As George Lucas once said, for some reason: history is like poetry. It rhymes.
Our industry is certainly rhyming with a historical era I love to pick on: the boom in tech investment that followed the one-two punch of the Dot Com Bust and the Financial Crisis. Companies and their investors were desperate to find that one silver bullet that would turn everything around. Back then, that silver bullet was Design Thinking. Big tech poached designers from celebrated consultancies and airdropped them into pure chaos without a plan beyond “fix the product for us.”
Ten years after design thinking’s heyday, there’s a new silver bullet that everyone has pinned their hopes on, without a clear idea of what exactly it will do. At this point in the hype cycle, you’ve seen the numbers for yourselves: billions in investments, and loud cricket noises for returns. Anecdotes of individual productivity gains are easy to come by, and yet they do not translate to improvements in the bottom line!
What’s actually making teams move faster is org design. Taking care of organizational dysfunction may be a side effect of “AI transformation” but it sure seems to be offering the majority of the benefits.
Consider this story when the right handpicked people completed in 3 hours, a piece of work estimated at 10-12 weeks. That story was from roughly ten years ago, well before any of the AI tools appeared on the scene. It didn’t rely on magical tooling, it just needed us to undo some organizational dysfunction and to put the right people in the room to collaborate on the problem.
The purpose of a system is what it does
“AI works” vs “AI doesn’t work” is an exhausting debate; no one is ever convinced of anything and the only winners of the arguments are keyboard manufacturers. I’m not going to go over the various hallucination rates and studies on how much longer it takes to check an AI’s output for errors than to just do it yourself. Even with 0% error rates, those figures would be besides the point. The biggest problem — and don’t take my word for it, listen to some Nobel laureate guy — is that the bottleneck lies upstream of your Jira ticket:
Most research on AI productivity is overblown because it overwhelmingly focuses on easy, well-defined tasks where context is clear. These are not representative of the economy, which simply isn't set up with so many of those.
We are not in the future. It is still 2016. It is still 1986, if you are so inclined. We have hardly made any strides in our ability to define tasks and their context. What baby steps we’ve made in the right direction have been swallowed up by a great cloud of noise. For every Toyota Way book written in the 90s, there is a Ford counterexample being eagerly practiced today.
AI has intensified this problem, as it intensifies all things to which it is applied. But the problem did not originate with AI. Before Claude could learn to talk like the median LinkedIn influencer, the median LinkedIn influencer had to arrive at that mindset under their own power:
Criticism of the […] slop genAI produces invariably fails to acknowledge how much slop-ification was already characterising so many discourses prior to retail LLMs: PEEL paragraphs, Powerpoint slides, TED talks, Big Consultancy strategy canvases, PR company produced apologies to customers, LinkedIn posts...
The bottleneck is the context in which the work takes place
For every step a would-be thought leader took towards “synergistic leadership” or “growth-hacking paradigms” they took a step away from reality. And corporate bullshit (not my name for it but a real term of the art, take it up with Cornell!) is not the only layer lubricating a product’s slide into slop. Robert Fabricant has a good layman’s overview of how little a modern product is influenced by a given design decision, and how much is a consequence of business and technology decisions that stack up over time without a single decision-maker deliberately steering things to be so. I recommend giving this one a read, as a means of cognitive estrangement from the ways of working we take for granted, because we are too close to them to realize how ridiculous they really are.
Unfortunately, many tech leaders do not take the time to reflect on whether the system actually makes sense. They shirk the responsibility given to them — to build an effective system — and instead whip the existing, faulty system to run faster and faster, until all the wheels fall off:
“Ship more faster now” is not a product strategy; it’s a psychosis. …
Reframe craft as care for collaborators, users, and the collective aspiration to create something useful, meaningful, and significant. Craft is a social act that binds teams around shared values and takes time.
But how do we make this reframe happen? How do we seize control of our discipline’s story?
That’s where my head’s been at for years — before the Picnic, before my work for NN/g, even before I started publishing with the UX Collective. But only lately have those thoughts come together into something actionable enough to share. I’ll be writing on this topic through July, but the capstone of these ideas will be my one-day workshop with Throughline on July 31st.
Flex those professional development budgets and join me in taking back control of the conversation.
— Pavel at the Product Picnic