The recent news and hype around AI motivated me to create a project on my personal computer with limited resources, using minimal code[0] to combine ChatGPT[1] with computer vision. This efficiently extracts structured data from virtually all websites I visit regularly, even though the resulting text may be somewhat poorly formatted. Despite the text's imperfections, ChatGPT is still able to provide me with summaries and insights. For instance, it can compare prices from various e-commerce sites, detect discounts and different payment methods, analyze soccer match results, and even generate humorous morning news summaries based on the top headlines in spanish!.

However, the recent "letter of warning against AI" penned by Elon Musk and other influential individuals has led me to question their motives. I suspect they may be losing the competitive edge they once had, as users can now effortlessly obtain information in ways that were previously unattainable. It's possible they are seeking to lobby for new laws that restrict this kind of data extraction. With such simplicity, I can envision a future where people no longer need to browse web pages, ultimately making ads and the whole business of treating your users like crap irrelevant.

[0] 10 lines to screenshot with puppeteer, tesseract to OCR and a clustering algorithm, add to this whisper and a nanotts to get an enhanced Unix philosophy+AI experience

[1] Im pretty sure i can use low resource models to do the summaries and insights, any recommendations?

* edit,im pretty sure that could have done this PoC years ago without GPT.