“Move fast and break things!” has always struck me as the silly mantra of a gaggle of immature brats whose “things” were always safe and in no danger of getting broken. Now, many of them have mega-yachts and bolt holes around the world, where they think nothing they possess can get broken. I hope they never find out otherwise, because their subjects will eventually rein them in. I am seldom optimistic, but I can be unnaturally hopeful at times.
About a year ago while looking for something else I saw that Forbes had tried semi-convincingly to split this baby without killing it. Earlier this year Evan Selinger and Albert Fox Cahn published Move Slow and Upgrade: The Power of Incremental Innovation (Cambridge University Press, 2026). I have a shelf of “little books” that includes books I recommend when friends and coworkers ask me where I get my “unconventional” perspectives on matters arising. This book will be added to that shelf.
The authors had me with their dedication “For those who do the hard work, walk the long path, even as others seek shortcuts.” I have been battling shortcut seekers for my entire scientific and academic career, sometimes successfully.
The case studies in Move Slow and Upgrade include:
- Zuckerberg’s Mythological Metaverse
- The Crypto Con
- Home Security Upgrades
- The Failed Promise of Covid Innovation
- Moving Fast and Breaking Schools with Remote Proctoring
- Upgrades in the Age of Generative AI
- Upgrading Hiring
These are followed by the concluding chapters:
- Cybersecurity: The Land of the True Upgraders
- The Upgrader’s Mindset
What follows are brief discussions of selected chapters in Move Slow and Upgrade. The key distinction in the book is between innovation and upgrade. Selinger and Cahn begin with the Innovation Trap, a good example of which is a new car with controls that are downright dangerous, mostly in the form of touchscreens that perforce require the driver to take his or her attention off the road. According to BMW this is because younger drivers (i.e., those under 45) love their screens. I suppose money has nothing to do with it. Fortunately, other carmakers are returning to buttons. This was a selling point for the Mazda MX-5 (jpeg) [1] I bought last year after my treasured 15-year-old MINI Cooper convertible was crushed in the front end (no injuries) by the inattentive driver of the typical hypertrophied American pickup truck (a Honda that is probably not sold in Japan) while I was sitting at a red light in the dark rain waiting to turn left.
In general, innovation is the generator of hype, usually unfulfilled, while an upgrade is motivated by the considered hope that the change will be for the better. Or in the words of Selinger and Cahn, “Innovations are when we use a dollar to buy a lottery ticket (actually $2, I think); upgrades are when we put a dollar in the bank.” What follows is a short discursion into the case studies, beginning with Zuckerberg’s Mythological Metaverse, [2] where:
Few innovation mirages are as enticing to technology companies as the “metaverse,” the all-encompassing digital platform that is so immersive that it rivals our physical world in what we consider real.
As my English friends might say, “Bollocks!” The Zuckerberg metaverse is, or was, a method to evade a reality in which you actually cannot do almost anything you can imagine. The hidden reality of the metaverse is that mostly what we have done is give up our privacy.
One of the more absurd notions associated with the metaverse and its companion virtual reality is that it would increase empathy among its acolytes. Studies show that it did, on the surface, enhance emotional empathy, which is “fast, automatic, and occurs spontaneously.” This is not news to moviemakers and television producers, who have been very good at producing tearjerkers for more than a hundred years. However, cognitive empathy is a more advanced mental activity, “a deliberate skill…that requires attention and effort to decipher the thoughts and feelings of the other person.”
The authors of that particular study propose this limitation occurs “because virtual reality leaves too little to the imagination…virtual reality might deprive us users the skillful mental effort required to put themselves in someone else’s place.” This is a feature, not a bug with the metaverse, and as Selinger and Cahn note, “the truth is that when technology brings us together, we still behave as ourselves.” They might have added that way too many of us have been taught, by poor examples, to behave as online avatars instead of human beings. Read the comments of the typical website. The “killer app” of the metaverse that brings people together by default has eluded our grasp, but its hype has made the masters of the tech universe wealthy beyond imagining.
And even Zuckerberg eventually came around to “Move fast with stable infrastructure.” That clanks loudly off the rim, and he never really meant it. What I have wanted to ask him is this, “Don’t you understand that the expression ‘that’s so meta’ is not a compliment?” Considering another of these companies, did the founders of Google really mean “Don’t be evil”? Perhaps, at first, but it did not last, as Rana Foroohar has told the story. Money talks, loudly.
I admit that I have never understood the Crypto Con. I am probably not alone. But I may understand that Sam Bankman Fried and his Effective Altruist (former) friends are simpletons who can and have done serious damage. At bottom, crypto is little more than “a shared database that allows millions of participants to do the work of tracking transactions, rather than using a single, centrally operated server…this distributed database of transaction is called the ‘blockchain.” Even a simpleton like me can see the utility of blockchain as defined. But how crypto can be a convenient and robust method of financial exchange is beyond my ken. What I can understand is that these cryptocurrencies are really nothing but faux wealth of an uncertain kind, digital tulips that will bever be useful as currency, no matter what people, including the current president of the United States, believe. [3] Crypto is an asset bubble that keeps inflating as long as “investors” believe someone else will pay more for the token tomorrow than they are willing to pay today.
