I often see discussions about the purpose and benefits of AI tools revolve around whether or not they can completely do your job.
The answer is often clear: no, they can’t. If you want, you can always find a way to make AI look stupid.
In a way, this is even helpful: you should always know the limitations of an AI so that you can use it accordingly.
But this is also where my approach differs from many others: If the AI can’t do something, I don’t give up.
Instead, I ask myself a question that I have always asked myself whenever I come across a new tool, a new service, or a completely new medium like the Internet:
What can I do with this that I couldn’t do before?
For example, when I went online for the first time in the late ’90s, I understood: I can publish my own content! That alone was such a radical innovation that I didn’t care what other people thought about the Internet at the time.
I feel the same way about generative AI. Instead of focusing on what it cannot do, I focus on what it can do. And then I look at what is particularly useful.
One small example from my own work: This article summarizes a Reddit discussion with hundreds of comments. It’s available in German and English. Claude gave me the first draft of the article. I edited and checked it, of course. DeepL did the translation – also proofread by me.
In other words: The AI didn’t do all the work for me, but without its help I would never have found the time. The result is only possible in this form thanks to these new tools.
Another example is HN Update: an automatically generated hourly podcast about the top stories and comments on the Hacker News site.
Or Ellipsis News: an app that compiles news on any topic – with the help of AI, of course, and created by a coding layman using Claude 3.5 Sonnet.
I find that exciting.