I've been really energized lately, there just seems to be so much going on in AI right now and it's opening up a ton of new opportunities. Of course, there are a lot of problems that need to be solved, especially around security, privacy, fairness, not to mention the cost of running these systems. It's exciting though and it reminds me a lot of what it was like working on the internet in the very early days.
First things first though. With all this new found energy, I've taken a step back from the tech grind and taken the opportunity to unplug a bit and just focus on a few very important things, part of this means getting the "honey do" list back under control. Stuff around the house needs to be fixed, cleaned, moved, painted, you name it and everyday I try to make a dent in the list. Today was interesting though, I was cleaning up a lot of old stuff that I have collected over the years that was down in our basement. Actually, this is a bit more of a "for me" project because part of the reason I want to clean up some space down there is to build a darkroom. I have always, like since high school, wanted to have my own darkroom. I took a darkroom class at the local university to make sure I remember how everything works and to get an idea of what I will need to buy (the fun part). BUT, before I buy anything, I need to make some space and when you are in your late 50s, have raised three kids and have a basement that just accumulates junk, this is a bit of a monumental task. I threw out the box of wires and cables that could be used to reassemble computers I bought in the 90s. I shredded old utility bills, got rid of old manuals and notes for programming languages that no one cares to remember and then I came across this one box that made me slow down for a minute.
The box had a lot of stuff from right around the time I finished university and from my first job. I found copies of the newspapers, froshbooks, yearbooks and other stuff I either wrote or edited while I was in engineering school in Toronto. I found some notes, network maps and drawings from when I was learning unix and discovering how the internet worked. Then, I found a printout from 1994 of websites, organized by category. It was a list of all the websites I used at work and personally on a daily basis. It's not necessarily ALL of the websites on the internet at the time, but in 1994, it made sense to just make a list of URLs and store them in text file. At that time I was working as a systems programmer/administrator in the physics department at the University of Toronto. Most of the work I was doing at would have involved working on mail and web server software as well as maintaining a fairly large network of various types of unix based servers, workstations, X-terminals, printers, network gear and wiring. This list is crazy though for a few reasons! One, some of the sites are still online like http://einet.net and http://www.x.org and these sites still kind of look the way they did 30 years ago! Two, the sites that are missing from the list because they did not exist yet - no Google, Facebook, Amazon, even Yahoo which did exist at the time was not on this list yet because no one had heard of it. In the early 90s, my friends and I would go down to the pub and debate whether "the net" was something big or just a fad. Having worked on the internet since 1989, I was certainly in the "it's something big" camp but I could have never imagined, when I made that list of websites 30 years ago, that the internet would ever look like it does now.
Fast-forward to today and try to imagine what the next thirty years is going to look like. We have the cloud, computing resources continue to become faster and cheaper and we have a new wave of Artificial Intelligence technology arriving on the scene. I can see where someone might even make a list, similar to the one I made in 1994, that lists the major AI companies that currently exist although this list would already be quite substantial (and nobody prints anything anymore anyway). Many people are complaining that it's all going too fast, the same way we did in the mid 90s about the internet. Maybe the only difference is that while internet development was very intense in the late nineties, we didn't really have an internet yet to advertise that fact. It's not like the internet exploded onto the scene in 1994 and was in everyone's Twitter feed. On the other hand, AI has exploded onto the scene and it's everywhere all at once.
One of the things I'm curious about though is how this new technology will change existing tools like email and the web that we have used for the last 30 years? Looking backwards, we can see how the emergence of the internet changed the way that we wrote letters and used phones. How our shopping experience has changed, how our social experiences have changed. In 1994, when I went down to the pub with my friends we would debate all kinds of things, like the viability of the internet, baseball, music, movies, and other important things. We would enlist people at neighboring tables, servers, and bartenders to help support our cause, like how many Stanley Cups did the Toronto Maple Leafs win and when was the last time they actually won it. I also remember a particularly long debate over the spelling of a word in Boggle which had half the bar arguing - half the fun of the game was to make up words and bullshit the definitions. We would debate things like that for hours that, today, you can just look up on your phone in a few seconds, kind of boring actually, what does an AI fueled game of Boggle look like anyway "Yeah ChatGPT, you are right again - armpitopia is not really a weird fetish"...sigh.
Maybe the AIs of the future will protect us from all of this technology though, just showing us the things we want to see and filtering everything else out? For example, if I have a personal AI, a virtual me, I can tell it that I need a pair of shoes. Since the AI knows all of my preferences, buying habits, and how much I can afford, it can go and hunt the internet for the shoes I want, in the color and size I want, for a good price. The AI would make the purchase, track the shipment and deal with the store's customer service (probably another AI) if there is a problem with the order to arrange a return. In reverse, when I don't need a new pair of shoes, the AI can filter my online experience in real time as I'm browsing the web, reading my email, probably even watching instagram influencers or streaming shows/movies to remove ads from shoe (or any other non-relevant) advertising. Can you imagine if you could only see the ads that you need to see when you need to see them - I would pay for that. In fact, I already do pay for that as I will pay extra for an ad free experience. Instead of running my own email server which is something I did for many years, I pay google to filter ads and spam from my email. Kind of funny how email has gone the same way as regular mail - mostly junk. When was the last time you actually got a letter in the mail that was worth opening. When was the last time someone actually sent you a handwritten letter? When was the last time someone bothered to write you a personal email that was not derived from a template...or written by an AI?
This brings up a few questions. The two major areas of concern are privacy and security. How much information will people be willing to share with their personal AI? Will you be okay with giving it your credit card number? Your shoe size? Access to all of your internet traffic? If you are, how do you make sure that all of this information remains secure? That some AI run by a rogue company does not scam your virtual AI? In fact, we are already seeing these types of hacks in action. This article in Wired talks about how an AI reading email can be fooled by prompts inserted into the email (https://www.wired.com/story/here-come-the-ai-worms/). You may be reading this, shaking you head and saying "No way - I will not be participating in this madness", exactly the same way that I refused to use my personal credit card while testing an electronic store on a website that I built (all by myself...from scratch!) in the late 90s. It's going to happen though. It's not a fad. This technology is here to stay and the problems you are seeing today will be resolved as the tech improves just as the security and privacy problems of the early internet have been resolved. Well ok, somewhat resolved. Keep in mind though that in the early 90s, servers had "guest" accounts that anyone could log into without a password. The concept of firewalls to protect networks was just starting to come online and encypted communication tools like SSL and openSSH were still quite a long way off.
A million caveats though which is nice as it gives me a lot to write about! Thanks for reading, sorry it was a bit long. Next time maybe have your virtual AI summarize it for you ;-) !