Patrick McKendry

It would be an understatement to say that AI is a hot topic right now. Since the release of ChatGPT at the end of November, its functionality and the wave of new use cases has been a near constant conversation in the business world. That all but accelerated with the release of GPT-4 to OpenAI’s chat products and language model APIs in March of this year.¹ Within hours, users leveraged the tool to create a website from a napkin sketch, recreate the classic app-based game Pong, and much more.² With the achievements of OpenAI’s GPT progressions, it is easy to see why Microsoft has invested in such a big way in the company.³ Not to be outdone, and to protect its search engine market share, Google has accelerated accessibility of its own AI-enabled chatbot, Bard.⁴ Facebook recently released LLaMa - its own large language model - to support NLP researchers, adding more fuel to the tech giant race.⁵ All to say that the evolutions happening as a result of AI enhancements are starting to come fast.

‍

Maybe too fast?

‍

In late March, as major tech companies locked themselves into the accelerating arms race of language AI superiority, tech leaders from around the world signed an open letter urging AI developers to pause further development on these models.⁶ The letter cites concerns of programmed bias in models, spread of misinformation, data security, and an unclear impact on human labor impacts as use cases grow.⁷ ChatGPT has also been banned in Italy due to its mass collection of data for model training purposes and potential non-compliance with GDPR for how it captures and stores personal data.⁸ Members of the U.S. Congress have also started to call for laws regulating the use of AI technology as the U.K. government published its own regulation recommendations last week.⁷

From the outside looking in, the AI race is an interesting case study in organizational momentum. For leaders, momentum can be a powerful tool to have your side. However, momentum should be approached thoughtfully as it can always risk snowballing out of control. The rapid surge of AI development in the past half year showcases what this can look like, both for good and bad. Here, we will talk about what momentum is and how to create it, the difficulty of overcoming inertia, and what happens when momentum grows out of control.

‍

Momentum - What is it?

Do you have momentum?

It likely needs no explanation. The concept of momentum is prevalent in our society. We talk about it in sports. We hear about the business cliche of “tailwinds” and “headwinds”. It is that common feeling that everything is going right or falling in our favor. As a leader, momentum is helpful because it can absolve a lot of sins. Proposals that are not quite complete get sign-off because initiatives for the team have tended to be successful. Project scopes that are not quite defined tend to be worked through rather than causing a full stop and re-scope effort. Missed moments of praise or imperfect global corporate memos never grow beyond “mole hill” status because good intent is asumed. Usually, you can just tell if you have momentum. And we certainly feel when we do not have it. Everything becomes a fight where you need to prove beyond doubt that things are under control and tests are likely to work out.

Alternatively, you can attempt to measure this more quantitatively with a team pulse survey. In this case, you want to measure for the factors going into momentum. Ask questions to get at these inputs:

On a scale of 1-10, how confident are you today in our ability to achieve our vision and goals for the upcoming quarter?
In your own words, describe our team’s vision.
On a scale of 1-10, how much do you agree with the statement “Our meetings and production work all add unique value.”

You will want to get a check on momentum fairly frequently. I would recommend a pulse survey quarterly or maybe monthly. For leaders, it should be top of mind as a question for weekly reflection.

User interactions with OpenAI highlight what momentum can look like for a business. OpenAI has been around for a few years, but many people only heard of them recently. Founded in 2015, the company released its original GPT model in 2018 alongside of a myriad of other machine learning models research across the years.¹¹ Their explosion onto the popular culture scene was a masterclass in creating momentum. A few months after releasing GPT-3.5 to its API products, OpenAI built and launched the PR behemoth ChatGPT. What followed was a flywheel of word-of-mouth growth leading to consumer-delivered product-market fit creating effortless public relations headlines. The service took off and became the fastest application to hit 100 million active users, reaching the mark in one-fifth of the time it took for TikTok.¹²

‍

Inertia - A lesson in physics and jumpstarting human behavior

Building Momentum

If we want to talk about building momentum, we really need to talk about overcoming inertia. I believe the reason most leaders feel unable to create momentum is that most leadership training is not focused on changing the mental state of teams. In those cases, some velocity is already assumed. But that is a huge situational difference because when we talk about creating momentum, the force output of commonly recommended actions is far less than what is typically expected. This causes many leaders to get discouraged when their typical tactics seem to not resonate and they stop short of ever overcoming the initial state of inertia.

Newton’s first law of momentum teaches us about inertia, stating that objects at rest will stay at rest while objects in motion will stay in motion. The math is not directly applicable, but the principle seems to apply fairly naturally to human behavior. Psychologists often refer to humans as “cognitive misers” to describe the behavioral tendency to avoid mental exertion and follow the simpler, existing path whenever possible.⁹ Think of it this way. Imagine you came upon a large boulder that was sitting still. If you give it one push, you have created momentum. But you may not notice your efforts because the amount of force you can create is much less than the mass of the boulder - especially with just one push. What happens if you push twice? You’ll apply twice as much force, and still the impact may be unnoticeable. There is momentum that is ever so slight but it is there!

Psychologists often point to the benefit of framing a situation as an emergency or critical point of failure to overcome inertia and make others believe that change is needed.¹⁰ This can work, but it is not reasonable to have this tactic as our only tool. We can also use positive feedback and energy to build momentum. This is good because this tool can be leveraged time and again without wearing out. Still, when applying this to our teams it needs to be an all-out effort. A few kudos and the occasional celebration will not cut it. Think again about the boulder. You cannot just push once or a few times. But, if you push consistently with all of your might, maybe - with a little bit of a slope to help - you can get it moving. To start from zero, any positive win should be called out and tied to the team charter and the new goals. This will highlight to the team that there are many ways for us to make progress and generate wins, even in little ways. If you haven’t had any momentum then the wins will be especially sweet and members across the team will work to be able to create their own. The variety of rewards will also be increasingly important here to offset the diminishing effects of high frequency praise. Additionally, some rewards may also help you to apply greater force of momentum - recognition from the CEO likely holds more weight than an appreciative email from you (no offense). And note, the excitement and praise must be genuine. That should not be hard as, frankly, it should be fun to start building momentum. The fun and positive energy that you bring to the team will be contagious and the group will follow your lead. Stick with it and you’ll, eventually, be on your way.

