Chris, hold my Diet Coke (sorry I don’t do the Dew, to much sugar and caffeine, I have enough problems). As you point out Apple has built their business model on hooking users into proprietary software and hard ware. the comparison to Microsoft as the polar opposite is well thought out, but not quite applicable as Microsoft makes only software for others that manufacture hard ware. Also you are quite correct companies change cables to obsolete the hardware, which really is annoying, speaking as someone who has a 16-prong VTR and cable but only a 1984 monitor with the correct receiving port. It is the comparison to medical records and data that does not work. You have countless clinics and hospitals that collect data but it is stove piped from city to city, county to county, state to state, county to country. Not to mention the vast array of systems and software that house the data. Perhaps one could start with the VA as they have data on veterans nation wide supposedly in a single medical records system. However you would run the risk of lack of diversity in the population as individuals that are veterans may not be a representative sample of the population as a whole. Just just my thoughts.
As usual my clever, experienced readers jump right past the end and into the future. I’m breaking this problem down into a series. First up: Is it possible for a massive AI to cure cancer? Yes! It’s possible. Next up: How would we build such a thing? We’ve started this off with the look at workflow; we’re going to next dive into agents, CAGs and RAGs. Finally we’ll look at the practical obstacles such as your point about the current stovepiping of data. Spoiler alert: Information Age solutions require Information Age data architectures; specifically, Distributed Data Management Systems such as PrivacyChain.
Chris, hold my Diet Coke (sorry I don’t do the Dew, to much sugar and caffeine, I have enough problems). As you point out Apple has built their business model on hooking users into proprietary software and hard ware. the comparison to Microsoft as the polar opposite is well thought out, but not quite applicable as Microsoft makes only software for others that manufacture hard ware. Also you are quite correct companies change cables to obsolete the hardware, which really is annoying, speaking as someone who has a 16-prong VTR and cable but only a 1984 monitor with the correct receiving port. It is the comparison to medical records and data that does not work. You have countless clinics and hospitals that collect data but it is stove piped from city to city, county to county, state to state, county to country. Not to mention the vast array of systems and software that house the data. Perhaps one could start with the VA as they have data on veterans nation wide supposedly in a single medical records system. However you would run the risk of lack of diversity in the population as individuals that are veterans may not be a representative sample of the population as a whole. Just just my thoughts.
As usual my clever, experienced readers jump right past the end and into the future. I’m breaking this problem down into a series. First up: Is it possible for a massive AI to cure cancer? Yes! It’s possible. Next up: How would we build such a thing? We’ve started this off with the look at workflow; we’re going to next dive into agents, CAGs and RAGs. Finally we’ll look at the practical obstacles such as your point about the current stovepiping of data. Spoiler alert: Information Age solutions require Information Age data architectures; specifically, Distributed Data Management Systems such as PrivacyChain.