Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
If AI brings the answers to us, then less traffic to websites. Then, those websites lose revenue, shut down. Then, where do these AI scrape data from? Maybe it will start scraping its own historical outputs in the ultimate narcissistic move. LOL. There is a part of me that still wants to explore and find the answer. This is nothing more than a curiosity to me at this point. It has great potential for some uses, though.
Good question. I imagine one possibility is AI search will be heavily monetized either through ads or making it not free, and websites would get a cut based on how much AI traffic they get. Each query would be a tiny micro transaction. So websites could continue to exist. But lot of questions there.
 
Crypto was the last meme, AI is the next meme. When crypto rose, GPU price rose. GPU price crashed when crypto crashed. The next meme is AI. All for the selling of GPU.
 
  • Wow
Reactions: gusmula
I mean, you can't really trust any answers from these things anyway, so you should still visit the websites they cite to verify that the bot hasn't lied or done stupid math.
Then that defeats the purpose of having the bot in the first place. It won't really help search, and just help students write their term papers, especially if it provides the references to them. I do worry for coders, as it can some rudimentary code and that ability could improve over time.
 
I mean, you can't really trust any answers from these things anyway, so you should still visit the websites they cite to verify that the bot hasn't lied or done stupid math.
True. As impressive as recent large language models are, this is still a tough problem. We don’t know how to train an AI to provide correct answers as its actual goal and instead can only imitate that goal with simpler proxies like “responses that sound convincing to some humans” which might provide correct answers. Maybe.

Computerphile has done great videos on how AI actually works and what the current limitations and challenges are.
 
Then that defeats the purpose of having the bot in the first place. It won't really help search, and just help students write their term papers, especially if it provides the references to them. I do worry for coders, as it can some rudimentary code and that ability could improve over time.
This right here. I'm worried about programmers as well, and I am one of them. In even five years, who knows how far along Chat GPT will be. I asked it to write a shell script to update PeopleSoft run controls by modifying the date field on the first of each month and it all but wrote the entire shell script for me. The only thing I had to do was type in the specific parameters needed for my use case. Crazy.
 
This right here. I'm worried about programmers as well, and I am one of them. In even five years, who knows how far along Chat GPT will be. I asked it to write a shell script to update PeopleSoft run controls by modifying the date field on the first of each month and it all but wrote the entire shell script for me. The only thing I had to do was type in the specific parameters needed for my use case. Crazy.
I asked ChatGPT to write a small Python program and was shocked at how good the result was. SkyNet, here we come.
 
  • Like
Reactions: gusmula
I asked ChatGPT to write a small Python program and was shocked at how good the result was. SkyNet, here we come.
Right!? Of course it has a long way to go before it can write massive programs without a lot of bugs/issues but it will happen sooner than later. Hopefully not before I retire!
 
  • Like
Reactions: gusmula and dbwie
Right!? Of course it has a long way to go before it can write massive programs without a lot of bugs/issues but it will happen sooner than later. Hopefully not before I retire!
I'm a psychologist who does therapy. AI does not deal with emotions well, but you can ask it for an empathic response and get something decent in return. I have 10 years until retirement, hope I can make it! Laughing, and sort of not laughing about this.
 
At this point I can’t blame them too much. Even with that restriction there is still a waitlist, and even with the waitlist their servers are apparently getting hammered since everyone is very interested in this right now. It may open up more later, or they may decide people want it enough to use Edge to get it.

Personally I’m terrified to try this with my personal Microsoft account. I don’t want the AI having access to the information in my account.

I assume since I have a MS Dev account they (AI) already have that info.
What I am concerned is filtering of my search queries via the AI. This looks more like an additional layer instead of a new method.

JMO
 
  • Haha
Reactions: gusmula
I'm a psychologist who does therapy. AI does not deal with emotions well, but you can ask it for an empathic response and get something decent in return. I have 10 years until retirement, hope I can make it! Laughing, and sort of not laughing about this.
I think you will be safe in your field for quite some time! :)
 
I'll admit it, like many of you I thought it was all hype when I first heard about ChatGPT last year. Then I tried it...

Large Language Models like ChatGPT and BingAI are literally going to revolutionize the world. If you haven't spent any time with one I highly suggest you open up an OpenAI account and give ChatGPT a whirl (or just spend some time with Bing although there are some caveats with this version) Siri is literally nothing by comparison.

