How do some great product ideas fail during tech hype cycles? A modern-day tempest retold from the trenches.
“Your tale, sir, would cure deafness” said no one ever, outside of fiction.
In this article we will see how the smartest folks in technology & business can fall for the promise of hype cycles, what happens when the music is playing, what happens when it stops, and what to do if you ever get caught in such a Tempest. Structure in 5 sections as follows —
Introduction (also see “Narcissism”)
Act I: The first winds of a hype cycle (with all the fizz & champagne)
Act II: The storm (and a worthy problem to solve)
Act III: The shipwreck (and the people behind it)
Conclusion
Introduction —
“Hell is empty and all the devils are here.”
I work in the field of security and my worldview leading up to this event had a very risk-based approach to it. My paranoia started in consulting where I had my fair share of long hours of particularly nothing and sometimes everything, “pls fix” pings, miles, sushi meals and hotel points. Then with an unpredictable event a realization hit me that I’d rather take real money and equity to build actual products as opposed to fixing decks for my clients. Please don’t get me wrong consulting accelerated my learning by exposing me to some of the smartest peers, thoughtful execs, and early experiences that a tech job would never have come close to offering all those in such a short span. On the flip side, I may not know where the skeletons hide but I have seen a fair share of blood including spot firings, lay-offs of managers and peers, tears at client sites (both of clients and colleagues) and all that goes on in a typical salesy environment. At an early dinner with a friend I remember uttering in a typical monologue “I have a love-hate relationship with consulting but I want to get into product management.”
With a few scars that accompany every consultant and after several hops of good and bad fortune, supported by well wishers (occasionally interjected by visa gods), I landed a role of a security product manager in a respectable tech firm in a bleeding edge product that intersected the Venn-diagram of all the envy-inducing buzzwords of the time — Responsible AI, Edge, IoT, Industrial Metaverse, GPUs, Security and Privacy. I was on cloud 9. I had finally arrived! (Actually, I hadn’t, yet.)
Act I: The first winds of a hype cycle —
“Misery acquaints a man with strange bedfellows.”
The COVID-induced demand pumped up the adrenaline in the tech world and projections of growth were everywhere. Reason almost always takes a backseat in euphoria. Blank checks were being signed for ambitious pet projects. Catered lunches were served. Promotions were handed out like candies. Nothing was hard-earned, everything was smooth flowing so long as you knew how to please key influencers. The chiefs were building empires and handing little fiefdoms to their favorites. To be clear, it happened everywhere in tech. People were dying in the pandemic but a career in tech was asymmetrically kind relative to other industries. More so, if you sat on the top of the food chain. Not being a tech employee was almost a sin in the pandemic. Why else would life be hell?
We all know the hype-cycle right? But the corporate version of a phenomenon similar to mansplaining is a(n ex-)consultant, showing a line that a toddler can doodle while half-asleep, so here we go!
Act II: The storm (and all its wrath in product lifecycles) —
“Now I will believe there are unicorns.”
Who wouldn’t want to explore in good times and exploit in bad ones? We can’t simply blame the tech leaders all the time. With top-class graphics for slideware, an adequately hyped-up brand for this new tech, accompanied by a fancy UI, and preview products that met 2–3 use cases exceptionally well in lab-controlled environments but failed most practical real-world applications, I arrived at the scene at the peak of inflated expectations when management was shipping dev kits for free on one hand and was talking about killing a few lines of product on another.
I don’t know what most of your weekends are about but as a millennial my socializing has come down to either gender reveals or single-digit-year birthday parties. There is a point here I shall get to soon but on one such weekend evening I found myself blindfolded, (which I am fine with except that I was) sticking a virtual pacifier on a virtual baby before me. Evidence below.
These exact same under-slept, well-meaning, good-natured, smart individuals, who are making more of themselves, get up on Mondays and make decisions for giant corporations in the hope that they will change the world for the better, for their next of kin.
Without the blindfold, if you handed a marker and asked any of my hundred incredibly smart team members building the industrial metaverse, “Where do we want to focus this quarter?” or “Where were we focussing on in the last quarter?”, the image below would look very similar to the one above. Except with a lot more shots missing the mark. Unfortunately, there are no pacifiers built for adults, but as we shall soon see a lot of crying followed as the imaginary blindfolds of surplus cash were pulled off.
The Problem Statement: Building an AI model marketplace is a very interesting and difficult problem to solve. Although all tech firms are trying to build the next marketplace for AI models, and some have claimed to have already done so there is no everything store for AI models, YET, where researchers can come and put their intellectual property up for sale; A secured free-market where the IP gets readily consumed across various use cases, industries, geographies and hardware combinations. There are several traditional challenges such as identifying key target industries, top use cases across them, hardware compatibility, getting early adopters, building a flywheel of network effects, and identifying a viable path to monetization to getting your first paying customers. There are also new challenges since AI is a blackbox and copyrights/IP laws do not apply as model parameters are essentially non-static ephemeral entities that invite a lot of litigation for infringements. Also, data residency, data ownership, and sovereignty are debates for very long tables around the globe.