How long will this bubble last? No one knows, but as Selinger and Cahn remind us, gold has been in a bubble for more than two thousand years of ups and downs, with more ups than downs in the recent goldbug past. And while “crypto enthusiasts may claim they trust in the math…their fortunes are resting on nothing more than the power of positive thinking.” [4]. As for the problems with crypto as a currency, mining is slow and energy intensive, payments are non-reversible and expensive. Sam Bankman Fried prattled on about how crypto would help the unbanked, but most of those who lack back accounts – except for the likes of Francesca Albanese, George Galloway, and Nigel Farage – don’t have enough money to open an account in the first place. Crypto has absolutely no value or place for them.
Home Security Upgrades are the equivalent of continuous glucose monitors used by wealthy HealthMaxxers who have no underlying metabolic condition such as diabetes. What Ring cameras mostly do is instill fear and reliance on a technology that cannot meet its hype. I have many suburban friends who have been told constantly they should be afraid in our very diverse community. [5] This only makes them buy more surveillance equipment.
The various failures of Covid Innovations have been covered here from many perspectives. We should remember that two things can be true at the same time: The Covid mRNA vaccines, which did prevent death and disease while not preventing virus transmission, were based on forty years of molecular biology preceding the pandemic. What was not done during the height of the pandemic or was “a sustained program to invest in the quality of our indoor air” which would have mitigated the spread of COVID-19 and also help prevent the next pandemic. Unfortunately, that is not how we think or work, seriously and incrementally. More’s the pity.
Naturally, some of the Generative AI chapter has been overtaken by events in what has often been referred to at Naked Capitalism as “an overly dynamic situation.” Yesterday’s Breaking Points discussion is illustrative. At the base level, AI is the non-human solution to the need for a dedicated personal assistant. Or as a married woman I worked with in the laboratory more than forty years ago (early 1980s, so consider the zeitgeist of Early Reagan) claimed in frustration, “What I need is a wife!” The larger problem with generative AI is that it necessarily entails cognitive surrender by those who come to rely on it for, well, cognition. Is AI a hyped innovation or a hopeful upgrade? For my colleagues and me, the AlphaFold is an unimaginable upgrade that enhances our cognitive function. For medical students and professionals, AI as presently conceived and delivered is mostly cognitive surrender that will lead to nothing good.
To go beyond Move Slow and Upgrade, Cory Doctorow’s conception of AI in The Reverse Centaur’s Guide to Life After AI seems most useful:
A centaur is a person whose work in assisted by technology, a worker whose tools make them happier and more productive. A reverse centaur, on the other hand, is a worker pressed into the service to technology, a person pushed beyond human endurance to work on a machine’s terms. Think of a warehouse worker made to work without food or bathroom breaks, a programmer made to crank out impossible amounts of code.
Or think of the person whose occupation is destroyed by the rise of AI. Unfortunately, as a path to the enshittification of human existence, AI as it is being intentionally developed converts people into reverse centaurs. Horses are noble creatures, yes, but a horse with human legs is useless. Unless and until the masters of AI reverse this trajectory, nothing good will come out of its wide adoption.
Finally, Cybersecurity is one of the few functional promised lands of the upgrader:
So many of the poorly designed innovations…in this book start with the simple assumption that a single change can solve a vexing social problem. Again and again, we see innovators overselling their inventions, claiming that a single app or algorithm can somehow fix deep problems rooted in entrenched inequity, outdated infrastructure, or any number of other multifaceted, interconnected systems. They move fast, they break things, but they rarely fix the problems. But cybersecurity professionals know just the opposite…They don’t just put a better lock on the door – they build a better door.
And they are honest, with an upgrader’s approach that would have been the better solution to each of the problems covered in this “little book.” In this they refract the light of good science. Very few scientific advances are leaps and those claimed to be otherwise are usually hype. Gene therapy was thoroughly hyped to the public in the 1980s and 1990s as molecular cloning became standard laboratory practice. Almost fifty years later, gene therapy is useful, but only for a relatively few diseases, and even with those the fix is not necessarily “simple.” The original gene therapy whiz kids moved fast, and they broke things. Few complex problems enmeshed in complex systems admit of simple solutions. It took a long time for the gene therapy project to recover from moving too fast, and even now moving fast breaks things and leads to bad outcomes.
The carpenter measures twice, cuts once. The good scientist does every control so she will know why the experiment did not work if it fails. Virtually all progress is incremental, just as most social problems can be solved only from an oblique approach. If the goal is to improve profits, the best route is to make a better product and treat your customers or clients with respect. Ido this and they will come. f the goal is to increase executive compensation, the stock price should be goosed in the short term using buybacks instead of upgrading the company. There are no shortcuts, even as we hear day after day that AI will transform society for the better.
And one final word of wisdom: Always upgrade your operating systems. The cybersecurity experts can’t keep you ahead of the bad guys if you do not let them.
Notes
[1] Analog controls were important in a car that could also get crushed in a similar accident. But I have no doubt that what I really wanted was a successor to the Opel GT of my late teens. I must confess that driving my quick little car with the 6-speed transmission does make me smile, as the last car I ever expect to buy. It also gets about 35 mpg overall and I drive only about two miles in a typical day.
[2] Which brings up the question, mythological or mythical? I would tend to go with the latter. Mythology can be useful human construct; mythical commonly just means “unreal.”
[3] Of course, the current occupant has used his crypto con to get much richer while his “investors” have gotten much poorer for buying in.
[4] The Trump family attended Norman Vincent Peale’s church when the president was a child.
[5] Property crime in my in-town neighborhood is much lower than in the suburbs, because we have people on the street at all hours, with the fire station around the corner and county, hospital, and university police officers nearby at all times. But more importantly, our sense of community is palpable. Neighbors actually know one another and pay attention.
By