There are, of course, limitations given the strength of the headwinds we might be facing. Yet, just as a skilled sailor can tack into the wind, a good manager can set relevant goals that can allow for progress against a challenging environment. And if progress can be made, then momentum can be built.

In the AI race, Google gives us a good example of what inertia looks like. Google is a powerhouse of machine learning capabilities. They created the transformer technology upon which OpenAI’s GPT model is built after all.¹³ However, today OpenAI and its primary investor, Microsoft, are taking the lion’s share of AI and NLP headlines as Google’s Bard chatbot generally finds itself on the receiving end of critiques. Google appears to be getting lapped in this race because it was slow to start and has been slow to move. The inertia for Google is that their massive advertising business was not at risk and continued to be the main driver of the business. There was no emergency to drive change and, I expect, little positive reinforcement at the organizational level to redirect team efforts to these new capabilities. The major turning point in the conversation was Microsoft’s entrance to the game and their promise to bring ChatGPT into the Bing search engine interface.

Snowballing - When momentum grows out of control

Controlling momentum

Creating momentum in your team helps foster a winning environment. We talked about the markers of momentum at the start - agreement for tests by default, greater appetite for risk, and things tending to fall in your favor. These are fun tailwinds and it is clear to see how we can benefit with some amount of these momentum behind us. Yet, too much momentum is not good. As leaders, our role is to maintain the appropriate balance. This may mean opening the throttle further than we had in the past as our team gets more capable. But it also requires that we introduce counter-forces to slow momentum when it risks getting out of control. We need to be considerate of the short and long term tradeoffs. Keep momentum on your side, but in check.

There are numerous tactics to control momentum and your options depend on how close you are to the brink. The primary goal is keep momentum above zero, and if you are suddenly very close this could fairly abruptly cutting high-risk projects with large potential negatives if they do not land. Shame on you if this is required. Ideally, there are other approaches. This might look like resetting test criteria and challenging the team to meet more rigorous standards for ROI prior to launching. This could mean smaller investments of people or money. It could also mean healthier margins on revenue generating activities. You could push for lower risk profiles to increase the likelihood of net positive payouts on testing. All of these approaches require raising the bar for your team, and you must be prepared to coach them to that delivery. If your team is not ready for the rigorous requirements delivered by counter-cyclical environments, shift their focus back towards more core production activities. This allows you to maintain momentum and continue to pile up wins. Yet, I will note that you will likely need to play a more active role as a momentum builder as the work will become inherently less exciting and the team, resultantly, less enthusiastic to celebrate that work.

The largest threat to momentum is almost always something external - economic factors that slow down your industry, a competitor taking a new tactical approach, or government regulations as example. These types of changes are unpreventable, but you can prepare. When monitoring momentum, survey the horizon for problems looming three to six months out. If you can see the storm coming, it is smart to navigate around it. This does not need to be a complete overhaul of strategy and there is not necessarily a need to warn the team and throw off their focus. You can, however, start to shift what behaviors you reward and adjust Big Rocks to play defense. Sometimes a change will catch you completely off guard. Meaningful macro changes should translate to resetting your primary priorities. It is worth not waiting for this. And, if you need to edit multiple times as broader factors continue shifting, be comfortable with making the edits. The alternative is to do work that does not account for the new environment. The risk you run here is that the team’s work will seem pointless and lacking direction. Without direction, we don’t have velocity. And without velocity, our momentum will start to slow as the hits from macro factors slow us down bit by bit.

The sudden loss of momentum is one end of the range of outcomes we need to try and control. The other is runaway momentum. Imagine that boulder again and you reach the top of a hill. As it rolls down the hill, momentum picks up and will grow out of control and unstoppable. Most likely, that boulder crashes into a number of objects, creating and taking a lot of damage before coming to an abrupt stop.

For OpenAI, we see a case of momentum clearly run out of control. Sam Altman has put up feeble attempts to slow the ChatGPT flywheel, citing weaknesses and simple errors that the tool makes.¹⁴ Altman has even cited concerns for needed regulation around AI, measures to reduce the spread of misinformation, and improved research for preventing cyber attacks.¹⁵ At the same time, OpenAI continues to update its language model products with GPT-4 arriving less than a year after version 3.5 and GPT-5 planned for launch by the end of 2023.¹⁶ It is this rapid push forward that has led government regulators and many in the broader AI community to push back directly against OpenAI and similar research labs, rather than seeking a more collaborative approach on a path forward. It remains to be seen exactly how strict and limiting regulations will be and the size of potential fines for privacy violations. This is the cautionary tale of heedless momentum.

‍

Momentum is the organization’s Sword of Damocles. From those on the outside, momentum is an incredible tailwind to have at your back. The benefits provided are numerous and constant. Yet for leaders, momentum can never fully be enjoyed. The potential for sudden disruption or out-of-control growth hangs over our heads, only a hair’s width away from catastrophe. Momentum is literally a force multiplier. Having it behind you is game-changing. But remember, everything should be enjoyed in moderation.

‍

References

‍

Momentum - Good or Bad? It Depends.