They can respond to complex input, actively engage with the user, answer complex questions, create convincing output in multiple languages, write code (!), summarize, review and even create content, and complete work in seconds that would take humans hours (if not more.)

When BingAI first released to preview last week it was incredible. Compared to ChatGPT which is not connected to the internet, BingAI could do complex searches, pull accurate information from the internet, do complex calculations, had better foreign language support, and, if left as it is would've been a must have for almost everyone. The things you could do with it were truly mind blowing, and is what finally sold me on the "fourth industrial revolution" an amorphous term I previously felt was just a buzzword lacking in substance

That said, this could've been an earth shattering moment where Microsoft seizes control of not only search, but a lot of office type productivity. Unfortunately as anyone who was in the preview around a week ago knows, the version of BingAI being rolled out today is vastly inferior to the one available initially. While the preview version had it's problems, and could be provoked into going off the rails, it made ChatGPT (already very capable) look like a child's plaything in comparison.

I'll say it once: Don't sleep on this. Especially if you do white collar work. Your future employability may be on the line.

AI assistants are literally on the cusp of going from a joke (Siri, Alexa, etc) that is sometimes mildly useful (set a timer/reminder/alarm) to something that will be indispensable to our lives and change the workplace forever.

ChatGPT just screams for regulation by various governmemnts.
Talk about an insane system. Literally.
 
ChatGPT just screams for regulation by various governmemnts.
Talk about an insane system. Literally.
And, governments themselves will be working on their own versions of AI. It is already happening according to some reports. Imagine an AI system reviewing reams of intelligence reports, finding possible issues of concern. Or scanning a foreign countries social media to help shape diplomatic responses to tense issues. When you start thinking about computers analyzing human language and finding patterns that we cannot find ourselves, the possibilities of how it could be used are mind boggling.
 
Last edited:
I assume since I have a MS Dev account they (AI) already have that info.
What I am concerned is filtering of my search queries via the AI. This looks more like an additional layer instead of a new method.

JMO

A dev account is one thing. Hopefully it’s not tied to the login of any devices with data that you care about.

Connecting an AI to the system that controls licensing, login, etc for computer systems is a little too close to Skynet.
 
A dev account is one thing. Hopefully it’s not tied to the login of any devices with data that you care about.

Connecting an AI to the system that controls licensing, login, etc for computer systems is a little too close to Skynet.

It isn't. However, I do have an exchange account, MSO and other items. While different accounts, my name is tied to each. These accounts started back in the days prior to Android and iOS :cool:
 
Yep. I feel like their software engineers have no time to improve Siri or fix bugs. They’re too busy writing code to suggest photos to use as backgrounds or featured photos.

It’s all about priorities and I just don’t understand how they clearly don’t see Siri as a top priority. Is it really possible they don’t know how bad it is?

How can they spend so many years and resources working on a car, and be completely unaware of the fact that people are reluctant to trust Siri with anything more important than a lightbulb? Talk about buying the cart before the horse.
 
Said without a hint of irony.

Screen Shot 2023-02-23 at 1.26.40 PM.png
 
Crypto was the last meme, AI is the next meme. When crypto rose, GPU price rose. GPU price crashed when crypto crashed. The next meme is AI. All for the selling of GPU.

Yes there is a multi layered corruption happening.

They hype stuff that they know is very buggy, can't scale well and is environmentally destructive but as long as it makes Nvidia and other chip shareholders richer they hype this crap.

Crypto people have always known that the real money invested in their pyramid scheme always goes out the backdoor to mining chip makers, miners and electricity companies. They have to make fake price charts to lure suckers to throw more money at them. They have always know things like bitcoin can't scale well and is clunky.

Something similar is happening with this AI wave.

They are describing magic powers to these chatbots that they do not have. They are terribly clunky and easy to confuse. They hallucinate random misinformation. They would consume an incredibly high amount of energy to be used as search engines. ChatGPT can't provide links to sources when they chat because the links aren't saved in the training data.

But they consume mountains of GPUs, disastrous for the climate and will result in mountains of e-waste and high energy bills for the rest of us.

But billionaire shareholders will be laughing at the mess they put in our laps.

They commit the speeding offences but we get the speeding ticket.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.