The Catastrophic-Job-Loss Doomers Are MILES Out Over Their Skis

Those who enjoy writing or speculating about technology -- and those who get paid to -- make this mistake over and over again: They overgeneralize special situations.

It doesn't matter whether you're closer to the bottom or the top of your organization's hierarchy; sudden, large-scale job loss has bad effects on everybody, full stop. Your job might remain, but other aspects of your life will suffer. There are quite a few smart, serious people who fully expect 20% plus of the worldwide economy to become unemployable inside of 10 years. I really like what Daniel Miessler is doing with his Unsupervised Learning newsletter, but here he is with this kind of prediction:

Source: https://danielmiessler.com/p/ul-416

https://danielmiessler.com/p/80-jobs-go-away

You might notice the BCG study that Daniel referenced. That's a special situation. That's skilled, competitive, elite knowledge workers (the BCG employees) trying out amazing, powerful, poorly-understood, unreliable software and gaining an advantage from it. Good for them! But that is a special situation.

Special situations don't often generalize easily or quickly. (That's what makes them special!)

Geoffrey Moore argues convincingly that powerful, desirable software is nothing without an ecosystem of support, including best practices, training, skilled vendors, and other supporting infrastructure. And for the most risk-averse organizations, the other thing they need is an overwhelming rate of adoption by others in their industry. In other words, the technology ecosystem could be perfect in every way on day one of its release and such organizations would resist adopting the technology for years if not decades, simply because not enough other companies have done so yet.

Efficiency or other forms of advantage are not enough to drive immediate, widespread adoption of any innovation, including AI.

As part of thinking this through, I had an illuminating conversation with ChatGPT about the University of Oregon's "speed freak" offensive strategy. This chat is worth sharing because looking at businesses through a non-business lens can be enlightening.

ChatGPT: The University of Oregon, under coach Chip Kelly, popularized an up-tempo, no-huddle offense that emphasized speed and rapid play execution. This "speed-freak" approach has since influenced numerous American football programs across both college and professional levels.

In college football, teams such as Baylor, Texas A&M, Oklahoma State, and Auburn have adopted similar hurry-up, no-huddle (HUNH) offenses, achieving significant offensive success. For instance, these teams finished among the top in total yards per game during the 2012 season. While it's challenging to quantify the exact number of teams that have adopted Oregon's specific approach, the principles of the hurry-up, no-huddle offense have become widespread.

Philip: If this approach has an advantage why have not 100% of teams adopted the technique?

ChatGPT: While the hurry-up, no-huddle (HUNH) offense offers advantages such as controlling game tempo and limiting defensive substitutions, it also presents challenges that deter universal adoption:

• Player Conditioning and Depth: Maintaining a high-tempo offense requires exceptional player fitness and a deep roster to sustain performance and mitigate fatigue-related errors. Teams lacking depth may struggle to implement this approach effectively.

• Complexity and Execution: The HUNH offense demands precise communication and rapid decision-making. Missteps can lead to procedural penalties or turnovers, making it less suitable for teams without experienced personnel.

• Defensive Adjustments: Defenses have developed strategies to counteract fast-paced offenses, such as simulating the HUNH in practice to improve preparedness. As defenses adapt, the effectiveness of the HUNH can diminish.

• Strategic Preferences: Coaches may prefer ball-control offenses that emphasize time of possession, aiming to keep their defense rested and control the game's pace, contrasting with the rapid nature of the HUNH.

• Situational Use: Some teams reserve the HUNH for specific scenarios, such as after a big play or when the offense is lethargic, rather than employing it as a constant strategy.

These factors contribute to the selective adoption of the HUNH offense, as teams weigh its benefits against potential drawbacks within the context of their unique rosters and strategic goals.

The mistakes that we make when we base our guesses about the future on an overgeneralized special situation include ignoring or underrating the importance of:

• Incentives that might be contrary to adoption of the new tech

• Unavoidable constraints that might interfere with implementation

• Completely different ideas about how to secure competitive advantage

• Personalities and culture that resist rather than welcome the innovation

• Turf wars that slow things down

• Mistakes and poorly-governed transformation initiatives

• Misinterpretations that cause mis-application of good ideas

• Differing context that causes something that worked great over there to fail over here

• Opposing forces that exploit things like increased legal liability

I feel bad when I have to take on this role. In my work and life I'm a heavy user of state-of-the-art AI tools. I love, love, love the new capabilities they bring within easy reach. I do not love the friction, complexity, and negative inertia the real world often presents. I am confident that AI will be a net positive for the world and will not involve catastrophic >20% unemployability.

But any vision of change -- any dream of improvement no matter how small -- has to work with rather than ignore this real-world stuff! Otherwise, it's not an inspiring vision; it's a fantasy.

Special situations can become the normal, general situation. Innovative stuff can become everyday ubiquitous technology. In our lifetimes, we've seen that with personal computers, the internet and World Wide Web, mobile phones, wireless networking, digital music and portable media players, the Global Positioning System, Genetically-Modified Organisms, solar energy generation, carbon fiber and high-tech plastics, robot vacuums, electric cars and bicycles, additive manufacturing, telemedicine, and wearable health monitors, just to name a few.

But the process of a special situation becoming the general, usual, normal situation takes time and effort and a symphony of factors working together. You know this from your own experience of trying to drive change in your organization.

The special case can be a fulcrum that helps drive innovation, but it's a long road from there to widespread implementation.

Despite my criticisms, I'm happy that Daniel Miessler and others are showing up with the message that they are. They're wrong in the right kind of way. If you squint to obscure the details you see this message: "The world is changing, and you can't avoid adapting to change with it. Don't complain; change instead!" I agree 100% with this message, even though I know firsthand the pain involved in saying yes to this kind of change.

I hope you do too. It won't involve catastrophic job loss.

Further Reading

• Evaluations are all we need by Rohit Krishnan explores how replacing human jobs with machines is a complex, technical challenge requiring careful consideration and systematic evaluation. It's a dense, difficult read only because the underlying subject is a dense, difficult topic (and, again, the difficulty of replacing an actual, real-world job with actual-real world automation is underrated by the folks who are predicting 20% plus unemployability).

• Bringing Elon to a knife fight by Jennifer Pahlka examines how entrenched bureaucracy resists disruption, offering insights into the universal challenges of institutional change. The article is about the American context, but the lessons are universal.

Did someone forward this email to you? Want your own subscription? Head over here and sign yourself right up!

Back issues available here.