Raw LLM Responses

@rosevee35 Great question. You AI sceptics use the same arguments for your own field: “Anyone can pick up a pen and start to draw, it’s already at the ultimate level of democracy” and “The amazing art is created by super talented and people that have trained they whole life to acquire their skills”. So right back at you, which one is it? From the perspective of AI generating art, you can very fast and easily generate a finished image, in the same way I can use my phone camera to instantly get a photo. For the camera example when the result wasn’t what I wanted, I can quickly tweak the result by moving the subject, moving the camera or adjusting things in the background, and afterwards I can tweak the contrast, vibrancy, saturation, crop the photo etc. So I can get an immediate result or put more time into it, and if I was a skilled digital photographer I might made my own computer scripts that can instantly apply a series of modifications at once, because I don’t like to waste time. Similarly there are many different models and AI tools available to an AI-artist. If you want it to look like a photography, you can actually write the model of the camera you would like to use, different values for FOV and such, and this will be respected by the model. You might have created a workflow where you move the image between different models because you know the strength and weakness of each model. You can have your own reference images, tell and tell the AI that you want to combine specific aspects of each into the final result. But the speed and cost and the result you get at entry level is unprecedented. And while I would have preferred if I could find a more original and even better analogy, digital photography and digital video is quite good, as you can also see the good and the bad that can come from making it easy to create in these mediums. Sure we now create more photographs than we will ever look at again, maybe devaluing each individual photo, we have made it possible for anyone to capture anything spontaneous without having to carry a bulky camera around or wait for the result to be developed. But I personally use AI most in programming. And even here we have entry level ways to utilise it, when you want to do a similar change in 10 files you can have it do it in seconds instead of wasting an hour of your time. Or when the thing you are about to program requires 6 different skills and you are not very proficient in 2 of them, the AI can fill in the blanks and you can read the code, ask follow up questions etc. But for skilled usage of AI for software development, you set up a system with multiple AI agents with different tools that can automate different parts of your development cycle. The most basic right now is an agent that looks at code a human (or AI) is suggesting should be added to a product and analyses if it could lead to bugs or is insecure etc.