Disney making $1 billion investment in OpenAI, will allow characters on Sora AI video generator

[flynnibus]

No one was trying to use the printing press to determine missile strikes, or asking what medications can safely mix, or financial planning advice, etc. Also, as best I can tell, no one formed an unhealthy fake relationship with a printing press because it constantly validated them every time they talked to it.

So you're against search engines too I guess?

In a professional setting an AI can be great when used as a tool and not a replacement, and most of the people I know who praise it are looking at it in that sense. Great, but for the general public, it is a mess with a ton of traps and downsides that are being largely ignored.

Not being ignored - being willing to address as bad responses they will continue to improve. AKA evolving vs trying to hold back the tide.

The kicker here is that these companies are paying a bunch of money for these systems and their end goal isn't to have a few optional services on the fringe.

The companies are doing it because they fear falling behind the curve.

 

[parksandtravel]

You made a customized, interactive location-based tour guide without AI?

You can ask AI to cite its sources or limit to specific sources. I try to fact-check whether I’m using AI or not. It may be “notoriously unreliable” for people who don’t understand how to use it well, but I’ve found it to be very reliable and accurate.

Just curious, how much would you pay for that if it wasn't "free"?

 

[parksandtravel]

Yes, you can also etch the message in stone using two pieces of rock. It just takes a lot more effort, and more time.

Using a tool, you can reduce your time needed, you can improve the detail, and possibly get ideas you wouldn't have considered.. all for very little input/cost and all you need to do is review and refine the results.

Calling AI a tool is a very nebulous description. Those other tools don't cause damage like AI does to the environment or strain infrastructure each time they're used to the same degree and often for very frivolous purposes.

 

[flynnibus]

Calling AI a tool is a very nebulous description. Those other tools don't cause damage like AI does to the environment or strain infrastructure each time they're used to the same degree and often for very frivolous purposes.

Do you think the search engine bots and indexes just fit within existing infrastructure from 1993? That all the stuff you take for granted today didn't require expanding capacities?

 

[Mr. Engagement]

Just curious, how much would you pay for that if it wasn't "free"?

I pay $25/mo. for the Business subscription tier for apps, greater memory/context, and so that my information isn't used to train their models. I primarily use custom GPTs for research, development, instructional design, and documentation.

For me, what I'm getting out of it seems well worth the price.

 

[parksandtravel]

Do you think the search engine bots and indexes just fit within existing infrastructure from 1993? That all the stuff you take for granted today didn't require expanding capacities?

It's not to the same degree at all. And you didn't address the other part of the damage it causes.

 

[MisterPenguin]

Most of AI usage is used to automate tasks and to make some tasks easier by just asking for it. This is what corporate usage is all about. And this will never go away. It's mostly manipulating data they already have and organizing it.

Then there's the AI of self-learning to do tasks, like driving or being a robot doing robot things. Since it's learning... it could also learn the wrong thing. This type of AI needs a lot of baby-sitting to avoid bad consequences. But it shines in low-stakes task like "remove the background of this photo and replace it with a blue sky and white fluffy clouds."

Where AI gets a bad rap is with the "shadiness" of AI demonstrated by search queries giving bad results and AI "hallucinating" when people use AI to do original research. The AI engine itself has issues with the hallucinations. And the bad search results is often from the vast information on the web that is tainted by humans promoting false information and conspiracy theories. The human bias creates biased data. AI then uses that biased data for biased results.

 

[parksandtravel]

I pay $25/mo. for the Business subscription tier for apps, greater memory/context, and so that my information isn't used to train their models. I primarily use custom GPTs for research, development, instructional design, and documentation.

For me, what I'm getting out of it seems well worth the price.

Got it, your situation is different than the majority of the public. If you said that previously then I must have missed it. Although even at $25 that is far below the breakeven point for them so I am curious what the people who advocate for it so strongly would actually be willing to pay if costs were not offset by investor capital.

 

[Mr. Engagement]

Got it, your situation is different than the majority of the public. If you said that previously then I must have missed it. Although even at $25 that is far below the breakeven point for them so I am curious what the people who advocate for it so strongly would actually be willing to pay if costs were not offset by investor capital.

Yes, I'm not really the typical user. And you're right, they're not making money from me. I'm sure prices will only rise as the technology improves and the use cases become even clearer.

