Saturday, 8.37pm
Sheffield, U.K.
Automation is no longer just a problem for those working in manufacturing. Physical labor was replaced by robots; mental labor is going to be replaced by AI and software. – Andrew Yang
A month ago I started to use AI as a research assistant.
I used it to read long reports and give me a list of key points. I then read those points and selected the ones that looked interesting, wrapped them in posts and published them.
This is what I found.
1. AI doesn’t make boring stuff more interesting
Some of the reports were about important but dull things. What’s happening in the metals and minerals sector? How are hotels decarbonizing?
These topics matter, but if you’re going to read a post like that you’re probably already interested in the subject.
This showed up in the viewing statistics. The posts didn’t get too many views. There was some interaction but that was mostly from experts. And experts in what you do are not your target market.
I did notice on someone else’s post that a unwary end user talked about her experiences. She was promptly buried by an avalanche of sales requests.
2. It does make you, the writer, more informed
It’s funny how often someone would talk to me and ask a question related to something I had just written about using this process.
As a professional you need to be informed about what’s going on in your market whether you use AI or not.
Writing regularly forces you to look things up and that discipline is valuable.
3. Summaries are ok but detail is magic
AI summarises material but detail is what connects us to the idea.
In a recent post, for example, the AI summary was “reducing BIC Cristal barrel weight from 4.4g to 3.1g through value engineering”.
But I had to read the detail to learn that they did that by making the interior profile of the pen a hexagon to match the outside.
It was previously circular.
That detail is more interesting than the phrase “value engineering”.
Specific and concrete details matter and that leads to the next point.
4. AI is useless with well written work
Some books are a one sentence idea stretched over 40,000 tortured words.
Others are 5,000 carefully crafted sentences.
The first type of book is not worth summarising.
The second cannot be summarised.
AI’s true value may be in improving mediocre work so that it is at least average.
I think we quickly start to recognize the difference between manufactured words and human ones – there’s a sameness to the output, a ultraprocesed quality to the taste of the words that puts us off.
Of course, it’s going to get harder to tell the difference, just like a plant-based burger is really no better than a mean-based one.
There will be, I suspect, a flight to quality.
This happened to me with Kindle Unlimited. The idea of lots of books was great but you only found rubbish on the package while the good writers kept their books off and you still had to buy them.
People we already trust will get more of our business.
In commercial writing, however, the stuff you don’t read on websites for example – all that will be AI generated.
Because no one cares about that stuff.
5. Why do you write anyway?
I don’t know about you, but I write to think.
A first draft is me telling myself the story.
This is often rough and rambling. Most of it is rubbish.
But…
There’s a sentence or two that might be interesting. A fragment that makes it into a second draft. An idea that sparks a question that needs more research to answer.
You could spend time with a prompt engine to do this or you could just do the work because it’s interesting and because you want to understand it better.
The key point is that all this “stuff” has to make its way into your brain for it to be of any value.
You’re the consumer. And so you’re the most important person in the room.
Final thoughts
There are some tasks, like coding, where AI has changed everything.
Programming jobs are the easiest to automate because you don’t need a large team when a small team with AI can produce more and better code.
I don’t think AI is going away. I do think individuals and teams that use AI to help them create better work will do well.
People that try and simply use AI as a replacement for great people or teams will end up with mediocre and average results that will be ignored.
And, in today’s world, being ignored is the worst thing that can happen to your business.
Cheers,
Karthik Suresh
Handcrafted insight 