Now that PaLM 2 is going to run a GPT4ish model on my Pixel, I wonder if there's the same need for a bulky laptop. I mean, I can code with Replit. I can use ChatGPT from my phone and that will have a code interpreter released broadly soon, along with a diverse ecosystem of plugins that enable web browsing, reading up to date documentation,or awareness of real-time events. I can use a prompt generator python script for Midjourney using the GPT4 API that I wrote with the assistance of GPT4 to take cool prompt ideas from Twitter as input and expand on it for practice. I could start a Discord channel, invite my friends, make a GPT4 api(or PaLM 2(Android app?))powered Sci Fi AI dungeon master, and my friends and I could build the project in little time with help from Clyde, Discord's built in AI that can literally collaborate with us on ideas or make suggestions or write cool and useful things, like personalized tutorials. I can make music with my Pixel with MusicML(albiet eery at times), and use beta apps that put the power of sampling the real world around me easy(something Dilla dreamed of, I'm sure). The major inconvenience is the small keyboard. I dislike typing on it. That's short lived though, since voice commands are on their way, and I direct people to perform different types of tasks at work, so now I'm just directing AIs to do different work. So many things I do today will have a version that exists in this new world of AI and AI-accelerated technological developments. Nvidia literally uses AI to design application specific integrated circuits for AI. From the ground up, the AI stack is AI accelerated. AI progress is exponential, not linear.
