Eli5: How did we go from zero ai, to one company with an ai that holds a conversation, and then seemingly immediately on to multiple companies with ai that makes music, photos, and movies in such a short period of time?

1.65K viewsOtherTechnology

For real. One day there’s nothing. Then chat gpt shows up. And then the next day ai can create ANYTHING and there are a whole bunch of companies with software.

In: Technology

39 Answers

Anonymous 0 Comments

There was not nothing, the entire AI tech tree has had a gradual decades long buildup. Its just that Transformer based Large Language Models and Diffusion models for image generation (which include a sizeable language Transformer of their own) are the applications that finally made the headlines, caught the public attention and blew the minds of people unfamiliar with the tech.

Anonymous 0 Comments

To add to all the smart stuff people are saying here, most ‘AI’ tools you see across social media are simply ChatGPT wrappers. Essentially third party suppliers that give their customers ‘fill in the blank’ prompts for specific purposes – CVs, marketing copy, blogs, you name it.

Anonymous 0 Comments

We did not. Since everyone has provided the answer already I’ll just add that this applies to a lot more than AI – just because it’s only discussed in the mainstream media today doesn’t mean it hasn’t existed or developed until that day.

For example: lab grown meat – there has been a lot of work on getting production of it cheaper and on a bigger scale. Sometimes it makes the news every other year. One day there will be a breakthrough and you’ll have a bunch of companies trying to catch up to that market by utilizing the technology, and suddenly, it’s big news. But in actuality 10 years ago this was already being worked on and the technology developed since then makes it easier for everyone to mimic the formula so to speak. And when you have a lot more attention and money on something then development becomes faster.

Anonymous 0 Comments

A new way of labeling data called the transformer model and hardware advancement is the answer. These 2 factors are what enabled AI to go from thousands of datapoints to billions of datapoints.

Anonymous 0 Comments

When it comes to software solutions you only ever have to solve it “once” then it can be spread and used world wide in minutes.

Its much much faster than any physical progress that has to be built each time while software only ever builds on itself.

Anonymous 0 Comments

At first, we had basic tools, and then someone came up with a fancy, versatile tool, like a Swiss Army knife, that could do a lot of things. That’s kind of like ChatGPT — it showed what AI could do in terms of conversation. Then, people realized, “Hey, if we can make an AI that talks, maybe we can teach it to do other stuff too!” So, they started tweaking and improving the technology. Just like you can add attachments to that Swiss Army knife to make it even more useful, developers added new capabilities to AI.  Now, about making music, photos, and movies — imagine you taught that Swiss Army knife to also paint, take pictures, and play music. It would become a super versatile tool. That’s what happened with AI. Once developers figured out the basics with conversation, they started adding more “attachments” or capabilities to make AI more versatile and creative.

Anonymous 0 Comments

Because one could just as easily say that we are still at zero AI.

The AI label is being put on everything at this point. I half-expect to see potato chips “flavored by AI” in the grocery store on my next visit.

We still don’t have artificial intelligence and we may still be very far away from that.

We do have some semi-impressive software tools that are being called AI, but they also could be called many other things.

In other words, AI is the marketing word currently being used to talk about the latest advances in computer software and tools. Some of those tools aren’t very reliable either.

Anonymous 0 Comments

That’s not the case. There’s a lot of music text and image generation before chatGPT. chatGPT just provides a better UI to bring clever generative AI to the eyes of the public.

Anonymous 0 Comments

Investors. Right now if you are looking for a cash influx in 2 hours tell them you have an ai product. Boom. Cash today.