What is AI actually good for?


“Let’s brainstorm the best ways to get the next 1000 email subs for VP”.

This was the message I sent to Claude over the weekend, using their new “even more powerful” Opus 4.8 model.

What followed was 90 minutes of me telling it the ideas simply weren’t good enough.

Despite…

… all of the training I’ve given it on my biz
… multiple “readjustment” messages
… telling it what kind of ideas I was looking for

The ideas it came up with simply werent good enough.

It’s ideas ot grow the email list of VP were the most milquetoast, generic, and outdated approaches. Here’s a small sample for you.

  1. It recommended the same kind of free asset Digital Marketer used to get 350,000 leads - upon questioning it, that was a case study from 2013 and used FB ads back when they were cheap as chips - DENIED
  2. Run a free giveaway was next - Something I’ve done a lot of and have never failed to be disappointed with the lead quality - DENIED.
  3. Ask subscribers to tell their friends - Which I will do, but VP is a brand new biz with a tiny subscriber base - DENIED.

The whole session solidified something for me about the usage and the approach to AI in your business.

What AI is good at…

AI and LLMs right now are essentially prediction machines.

They have a huge amount of data and have understood the patterns between each element within. All they do is predict what’s most likely to come next - whether that’s an action, word, piece of code - based on their training data.

A prompt or skill will guide it and shape the way it picks what’s next. But, essentially, it’s choosing what works best in the majority of other, similar situations.

To simplify, if I was to say 1, 2, 3, 4, 5…

It could guess 6, 7, 8, 9, 10.

In most cases, that’s gonna be good enough. it actually makes AI really good at very templated, generic things.

So, for example…

  • If you want to do more social content. AI will be able to take a thought from you and reformat it into the same templates all of those big influencers use for every post.
  • If you want to draft something like a VSL, it’ll be able to hit the story beats the biggest in the biz have created.
  • If you want to create a simple website or element like a popup, it’ll knock one up for you right quick as it’s got hundreds of examples to draw from.

It’s insanely effective at doing this lower level grunt work.

But, that is where the real value of AI ends in my opinion.

I’ll caveat this here by saying this is still insanely valuable. A ghostwriter for social might cost you $500 - $5000 per month. Claude will cost no more than $100 / month.

Same again for developers and designers doing low-creativity work (templated ads type thing).

What AI is terrible at - and should stay with YOU

By design, AI is not good at thinking outside of the box and being creative.

As evidenced by my request for effective marketing strategies.

It basically came up with the stuff that’s ranking #1 on Google for “growth strategies” created to sell a BS service rather than a case study of something new and unique.

And again, that extends to everything AI does.

It defaults back to the most common.

  • For growth strategies, you get what everyone else is doing and so just become one of the crowd all trying to get results by doing the same thing - which makes getting that result harder
  • For design, everyone ends up with the same design which means those visually arresting ads/images you need are impossible cause you’re playing into the old “banner blindness”
  • For code, it’s not going to create a brand-new, never-before-seen service, feature, or approach that turns an industry on its head.

In essence…

AI lacks soul

It is a terrible creative thinker, and an even worse taste maker.

That stuff has to stay with you if you want to stand out.

I’m not saying it’s useless.

Far from it.

AI is great for things that are, essentially, creative chores, the kind of things that sap you of energy but aren’t really that impactful.

For example, years back, we did a huge breakdown of Noom’s marketing that still does really well to this day.

That was a solid month of work of research, analysis, creation, editing, hair pulling, and frustration.

But, there’s nothing else out there like the Noom study we did.

You know where I fell short on that?

Promotion.

After doing the 10,000+ word report, I couldn’t find the energy to turn that into…

  • Linkedin posts
  • Tweets
  • instagram posts
  • etc

I did a few, but it was inconsistent and honestly, low-effort.

This is where AI could have helped, with the “creative chore” of taking an original piece of work and creating multiple templated assets from it.

I could have crafted a skill trained with Social templates, fed it the original and had it craft two dozen posts within minutes.

These posts would be generic in their format (the kind of formats proven to work on a social platform - same shit any ghostwriter or mass producer of content is using), but unique in the content and idea because they’re pulling from my original piece.

AI is incredibly useful for you and for anyone running a business, but ONLY if it’s not being used to take away from the kind of thought processes that will help you stand out from the competition.

We’re entering an era where everyone is gonna be using AI for everything.

If you also use it for everything, everything you do will be the average of everyone else’s work.

AI should not be doing your thinking for you.

It should be taking your thinking and running it through the processes that are little more than time sinks and mental burdens.

Giving you more time to work on the truly creative things.

Any Qs, drop em below.

And if you wan help to implement this stuff in your business, consider joining us in GM+ here - https://growthmodels.co/community

Pete "using AI the right way" Boyle

Subscribe to Growth Models