AI versus Marshy #3: Is keeping up with it a problem?, Photoshop still requires skill, + hackers


Hey Reader,
​
Welcome to another week of AI versus Marshy!

This week we cover:

  • šŸ“ˆ Is there a first-mover advantage to AI? Probably not.
  • šŸŖ› Tool of the week: Adobe Photoshop’s Generative Fill
  • šŸŒŽ OpenAI + Bing is pentesting the entire web

I will also throw in a not-so-humble brag about shipping videos if you make it to the end.

Let's tuck in!

I can’t keep up with AI - will this be my undoing?

There’s a podcast/Youtube out there called Marketing Against the Grain.

I wouldn’t recommend it to most of my readers as its super-techy and in the weeds.

This recent interview with growth marketer Guillame Cabane from Hypergrowth partners made my ears prick:

Host: One of the things we’ve said on the show is that the first-mover advantage has never been more prevalent in the AI era. The first-mover advantage is even more so in that era compared to any of the other era because everything gets commoditised so fast.
​
Do you agree with that?
Guillame: Yeah, absolutely. Absolutely. Which is why you need to be able to move fast and test things repeatedly.

But does it?

Data has taken a long time to monetise.

  • Uber for all of its disruption still hasn’t monetised to profitability.
  • Google wasn't the first search engine.
  • The iPod wasn't the first MP3 player.

Is there really an advantage for people to move fast and break things?

Let’s look at this with a first principles approach.

Farnam Street is an incredible blog and nerds out on thinking models - it’s worth reading (and absorbing again later) their take on how to do this.

We can question how we’re feeling about something by asking why we think one way over another, what are our sources of evidence, and understanding the consequences of being right or wrong.

Another method (and one featured in my last book) is the 5 Why Technique - when done well, you can uncover what the real reason for wanting something is.

It’s usually not ā€œkeeping up with the latest trendsā€ and closer to ā€œso I don’t feel silly or left behindā€ or ā€œto help me feel like I’m doing SOMETHING when I feel powerlessā€.

So do YOU need to start using AI tools for your business this week?

Break the question down and see if its necessary.

  • Do I believe its necessary?
  • Do I absolutely know this to be true?
  • What other sources point to this being true?
  • How are other people thinking about it?
  • If I’m wrong about this what is the impact?
  • How about the consequences if I’m wrong?
  • So do I believe this is necessary now?

The chances are no.

It’s easy to get caught by hype.

And the hype machine is nowhere near the peak yet, so check yourself before getting swept away and think from all the angles with a strategic lens.

Tool of the week: Playing with Adobe’s Generative Fill

I’ve never been able to justify the cost of Photoshop due to:

  • a) Not being a full-time graphic designer
  • b) And sucking at it

I make do with Pixelmator Pro and Keynote (a slide maker) and its got me this far today.

However, I was really curious about Generative Fill after seeing Adobe’s slick demonstration video and wanted to see if I too could move a deer into a wet alley at night.

So I signed up for my 7-day free trial to explore this idea.

The answer is no - I can’t.

I did click around a bit and asked it to ā€œplace lots of greens and zeros and computers near the centre of the screen for someone to be be bashing their cyberdeckā€ - which isn’t worthy of sharing.

My steampunk version wasn’t that much better.

This could definitely be a PEBCAK thing.

The world’s largest cybersecurity experiment

Something intriguing about the rise of generative AI is that we really are playing with something that’s grossly unpredictable at best.

Advocates for the technology point to the outsized benefits are worth the unexpected risks that emerge as the progress roars forward.

And because it’s hit the web and starting to be noticed by consumers - we haven’t witnessed things going wrong at this scale ever before.

Google had the luxury of millions of tests before pushing a change to its search engine live. Facebook would regularly roll out changes to 1% of its audience (or test them in Australia first funnily enough) before making them available to all.

Because OpenAI and ChatGPT isn’t working this way, we see quirks pop up, get an immediate course correction, and the machine keeps moving.

I use the paid version of ChatGPT and a funny note appeared saying I can’t use the Bing plugin at the moment.

This plugin lets you use ChatGPT with the current web to search for information and bring it ā€œinā€ to your chat window.

An example I’ve used recently was - ā€œtell me the contact pages for golf clubs near Northbridgeā€.

The popup shared a page that said:

So you can see a URL’s full text?
Big whoop.

What you might not know is that some websites parse information from the URL to their own database. Some marketers for example will put a user’s email address in the URL bar to track who arrived at their page and how to contact them further.

Now - how was this uncovered?

Would a sophisticated hacker be able to exploit this functionality at pace and gather enough information for malicious intent?

Probably.

It was a blink and you’ll miss it moment - but we’re entering the web’s largest open-sourced pentest* and I’m not sure what casualties there’ll be.

*a simulated cyberattack to uncover any vulnerabilities before bad actors can find them

–

I’ve bashed out a week’s worth of videos with speedmarketingdemos and aiming to get 30 under the belt by the end of the month.

My goal is to build up my speed, punchiness, and confidence delivering at this point and think by then I could add more services (on-demand marketing coaching/encouragement and done-for-you marketing asset creation are the ideas winning at the moment).

If you’ve got more ideas or suggestions I’d love to hear them from you.

We’ve got this šŸ’Ŗ

-Marshy

​

Early Growth School

Continuous tips, lessons, advice, and direction on growth for your early-stage startup by a growth marketing pro.

Read more from Early Growth School

a braindump. Hi Reader, I disappeared for a while. I’m sorry - late last year I joined a startup, then earlier this year went full-time contract on it with a high-risk, go big, or go home kind of bid. We’re now in the latter stages of that bid and at the ā€œomgwtf I’m not sure if this is going to work outā€ stage, and look: I got what I signed up for. It’s been eye-opening and gratifying jumping onto the tools and building a growth engine for another business and if it doesn’t work it won’t be...

the results are in. Last week I said I had a bit of spare time and wanted to test more video with shorts. This is what happened in the last week: Not bad right? Well actually in the scheme things - it’s unprecedented. Here’s the lifetime view count of my channel: The number’s pathetic. It’s never been a focus, and while I love presenting, have never been inclined to go hard at it because I have other things to do (like work for clients, raise my family, get sober and completely transform my...

short example

a tierlist for growth channels. Have you heard of tier lists? They’re very common in gaming world. Influencers like to put together lists ranking characters, items, or selections and opine on them. I had a quieter week as I work through a project exension and thought: Do you know what the world needs? A tierlist for growth channels for startups. But not just any tierlist, but grabbing and ranking the channels from the book Traction. Written in 2015 Here’s the channels from the book: •...