Dial-Up AI

I spend hours every day talking to AI through text prompts and screenshots while building concepts, and it definitely feels like using dial-up internet from the past. The AI is smart, but our communication methods are very slow. I type explanations, upload images, and describe what I want. And the AI responds.. I correct it, upload more screenshots to show what went wrong and we go back and forth with each back and forth taking forever. But there’s a bigger bottleneck: even when the AI gives me the perfect code, I still need to see the results with my eyes, run the application, check if it works (at least human-ly), take screenshots to show problems, then explain again.

Eventually we’ll have direct brain-computer interfaces i.e. transferring complete context directly to the AI without typing or screenshots, just thoughts flowing from the mind to machine. AI will probably interface directly with your visual system too, letting AI systems see the application directly. Similar to dial-ups in the past, AI development feels very slow right now. The intelligence is getting better and faster every day, but we are limited by keyboards and screens and our slow human hands.