ChatGPT Introduces Lower-Priced Subscription Tier, Ads Coming Soon

[BeatCrazy]

Genuinely curious but what do you people use AI so much for that you need a subscription in the first place? I ask it occasional questions that should be common knowledge and thus saving me a Google search and many clicks.

You can do many amazing things with AI besides as an enhanced search agent. A few days ago I had it build me an entire Excel business model with inputs and outputs. Something that could have been done manually over the course of several hours, but Copilot built it in minutes. If you weren't an Excel wizard, you probably couldn't have even done it yourself.

 

[jole]

On the positive note: I like that they are giving customer a choice.

 

[Regular_John]

"it ensured that conversations with ChatGPT will be kept private from advertisers, and that it will never sell user data to advertisers."

<insert anchorman I don't believe you gif here> 🙄

 

[Frantisekj]

Let’s not forget the fourth tier, which is completely FREE if you don’t use it at all.

May also shows that their business model is not effective and needs more income to sustain that or persuade investors that their investment are not in danger.

 

[JSRinUK]

I’ve gone from being very enthusiastic about AI (as we call it) to being very jaded.

As a fun exercise, it can be amusing. It can help learning things like coding. It can generate stupid pictures. All fine.

But when the companies behind them try to force feed it into every aspect of our lives where it doesn’t need to be, it starts to feel insidious. When AI becomes the “single source of truth” yet a truth that is manipulated to appease the government of your country, it’s clear we’re heading down a dark path.

For a brief time, I paid for Grok. I don’t any more. I wouldn’t dream of paying for access to an insidious “chatbot”. I do currently pay for Claude because I find it useful upskilling my coding knowledge.

But, for anything else, I hold it very much at arm’s length.

We have the ability to run our chosen LLM locally these days. We can run local AI tools like image generation. I find it difficult to justify a purpose for frontier AI. Yet so many people and businesses already appear to not be able to manage without it. I hope that’s just an “appearance” and not indicative of the real world.

 

[Frantisekj]

I’ve gone from being very enthusiastic about AI (as we call it) to being very jaded.

As a fun exercise, it can be amusing. It can help learning things like coding. It can generate stupid pictures. All fine.

But when the companies behind them try to force feed it into every aspect of our lives where it doesn’t need to be, it starts to feel insidious. When AI becomes the “single source of truth” yet a truth that is manipulated to appease the government of your country, it’s clear we’re heading down a dark path.

For a brief time, I paid for Grok. I don’t any more. I wouldn’t dream of paying for access to an insidious “chatbot”. I do currently pay for Claude because I find it useful upskilling my coding knowledge.

But, for anything else, I hold it very much at arm’s length.

We have the ability to run our chosen LLM locally these days. We can run local AI tools like image generation. I find it difficult to justify a purpose for frontier AI. Yet so many people and businesses already appear to not be able to manage without it. I hope that’s just an “appearance” and not indicative of the real world.

Probably too early to call it AI. It is great for some purposes but not all, differs model to model a also results can vary very much based upon who is asking. You has to be one who controls outcome to minimize risk of nonsense.

 

[ChrisMac47]

"Artificial Intelligence, AI" is a terrible description, there is no intelligence whatsoever. "Machine Learning, ML" is better and as far as I'm concerned the machines are still learning! Even simple searches are frequently inaccurate or incomplete, so why would I trust ML to do more complex tasks?

Let's all re-visit this in 5 years time, after the bubble bursts and people have had time to reflect.

 

[Zondar]

Folks, this isn't hard. Install Ollama (ollama.com), then pick a model that fits in your memory, and you have your very own, fully private LLM at your disposal.

Even models small enough to fit on a low-end laptop are becoming quite good. For example, on my M4 macbook air (16 GB), the gemma3 12b model works smoothly and provides excellent, comprehensive responses. There are many others, including those tuned to specialized tasks. Of course, if you have a studio, or any other mac with more unified memory, you can fit larger models too.

The above is very easy to put into use. Try it! 🙂

 

[aforty]

You can do many amazing things with AI besides as an enhanced search agent. A few days ago I had it build me an entire Excel business model with inputs and outputs. Something that could have been done manually over the course of several hours, but Copilot built it in minutes. If you weren't an Excel wizard, you probably couldn't have even done it yourself.

Thats cool and yeah I use it for work too, I guess I phrased my question poorly as I meant for personal use that would require a personal subscription that isn’t paid by work to do work things.