A profitable use case for AI is like:
A customer service bot that runs on your own local machine. You can configure all of it and let it query your inventory to tell customers whether something is in-stock. You can program in thousands of frequently asked questions, let customers look up order status, find tracking numbers, hours of operation, account information etc. It's a sophisticated chatbot that does everything a customer service worker does.
The best way to implement this sort of thing is probably to set up a camera and record what actual workers do at their desks, then use that data to program the AI.
These use cases wind up being extremely specific: tax accountants, customer service workers, call center employees, purchasing agents, travel agents, match makers, consultants, research assistance, business process analysts, intelligence, and so forth. You do this by literally watching the employee work and then making the bot do the same thing.
They will continue to piss away billions until the jig is up and retail investors are left holding the bag, which point the industry may actually pivot to something useful. Making people's jobs obsolete is a specific process on a job-by-job basis where you watch and mimic. True off-the-shelf products come after an exhausting process of watching thousands of professionals work their jobs. Eventually the bots are talking to each other, negotiating prices, haggling, booking travel, finding matches, and streamlining the whole process. AGI is going to require this huge on ramp of real world experience.
No comments:
Post a Comment