Gates and AI - November 2023
Bill Gates has a great breakdown of where he sees AI going in the next couple years. He covers the positives and the negatives both a bit, which is something I always appreciate about his analysis. I find it interesting that nowadays I frequently refer to "my assistant" when having conversations. Just yesterday I was in an architectural discussion and I asked the agent to verify if I was remembering some stuff correctly. And my agent and I delved deep into SAST and SCA as well, which are some concepts I haven't worked much on. Somewhere along the line I stopped posting stuff to Facebook to my "AI Overlords" album because it's very accurate to say I use it daily. It doesn't surprise me like it used to, now it's just...part of my day.
One of the things I find frustrating about the advent of AI like this is the way people react to it. Some people, like me, embrace it. I have it installed on my phone, I've used it in stores. I have one friend who even more fully embraced it and built a business case at his job so he could expense it to his employer. Other people have probably never even looked at ChatGPT and all they know about it is whatever they've gleaned from news reports. Others seem to actively dislike it or focus on the flaws and imperfections. The last group frustrates me. It smells of Luddism to me. (tangent - I actually had "Ludditism" but I asked my assistant if that was right and they gave me "Luddism" instead and gave me 3 paragraphs on where the term came from). What I find most interesting is that some of the folks who disparage or dislike AI are in my profession which most would think has a fundamental need to understand and embrace new technology. This is entirely incorrect by the way - lots of people in technology aren't interested in progress or learning new things and even actively push against new stuff. They...aren't curious. I don't get it, I don't like it, but it is surprisingly prolific. In contrast, I've been raising my kid with as curious and adaptive a mindset as I can. I think one of the most useful things a person in this world can be is curious...and I think that will only grow more important in the future and I think that AI agents are going to be a huge factor of that.
And yet I won't let Squirt use AI unsupervised, which I think is interesting. He's coming up on being adulthood fast and he's seen me using it and asked about it and I said "no" without hesitation. "Why?" he asked immediately because that's how he's been raised and how I want him to approach the world. "This doesn't make sense to me, please explain." There's that curiosity. And I didn't have the best answer at first. So I've thought about it a bit.
My first answer I gave him is that I want him working out more for himself first. I want him to figure these things out on his own. I frequently compare my childhood to his and one of the things I've noticed is that things come easier to him because of technology. He doesn't have to only learn how to read visually, he can listen to audiobooks. He hates commercials because streaming services don't make him watch them generally. He has endless options for entertainment and keeping his mind occupied compared to my childhood. He doesn't need to work as hard at stuff, and that's not a bad thing - progress is about making things easier. So I create hurdles for him. Take away his audiobooks sometimes, make him sit down and read visually. Watch a movie with commercials. Restrict his electronics usage. Make him ride in the car with me without a tablet and look out the window instead. In my head these things are making him a better person because a little bit of a struggle is a good thing.
My next answer was a variation of "people are stupid." What this boils down to are my concerns with the internet in general. I spend a ridiculous amount of time on the internet and...people are fucking stupid. Not me, of course. But everyone else. And ChatGPT is just a distillation of the internet which is really just a distillation of people in general. It's too easy for all of us to put stupidity out there and that stupidity has invested ChatGPT and...I want him to be older and more mature and more ready for that. I've WATCHED his mind and critical thinking abilities expand...and I want these things to be wider and deeper before he starts swimming in humanity's sea of idiocy.
The next answer was...ChatGPT is POWERFUL. Scary powerful and scary easy. And everyone knows what Spider-Man says about power, right? When Squirt’s helping me build things he can have a hammer and nails, but he's not getting a nailgun. I'm not handing him my table saw or my circular saw. These are wildly useful tools, allowing me to do things it would take me vastly longer to do without them. But it takes nothing but a little inattention to take a hand off or a punch a nail through your foot. Adults do that all the time, I'm not trusting the child with it. And the same is true for ChatGPT. I don't have specific scary things he might do that would be harmful but I know he can do so much more with it than without.
Another thought, and this dials back in a bit to Gates' "agents" in the piece, is that there is an element of dehumanization to ChatGPT. This is something I've noodled a bit in my head, and it's a theme in Science Fiction as well. A power tool is simply a force multiplier. It makes you stronger. But an agent is so versatile it can make you "stronger" but it can also allow you to offload cognitive load. It doesn't just make you think faster, it can think for you. Depending on how you use it you'll find yourself forming opinions or making decisions without having actually done any thinking yourself. You may have just translated the question out through your fingertips and then taken the prescribed action and executed on it. You've delegated a decision out to a distillation of the stupid sea of humanity that has no consciousness or soul. Where does that end? Will you ask it for relationship advice? Decisions about friendships? What will you not ask it about, and why not? If it gets smart enough and you use it enough how much of "you" is even going to be left? A question for poets and novelists and philosophers I suppose.
I didn't intend this to become a treatise on parenting in the modern age as well as introspection on technological progress but here we are. The world is shaping up to be a fascinating thing and I think as different as my childhood and my kid's childhood is...that margin will be vastly larger when I compare my child's early adulthood to mine.
And because it amuses me, here's ChatGPT's summary of what I wrote:
The author reflects on the current state of artificial intelligence (AI), sharing Bill Gates' insights on its future. The author mentions using AI, particularly ChatGPT, in daily life, embracing its convenience. However, the author expresses frustration with those who actively dislike or focus on flaws in AI, likening such skepticism to "Luddism."
The author discusses their parenting approach, allowing technology but setting boundaries. They cite concerns about the ease of access to information and express a desire for their child to face some challenges for personal growth. The author also notes concerns about the internet's content and the potential dehumanizing aspect of AI, raising questions about the extent to which individuals might delegate decision-making to AI agents.
In the context of raising a curious and adaptive child, the author acknowledges the potential impact of AI agents on shaping opinions and decisions. The piece delves into the complex relationship between humans and AI, pondering the role of technology in personal development and decision-making.
