"You kinda have to get on board..."
The crew on the Accidental Tech Podcast got into the ethics of AI a few days ago, and co-host Marco Arment (who makes the very cool Overcast app I shared this clip from) basically verbalized my feelings about AI better than I ever have myself.
Listen below:
As someone who's produced creative work, and who has worked alongside creatives — human beings — I have a lot of feelings about all of the knowledge and art models have sucked up to fuel their "brains." And I'll be the first to tell anybody that nothing they produce is wholly unique. It can't be, because it is, at least in some small way, based upon the work of someone else, and it crucially lacks the ability to do something totally new that the world has never seen before.
I also recognize, however, that there are some things you don't need to be creative for, and tools like ChatGPT or Claude or Gemini are quite capable at helping with those things.
I'll give you an example (or a rant, maybe).
In my current role, which is highly classified for some reason, I have zero access to Google Workspace tools. I used to build all sorts of spreadsheet contraptions and scripts in Google Sheets. Now I have Excel, which lives (as does that company that makes it) solely because Microsoft has an in on the enterprise and moving away from it is time consuming and costly.
Google Sheets is Excel that experienced evolution. Excel, then, is basically the Neanderthal of spreadsheet applications, where it hasn't discovered fire yet and believes the secret to keeping warm is to find a nice cave and wrap yourself in the hide of a bison.
Everything in Excel is tedious. If you've used a web app in the last 10 years and had the joy of experiencing an understandable user interface, or unspoken yet intuitive actions such as grabbing a column and dragging it somewhere, well — sorry. Excel never came along for that ride. It's never had to, because, come on: are you going to switch or something?
An LLM is how I survive Excel on a daily basis. It's how I figure out how to delete a bunch of checkbox controls because, who knows why, those controls only show up in the desktop application and not the web version. It's how I learned how to place those controls in the first place because a checkbox isn't a thing you can just drop into a cell — it's an object that has to be built and managed from some completely different interface.
Excel has best ways, best practices for how to do certain things. It's when you combine those things together that you can actually start to get some meaningful work done. For a scenario that is anything but the most ordinary, good luck searching Google for a page about it, or crawling through Reddit to find a post by someone in your exact same shoes.
ChatGPT or something like it, though? It's read enough about how Excel works to know where the tools and actions are, and how to complete certain tasks. And it can string it all together for you. It can put the puzzle pieces together to do the thing you want.
And not just in Excel, but in anything you do that is niche enough to be impossible to search for, yet you know can be done if you just had someone on the God-mode expert level to help you. The fragments are out there, waiting to be pieced together. It doesn't just point to a maybe like a search engine — it combines those fragments and builds you a custom answer you can use. Give it access to a CLI and it can even skip the answer part and do the actual building for you.
That is where the power lies. And your power lies in your ability to understand it and leverage what it's capable of. If you can do that, you are immensely more valuable as a human in the world's workforce. If you refuse and sit on the sidelines, you'll in essence be mailing your resume. It'll be received three days after someone who sent an email or clicked "Easy Apply" on LinkedIn already got the job.
All this — it probably isn't going anywhere. There's been discussion about bubbles. Personally, for myriad reasons, I think there is one! But that deals in the shorter-term and doesn't affect the trajectory for where things are headed in the next 5-10 years.
The dot-com bust of the early 2000s sent a lot of companies into the bin. The internet itself? It did not go anywhere at all. In fact, that flurry and crash set the stage for the companies that have defined the past 15 or so years of our technology lives: Google, Facebook, Amazon, Netflix, etc.
All of those companies became mammoth in this stretch. Early on, excitement got the better of everyone and we overdid it, badly. But the underlying foundation was solid and the trend was there. Today, we do just about everything online.
AI will undoubtedly hit speed bumps and it may even do worse than that for a period. But, choose your metaphor: the genie is not going back in the bottle, the toothpaste is not going back in the tube, and so forth. Someday, we'll look back and remember life before it as we look upon life with it. As with the internet, there isn't an "after."
You don't have to like it. But I implore you to learn it.
