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
It wasn’t overnight. Eliza was around forever. Chatgpt just was trained using the internet so it wore useful and generalized than most.
All this started in the 70s already. We had the theory for natural language processing, neural networks, reinforced learning etc. But the technolgy was not there. The most powerful computers weren’t capable of running these algorithms in a way that could achieve more than mere expert systems or chess programs. AI research was therefore for some time a niche. But Moores Law was in full effect the whole time, so now we have the computational power to basically scrape the entire internet (also a factor, as the current flood of content just didn’t exist until 10 years ago) and train neural networks with billions of parameters.
It was there the whole time, but just right now several factors and technologies are converging to make it possible (highly powerful computer systems, data mining in a vast internet, easy internet access). And of course there is now an incentive to market it. You could not monetize things like the Deep Dream algorithm, but now that you can get “useful” things from those systems that you can make money from, every investor is of course throwing their resources at it. But the theory has been there for about 50 years.
>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?
By completely ignoring the decades of AI research and development that lead up to it.
Tensor Flow and the focus on neural nets was big push back about a decade ago. I mean, jesus man, did you not hear about [AlphaGo](https://www.youtube.com/watch?v=WXuK6gekU1Y)? Deepblue?
How the hell does this sort of false premise get upvoted?
AI is a tool. When you use a tool to make more tools, research and development become faster and better resulting to better AI tools.
This also explains why people are concerned about AI development. If it develops beyond our control and it becomes self sufficient, it CAN result to human extinction.
Welp, little Billy, you remember when you first started seeing those CAPTCHAs everywhere asking you to click or type specific things? Welp, little Billy, that’s been AI the whole time, and *you’ve* been training it! Yup. You’re complicit in the downfall of humanity. Tough vitamin to swallow, I know. That’s why we buy the chews, but I digress.
Marketing teams like to latch onto buzz words. 10 years ago everything was cloud. 5 years ago it was block chain. Now is it glorified chat bots being called AI when they aren’t AI.
To quote C3-PO… “shut me down, machines making machines”
Once you get an AI with infinite access to every bit of digital information ever put on the net. It can do things sooooooo much faster than humans give the same timeframe.
So we get constant ai improved ai’s.
We didn’t go from “zero ai”. AI has been around for a very long time, it’s just finally getting mass amounts of data to be used for the things that we see it being used for in the mainstream.
Yes, computer scientists have been working on AI since computers have existed. My impression is that the big factor that eventually led to all the current success was when scientists began to borrow ideas from how our brains work… and created “neural networks “. Is that right?
I had a conversation with a computer around 8 years ago, we talked about HAL and why he was being such a lil b, I remember being absolutely amazed at how natural it all seemed to be!
Latest Answers