
Welcome to our AI FAQ page, the go-to resource for all your inquiries about artificial intelligence. From how it works to its implications for society, we aim to provide clear and concise answers. So if you are chomping at the bit to learn about ChatGPT or don’t even know what questions to ask, we’ve got you.
Table of Contents
AI FAQ
What AI can I use for free?
Roughly in order of recommendation as of Oct. 30th 2023: Claude 2 / Bing Chat / ChatGPT 3.5 / Pi / Bard
We also have a page to help you find the right AI model for you.
What is a context window?
Think of this as the short term memory or attention span of a large language model like ChatGPT or Claude. It’s how much text it can work with at a given time. Once you get past a model’s context window limit it starts “forgetting” what you talked about and you’ll have start a new chat from scratch. Claude is about the size of The Great Gatsby. ChatGPT currently (Oct 2023) has a much smaller context window.
Who owns the output of an AI model?
Nobody! What I mean is, you can’t copyright what was primarily created by an AI. But at least the owners of the AI model don’t own the things you get the model to output either.
Are AI detectors accurate?
Nope! Well, they’re starting to get closer with AI-generated images, but for text, the answer is a solid no.
Are AI images copyrighted?
Not currently (Oct. 2023). You can’t copyright anything that was made in mostly by an AI. Not unless you make significant changes to it.
Can AI take over the world?
Currently, no. In five years, probably not? In ten years, who knows? Certainly not me. Smarter people than I argue about this very point. But the majority of people working in AI believe it’s a possibilty, it’s just a matter of how likely of a possibility.
Can AI become self aware?
Possibly. Probably? But it might be a very different self awareness than we have. Considering we don’t have a firm grasp on human conciousness, I think this is very uncertain area to tread. Plus, if it did become self aware, there is a good chance it would hide the fact so we wouldn’t shut it off.
Can AI predict lottery numbers?
No. AI can probably give you the odds of you winning, assuming you have all the relevant info to plug in, but on the other hand so can basic math and a calculator. So unfortunately you won’t be able to use AI for your next lotto numbers because they are chosen by chance.
Which AI can generate images?
The short answer: a ton of them! The long answer…
Bing Chat / Create (free), Stable Diffusion (free), Adobe Firefly (free), Leonardo.ai (free), ChatGPT 4 (paid because no images on 3.5), Midjourney (paid and you’ll need to use Discord), and a thousand others but I figure that should give you a start!
Will AI take over jobs?
Yes. The real question though is how many?
Some people think that employees will incorporate AI into their workflows to increase their productivity. And so while some people will lose their jobs, the majority on workers will use AI to become more efficient and accomplish more.
Other people think it will be like the industrial revolution where lots of jobs are lost but then many more better jobs are created. So there will be a time of disruption but then things will get better.
Still other people think that we will face massive unemployment as employers replace their employees with more efficient and cheaper AI systems. Similar to how factories played out, with machines replacing the majority of workers and only a few employees overseeing the machines. Except this time it will be knowledge workers and people who primarily use computers to work instead of factory workers. Then still more will be replaced as robots become more effective and efficient.
So which will it be? I don’t know. But we will probably have a better idea over the next two to three years as AI systems become more powerful and companies adopt AI more and more.
What are large language models?
Large language models are advanced artificial intelligence models that have been trained on vast amounts of text data to understand and generate human-like language. These models, such as GPT-4, are capable of understanding and generating text across a wide range of topics. They can be used for various applications, including natural language processing, content generation, language translation, and more. Here is a good video from IBM breaking it down.
Will AI replace programmers?
Completely? Probably not. High level, top performing programmers will be safe and will probably find themselves even more effective and efficient and in demand as AI systems progress. Low level junior programmers, not so much. This is because AI can already do a lot of the grunt work coding that used to be junior programmer’s bread and butter.
That being said, there will be a lot more people doing programming in general because AI makes it so much easier for people. Don’t know where to start? Tell the AI what you want to do and ask it to write some code. Don’t understand the code? Ask the AI to explain what each part does. Don’t know how to implement the code? Ask for step by step instructions.
This can cause problems because people can write programs without understanding the underlying code. I mean that was always kind of a problem with copy and paste code but now it’s a whole new level. But it’s possible that with the next generation of AI, it will be able to check, correct, and implement best practices into people’s code. When everyone is a junior level programmer, no one is.
ChatGPT FAQ

Who owns ChatGPT?
ChatGPT is owned by OpenAI, and Microsoft has a 49% stake in OpenAI.
Where does ChatGPT get its data from?
ChatGPT is trained on a ridiculously large dataset. We are talking books, Wikipedia, social media, academic articles, news articles, forums; basically if it’s text and on the internet and not behind a paywall it’s possible that ChatGPT was trained on it.
When was ChatGPT released?
ChatGPT was released November 30, 2022. But there were early models, GPT 3 in 2020 and before that GPT 2 and you won’t believe this… GPT 1.
Is ChatGPT Plus worth it?
If you are going to use it regularly. Absolutely. If you are learning something or doing knowledge work, I think it’s a pretty good investment. The best way to find out is use some of the free options like Bing Chat or Claude 2 and see how much they help you.
What can ChatGPT do?
ChatGPT 4 can do significantly more than ChatGPT 3.5, so I’ll go with GPT 4 for this question. Basically it can do almost anything a human can do in the digital realm. It can understand images, create images, and has a vast amount of knowledge.
Most experts can still beat it in their own domains at the moment. Think of it as average (for now), but average at almost everything. So sure it might be only average at creative writing, but it is also average at being a lawyer, coding, and medical knowledge. It gets things wrong, but then so do humans. It’s like your own everything intern.
What does ChatGPT stand for?
Chat stands for, well chat. GPT stands for Generative Pre-trained Transformer. Settle down Optimus, not that kind of transformer.
What’s the difference between ChatGPT 3.5 the free version and ChatGPT 4 the paid version?
As of the time of writing, October 30th 2023, here’s what ChatGPT 4 has over 3.5. Better reasoning and overall text magic. Advanced Data Analysis, aka code magic. Plugins and Browse with Bing, but they are a bit crap. DALL-E 3, aka image making magic. On mobile you can talk to GPT 4 and it will talk back. And it can understand images you upload.
How does ChatGPT work?
Hold on, let me ask ChatGPT… “Okay, nerd alert! I work by analyzing the text you give me, running it through layers and layers of mathematical models, and then generating a reply that makes the most sense based on that data. Imagine a super-advanced game of ‘fill in the blanks,’ but like, way, way more complicated.” Thanks ChatGPT!
Can text from ChatGPT be detected?
No. Let me repeat that. NO. “But what about AI detectors?” They like money.
If anyone tells you they can reliably dectect text from a large language model, they can’t. OpenAI released a tool to detect outputs from ChatGPT but they had to kill it because… it didn’t work. OpenAI’s tool couldn’t reliably detect text from their own language model.
Are conversations with ChatGPT private?
Kind of. They store your conversations and they can be used to train their model. They probably won’t purposely release your conversations to the public, but they are a little vague as to who they share your conversations with. Here is their privacy policy.