The "AI-first" trap 🪤
When all you have is a hammer, everything starts looking like a nail
New technologies are always exciting. They push the limits of our imagination, spark a desire to innovate, and create the next-generation of companies to disrupt the status quo.
But sometimes, we fall in love with technology too much. We stop thinking clearly, we see problems where none exist, and we end up forcing a square peg into a round hole.
Maslow’s Hammer
I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail.
– Abraham Maslow in The Psychology of Science (1966)
The law of the instrument, or Maslow’s Hammer, is a cognitive bias that involves over-reliance on a familiar tool. When a tool becomes a starting point, we’re much more likely to apply it everywhere, even when it doesn’t make sense. We overuse the tools we know, instead of evaluating other alternatives that might be simpler or more effective.
Some would call this “solutions in search of a problem”.
Don’t build a Juicero
Juicero was a startup that developed a $400 fruit and vegetable juicer that required a Wi-Fi connection and mobile app to operate. The company also sold proprietary juice packs for their juicer, and each had a QR code that would be scanned by the machine to verify it was a Juicero-made pack and not some third party. They raised $120 million from investors including Google Ventures, Kleiner Perkins, and Thrive Capital.
In 2017, 4 years after the company was founded, a Bloomberg report showed that the Juicero Press was pretty much a useless and over-engineered device. They proved that you didn’t even need the $400 device, because their juice packs could be squeezed by hand and give you the same result.
If we back-track to 2013, the tech world’s attention was around “Mobile” and “IoT”. If we look at this list of “Top 10 Strategic Technology Trends For 2013” from Gartner, we’ll find them them at the top of the list. To be fair, these were important trends to take note of. The Ring doorbell, founded in 2013, got acquired by Amazon for $1 billion five years later even after failing to secure a $700,000 investment on Shark Tank.
If there’s one thing to learn from the opposite fates of Juicero and Ring, it’s that success depends on your ability to define a real problem and solve it with the most suitable technology.
The next time you’re designing and building a product, ask yourself: “Is there an easier way to squeeze the bag?”
Fall in love with the problem, not the solution
The world is filled with problems, both big and small. Most of the hate and criticism that AI receives is because it’s being directed at non-problems, especially when it comes to creative work that people enjoy doing.
What if we started by understanding and listening to the real problems we want to get rid of and see if AI can help us there? Isn’t that what technology was meant to be for?
Instead of falling in love with AI or any technology, we should obsess over our ability to notice and observe the problems that are worth solving.
Practical Advice
If you want to avoid the AI-first trap, try these out:
Do a task analysis: break down the end-to-end process of how people complete a specific task. Here’s a guide from Nielsen Norman.
Define the job-to-be-done: Clayton Christensen’s jobs-to-be-done framework is one of my favourite ways of defining the purpose or desired outcome of a solution. Here’s a guide from the Christensen Institute.
Start simple before increasing the complexity: Gall’s Law states that “a complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works and cannot be patched up to make it work. You have to start over with a working simple system.”
The simplest way to solve a problem is often the simplest to implement, disseminate, and maintain over time. If you start with something complex, the odds of failure and costs increase.
It reminds me of a quote from football legend Johan Cruyff:
Playing football is very simple, but playing simple football is the hardest thing there is.
– Johan Cruyff
Agreed: https://open.substack.com/pub/ahopefulfuture/p/abundance-first-a-reasoned-case-for?r=ueg3l&utm_medium=ios
This is so critical to talk about, and needs to be talked about a lot more. Especially with AI where I think the consensus is to just use the brute force method in building a more general AI model. What would be true genius would be a model that solved for one specific problem.
The real genius of simple products, like the hammer 😃, or I like to use Calendly in the digital space, is that someone had to say that all of those features are good and we will put them on the backlog. Eventually, various features thought to be good were removed during a prioritization session. My guess is that people proposed hammers with all kinds of bells and whistles too, but no one was willing to buy.