On tech for tech's sake
| 4 min read
You may well know that I have a newsletter.
My reasoning was simple: I spend a lot of my time reading tech news, so why not share my thoughts on it? I already use Readwise (RIP Omnivore) to make notes on things I read, so surely it would only be a little effort to turn that into useful content for others.
Having had the thought, some hours later I had the perfect system. Using an automation platform and the OpenAI API, I had a completely automated workflow that, once a week, took all my notes from tech articles I'd read in the past 7 days, structured them into an email, generated an intro and outro for the whole thing, and then sent that to me so that it could make final edits before sending it to my adoring fans.
There was just one problem.
It sucked.

The system itself worked flawlessly. And if I'd wanted an AI generated round-up of the things I'd read that week, this would have been fine. Perfect, in fact. But I used the system for about a month, and there wasn't a single thing the AI wrote for me that I didn't rewrite into my own words.
Yes, I could refine the AI Assistant, feed it examples of my writing, and make it better at mimicing me. But the thing is, I like writing the newsletter. What's more, I don't want people to read AI-generated thoughts on articles that apparently have a 50/50 shot of being, themselves, AI-generated.
There wasn't a single thing the AI wrote that I didn't end up rewriting.
When you take that bit out, all the automation is really doing is sending the content through half a dozen different systems to wrangle it into a format that will play nice with the next system in the chain. Aside from that not really making sense, it's terrible practice. If any one of those services changes the way they work, it could break my entire flow.
Here's what I do now:

Notice anything? It's basically the exact same thing I was doing, without the automation block. I'm doing, essentially, exactly the same amount of work with or without the automation.
The human desire to over-use technology
Don't get me wrong. There are plenty of things that automation is good for. And being able to use ChatGPT and others to assist with some of life's more mundane office tasks is great.
But just because these technologies exist, it doesn't mean we have to use them. For me and my newsletter, there isn't really any benefit for me in overengineering a solution. Cramming new and shiny technology into places where there is no need for it just doesn't make sense. And that's before you start to think about the moral and ethical implications of some of the technology we use.
There's often a real fear that if we aren't at least dabbling the latest and greatest, we'll fall behind to the point of obselecence. So let me be one voice on the other side of that argument. It's ok to not always be at the bleeding edge.
I'm not saying stick your head in the sand - there's a balance. Be aware of the landscape and what's going on, and then judge for yourselves (with some professional help if needed...) what the right tools for the job are.
But if you're looking at a new tool that someone is selling you, and you can't figure out where it makes sense to put it in your organisation, there's a good chance that's because it doesn't fit. And that's ok.
Want to read the newsletter in question? Subscribe here. You can also catch up on past issues here.
Want to chat about anything in this post? I'd love to hear from you!
Feel free to drop me an email
Or find me on Mastodon