 

[flynnibus]

It's not to the same degree at all. And you didn't address the other part of the damage it causes.

It is the same. It's the same data center, same compute, same resources. What is happening now is rapid expansion to meet demand - not something radically different. The silicon is different - what they consume, what they produce, is all the same. What has changed is the industry is cramming more and more density in.

"damage it causes" -- Dude, your Google search engine uses the same cooling, the same fire retardant, same backups, same network, etc.

The 'dot com boom'? Maybe you've heard of it? All the rapid explosion of infrastructure started back then - and was a lot more disruptive as network 'highways' were being built everywhere. Now, thanks to dense wave multiplexing we dig less, and put more compute into the same footprint, generate less heat, and cool it more efficiently.

Do you have any actual specifics, or are you gonna stick with the doom and gloom the and the AI stole my baby?

I didn't know youtube and search were only used for totally above the line purposes..

 

[Dranth]

So you're against search engines too I guess?

Search engines in general, no. I was a big fan of how they started. I am not as enamored with what they have become.

That aside, the comparison is a bit of a stretch in my mind. Search engines are essentially a card catalog for the web. Sure, AI could also be used that way, but that is barely scratching the surface of what it could potentially do.

Not being ignored - being willing to address as bad responses they will continue to improve. AKA evolving vs trying to hold back the tide.

I don't think it is being ignored by the companies that make them, or at least most of those companies. They obviously want to make it better if they want any chance to "win" the AI race. That doesn't help with the growing number of people who treat the current iterations as an authoritative voice on whatever they ask about.

The companies are doing it because they fear falling behind the curve.

I understand why they are doing it, I just think it is a mistake for many companies or completely unnecessary for others. At least for now.

Overall, I think most can agree that the upside for AI is HUGE, but so is the downside. Personally, I just don't trust many of the people in charge of both creating it and regulating it right now. That doesn't mean I am against it in concept, just that a wild west style, no holds barred, race to the finish line isn't the best choice for developing and deploying something like AI.

 

[flynnibus]

Search engines in general, no. I was a big fan of how they started. I am not as enamored with what they have become.

That aside, the comparison is a bit of a stretch in my mind. Search engines are essentially a card catalog for the web. Sure, AI could also be used that way, but that is barely scratching the surface of what it could potentially do.

It's not a stretch at all. The angst being shown here... people offloading thinking to tech, and the 'cost' of providing that tech are direct parallels. Both involve huge changes in access to "information" and neither just appeared out of the blue, nor are they free to the planet to sustain.

Are we upset people have become so dependent on search engines they don't even type out URLs anymore? Or that people don't actually read the source - they just goto the wikipedia or search versions of things? All of these are similar in that technology is replacing work -- and the difference is most people are OK with the current status quo.. even tho it's not perfect by any means. You can still get crap pages... people still publish bad information... People still need to VET what find online.

I don't think it is being ignored by the companies that make them, or at least most of those companies. They obviously want to make it better if they want any chance to "win" the AI race. That doesn't help with the growing number of people who treat the current iterations as an authoritative voice on whatever they ask about.

You mean like the people who watch some propaganda TV show? Or who rejurgitate some non-sense they found on a website? Or who broke something trying to follow a poorly put together youtube howto they never should have been attempting?

All of this re-enforces what was said before... The human is still responsible for their reputation and reviewing output. Failing to do that and leading to garbage is not an AI problem - It's a human problem. It doesn't matter if they are repeating pesudo-science preached by some con-man, or the latest fake-news twist on things. This is a human problem if they pass along bad stuff.

I understand why they are doing it, I just think it is a mistake for many companies or completely unnecessary for others. At least for now.

Overall, I think most can agree that the upside for AI is HUGE, but so is the downside. Personally, I just don't trust many of the people in charge of both creating it and regulating it right now. That doesn't mean I am against it in concept, just that a wild west style, no holds barred, race to the finish line isn't the best choice for developing and deploying something like AI.

Like when search results weren't unbiased? Or how people would game the internet with bad info intended to hurt others?

This is more about how quickly new technology hits the mass market now (including the idiots of the world) when it used to be much slower and more restrictive. The ability for new capabilities like Sora to be just a few clicks away for ANYONE is really what throws gas on the fire.

