, but right now, the building of data centers is causing water shortages and power shortages in parts of the world, including in parts of Spain, unless you think that's all fake news. To me, humans shouldn't be suffering because of AI and I'm not convinced that we will be able to overcome all of the problems related to building all these data centers that are primarily being used for AI. Humans are really good at fucking up the planet, destroying other species as well as their own environment. There is overwhelming evidence for that, and I'm pretty sure that AI would agree with me on that. Plus, AI, as others have said, steals information from other sources, regardless of it's to make art, music or report the news. That was part of my point. Even AI knows that it has the potential to do a lot of damage to humanity. Time will tell how much.

 

[bigfield]

But what we've learned is that writing a document from scratch is always more work than having a document generated and then proof reading it, and leads to a better end result. Writing it from scratch will introduce more errors. Andf this increases productivity across the board.

Basically, humans are not good at doing boring things. As soon as we are bored quality goes down.

I wonder how much of this boring document writing is just bullshit work, and whether there's any value in making it more efficient.

 

[DrZoidberg]

But what we've learned is that writing a document from scratch is always more work than having a document generated and then proof reading it, and leads to a better end result. Writing it from scratch will introduce more errors. Andf this increases productivity across the board.

Basically, humans are not good at doing boring things. As soon as we are bored quality goes down.

I wonder how much of this boring document writing is just bullshit work, and whether there's any value in making it more efficient.

There's a lot about productivity that's basically voodoo.

Here's a fun annecdote. I have a friend who is a Zen buddhist monk. They had a new abott at his monastary. He changed the morning meditation. He removed the chanting. He had everyone sit in silence. Over time discipline fell apart all over the monastary and things got chaotic. Everyone knew why, becuase it was the only thing they changed. So they put the chanting back and then things went back to normal.

Change often have unintended consequences. We'll just have to wait and see. But I think the new, AI assited, way of writing documents will only lead to positive things.

 

[bigfield]

But what we've learned is that writing a document from scratch is always more work than having a document generated and then proof reading it, and leads to a better end result. Writing it from scratch will introduce more errors. Andf this increases productivity across the board.

Basically, humans are not good at doing boring things. As soon as we are bored quality goes down.

I wonder how much of this boring document writing is just bullshit work, and whether there's any value in making it more efficient.

There's a lot about productivity that's basically voodoo.

Here's a fun annecdote. I have a friend who is a Zen buddhist monk. They had a new abott at his monastary. He changed the morning meditation. He removed the chanting. He had everyone sit in silence. Over time discipline fell apart all over the monastary and things got chaotic. Everyone knew why, becuase it was the only thing they changed. So they put the chanting back and then things went back to normal.

Change often have unintended consequences. We'll just have to wait and see. But I think the new, AI assited, way of writing documents will only lead to positive things.

Perhaps not your intention, but it's funny that your productivity anecdote is a story about a place that produces nothing.

 

[pood]

More stupid AI bullshit