Put simply, you cant really go to one website and shop for an AI model you can download on your phone to count the number of sheep jumping off a fence although there are several dozens claiming that they do it, you will quickly find yourself buying other software and hardware you did not need or ask for. All you want is an AI to count the sheep jumping over the fence so you can sleep in peace!
While ideas are infinite, resources are not. Poor judgement catches up to every organization that indulges in lazy decision making and every person in the orbit suffers. No amount of slideware can substitute for a lack of strategy and planning. If x is the product, in our case it was indeterminate. Security is all about the product, the data within it, and the users it interacts with. It is essentially a function of x. f(x) for the geeky. I was trying to solve for f(x) when x kept changing values depending on who you asked, the day, or the time of day. Of course, I was struggling.
Act III: The shipwreck (and the people behind them) —
“Good wombs have borne bad sons.”
This brings me back to where we began, the paranoia that ghosts every (current and former) consultant. Consulting, put bluntly, is selling services largely in the form of human hours. Every consultant knows that while impact can be fictional up to some layers of middle management, money is very very real in the top rungs. Yual Harari can get all nuanced about this and disagree however, every consultant’s career path is clearly laid out. Earlier in your career, the more billable you are, the more successful you are. As years go on, the more billable hours you can sell, the farther you go. It’s up or out. Raised in this cut-throat environment, I had started raising alarms and even ruffling feathers expressing my concern over the longevity of the product and by extension, our careers. This was before the mass layoffs were announced. Tech had not seen lay-offs for a while, for a decade stocks had gone only up, up and up. The memory of the turn of the millennium was wiped off in champagne and covered under hoodies and goodies. Everyone brushed me off when I expressed concern. The key learning here is the Persian messenger syndrome is very very real. Maybe I shall learn someday.
The future became clear gradually at first, then all at once. The seasoned A-players started leaving. The food and champagne were long gone and hiring budgets started freezing as if all CEOs and CFOs met and agreed unanimously at the last summer potluck for the year after the reopening. Consultants hid the phrase they had invented just a year back. Remember “new normal”?
One of the popular girls grabbed a cavalier said in what I imagine was a valley accent “You know what be really cool?” “LAYING OFF WORKERS!!” they screamed in unison and started throwing pillows in the middle of the potluck. Why not? The CEOs, some of them brown folks like me, stood at the corner and snickered as their mouths watered and dollar signs began popping in their eyes. Consultants proposed another phrase as an expensive bottle of champagne popped, “Year of efficiency? Sound good? Let me text my team this fine Sunday afternoon and you shall have something tomorrow morning before you are done with your infrared sauna!”
Back in the land of minions, middle management, where the most insecure bunch of humanity resides, got a whiff of the rumor and started to agree on the concept of quiet firing, marking targets in close collaboration but continued to disagree on almost everything else. They also preferred their KoolAids in different flavors frequently.
The HR function, except for the poor ones in recruiting, suddenly found themselves rediscovering their sense of worth. Budgets were cut, and mental-health-related monologues by leaders were replaced by gratitude-related ones. Those who couldn’t quit started quiet quitting.
Gen Z dared to do what millennials only thought of and actually pointed out the Boomer hypocrisy. Or in their elegant words, “it all looked sus”. I don’t know how it was in the past millennium, but Gen X probably had quite an identity crisis as Boomer leaders decided not to retire before the casket after surviving the pandemic. Frankly, going by the media, Gen X was never a trending hashtag. No one particularly thinks of them, including themselves.
Satire aside, very competent, smart, hardworking individuals were displaced. Products with great potential were killed.
Conclusion
“What’s past is prologue”
Big tech is probably where people trade their dreams for golden handcuffs. Again, I wear these and I wear these proudly on most days.
But if every rational thought counts and if execs are paid for the quality of their decisions, was it wise to spend so much on hiring during the pandemic? Was it wise then to do the lay-offs to show QoQ growth on EPS? How were voices of reason muffled in the layers of bureaucracy? Does the pendulum have to perpetually swing between fear and greed? What happens when another wave of demand for labor creates another bid war talent?
What we seem to learn from history is that we learn nothing at all. Anyways GenAI is here. Everyone agrees it is not a hype. This time it is different! Bring out the champagne! Where are the recruiters at? All gone? How? Never mind, let’s pull an offsite together in Vegas or Davos and brainstorm this. The consultants can build the decks off our wine-stained napkin drawings.
More from Urban Malgudi here. You may also like some of these popular posts related to business & investing:
Data is NOT the next oil; Sand is!