TLDR - The gen pop will always be stupid and careless - it's not an AI issue as much as it's a human issue.

 

[Dranth]

TLDR - The gen pop will always be stupid and careless - it's not an AI issue as much as it's a human issue.

I did read your whole response but didn't want to make this too long by replying to each section.

I think the search engine to AI comparison discounts the abilities of an AI, but I get your point. Other than that, I might quibble with an example or two, but I largely agree and think most others would as well.

The issue I believe most have isn't with AI itself, though it gets shorthanded that way, instead it is the combination of the lack of safeguards and the fact that as a race, we are absolutely stupid and careless. Knowing that about ourselves, it seems absurd that we currently have no control on its development, and little to no controls on its use. I think most would be more optimistic if we had either fast progress with some solid regulation or slower progress with less regulation but more time to adopt.

Something with this much potential, good and bad, shouldn't be crashing through the walls Kool-Aid man style.

 

[Pizza Moon]

OpenAI is in trouble.

Likely XAI and Google will be the American winners, anyway.

Infrastructure won’t go to waste though, it’ll just get purchased when OpenAI inevitably fails or faces internal overhaul to right the ship, if that’s even possible.

 

[Mr. Engagement]

The technology I'm really worried about is those newfangled internet forums.

I mean, what good could come of a bunch of random people posting their opinions and photos from other websites?

 

[networkpro]

Calling AI a tool is a very nebulous description. Those other tools don't cause damage like AI does to the environment or strain infrastructure each time they're used to the same degree and often for very frivolous purposes.

Oh like cities, the internet, soap operas, and dollar stores ? All are "tools"

 

[parksandtravel]

Oh like cities, the internet, soap operas, and dollar stores ? All are "tools"

I don't follow the comparability in your analogy, so I'll happily agree to disagree!

 

[flynnibus]

The issue I believe most have isn't with AI itself, though it gets shorthanded that way, instead it is the combination of the lack of safeguards and the fact that as a race, we are absolutely stupid and careless. Knowing that about ourselves, it seems absurd that we currently have no control on its development, and little to no controls on its use.

I don't really understand this concern..

AI isn't showing up anywhere unintentionally. It's not creeping out of any boxes on it's own. It's not developing itself. The usual stunts about making the AI chatbots do something stupid, etc is more an exercise on demonstrating weaknesses in either the prompts or in their learning patterns.

Liability is still with the humans. Just like we expect a business to put someone trained on the other end of the phone - we don't require regulation to say they are some competency except in very specific fields, typically those that require licenses. And all those constraints are still there.

The majority of agnst about AI is the unleashing of bad output on unexpecting audiences -- This again is a human driven outcome - not something AI did on its own. What is happening now is people are upset with AI because people are using it in ways that negatively impact them.

- Customer Service is getting harder
- People are being mislead or decived by people doing nefarious generative stuff
- People are facing career pressures due to how AI will reshape how work gets done
- And the unexpected speed that these kinds of things are happening

So people are getting exposed to bad outcomes - and it's "everywhere and getting worse"... so they have a bad impression of AI and many start repeating the troupes about it.

But none of these topics are about regulation and they will never be part of government intervention to mandate it doesn't happen. Just like the government never stepped in and said "you must provide live customer service agents!"

Areas where we really need protection and evolving conventions on restricting usage really aren't a problem area yet -- because people are being cautious generally in those areas. The rouges who aren't, like DOGE bros are seen for exactly what they are.

Right now people are more upset with poor experiences DUE to AI involved outcomes than they actually know about AI itself.

I think most would be more optimistic if we had either fast progress with some solid regulation or slower progress with less regulation but more time to adopt.

Look at areas like self-driving - that's got plenty of stuff slowing that roll... what are the areas concerning you?

Something with this much potential, good and bad, shouldn't be crashing through the walls Kool-Aid man style.

It's not really - what is different is every tom, dick and harry has it at their fingertips.

Just like the social IQ went down as soon as everyone had a camera+high speed internet in their pocket.

 

[CoastalElite64]

 

[flynnibus]

In other news... NIMBY is still alive and well

Asking anyone to put something they don't want themselves.. what answer would you expect?

In late breaking news.. the general public are horrible long term planners. News at 11.