Strong Opinions Loosely Held Might be the Worst Idea in Tech

Nobody does it better

I misspent just about every night for 7 years at an old school pool hall in Milwaukee. This was an “action” room, where a regular lineup of shady characters was perched, ready to relieve you of any excess cash that might be burdening you. Most of the betting was on pool, but it wasn’t unusual to see wagers on cards, dice, sports, weather, or the time it would take a cup of water to boil in the snack bar microwave. Basically, if two people could be found with opposing viewpoints and an American dollar, it was on.

Betting on pool is a highly calibrated affair that makes it at least theoretically possible for a rank amateur to arrange a fair match against Efren Reyes. For example, in the game of 9-ball, you can adjust the number of games each player needs to win a set, which balls win the game for each of them, who gets to break, etc.

Lenny’s Wisconsin Billiards, c. 1995

When you bet on games of skill, you learn that the expected value of your bet depends as much on how accurately you assess your chances, and thus the handicap you negotiate, as it does on the skill itself. Gamblers that hold an inflated view of their chances (i.e., almost all gamblers) will lose money eventually, and the feedback loop is generally short. Overconfidence is folly and will be ruthlessly exploited. If I think I can beat Efren with the 5-ball and the breaks, I and my wallet will be corrected rapidly.

Toxic Certainty Syndrome – Could You Be Its Next Victim?

And yet, in the tech industry, with our motto of “strong opinions, loosely held” (also known as “strong opinions, weakly held”), we’ve glorified overconfidence. The same foolishness that would cost your money and pride in a pool hall is the stuff of legend in Silicon Valley. Stroll through an engineering office and you are likely to hear the (mostly white, mostly male) denizens making statements like:

“Only an idiot would use MongoDB.”

“We should absolutely build this using GoLang.”

“All of our users would kill for that feature.”

“Linguini’s is the worst restaurant in the world and I’d rather eat dehydrated bat vomit for the rest of my life than have lunch there again.”

The idea of strong opinions, loosely held is that you can make bombastic statements, and everyone should implicitly assume that you’ll happily change your mind in a heartbeat if new data suggests you are wrong. It is supposed to lead to a collegial, competitive environment in which ideas get a vigorous defense, the best of them survive, and no-one gets their feelings hurt in the process.

On a certain kind of team, where everyone shares that ethos, and there is very little power differential, this can work well. I’ve had the pleasure of working on teams like that, and it is all kinds of fun. When you have a handful of solid engineers that understand each other, and all of them feel free to say “you are wrong about X, that is absolutely insane, and I question your entire family structure if you believe that, clearly Y is the way to go”, and then you all happily grab lunch together (at Linguini’s), that’s a great feeling of camaraderie.

Unfortunately, that ideal is seldom achieved.

What really happens? The loudest, most bombastic engineer states their case with certainty, and that shuts down discussion. Other people either assume the loudmouth knows best, or don’t want to stick out their neck and risk criticism and shame. This is especially true if the loudmouth is senior, or there is any other power differential.

The loudest, most bombastic engineer states their case with certainty, and that shuts down discussion. Click To Tweet

Members of your team that are from groups historically underrepresented in tech may be less likely to have experienced the collegial, open debate environment, and may feel uncertain of their position. This means you might not hear their ideas. Given the extensive research that shows diverse teams make smarter decisions, this is tragic.

Even if someone does have the courage to push back, in practice the original speaker isn’t likely to be holding their opinion as loosely as they think. Having stated their case, they are anchored to it and will look for evidence that confirms it and reject anything contradictory. It is a natural tendency to want to win the argument and be the smartest person in the room.

Finally, in most cases, the feedback loop on these decisions won’t be closed quickly. Unlike my foolish 9-ball matches, a major engineering choice might not bear full fruit for months or years, by which time we’ll have forgotten how we made the choice and what the alternatives were. There is great value in thoroughly evaluating the key alternatives up front.

This (Actually) Won’t Hurt A Bit

Fortunately, there is a remarkably simple solution to the problem, which I picked up from Dan Shapiro, our CEO at Glowforge. He tells me he inherited it from our company’s co-founder, Tony Wright. The behavior change required is very small. You can easily model it for your team by adopting it yourself, and you can easily and gently guide other people into making it a habit. It isn’t even deceptively simple, it is just … simple.

All you need to do is add a degree of uncertainty to your statements. For example:

“I’m 90% sure we shouldn’t try to build our own social network.”

“I’m 50/50 on whether to do this with Cloud SQL or Cloud Datastore.”

“I have a low conviction hunch that the airplane icon will work better than the gift box.”

The effect of adding this qualification is twofold. First, you’ll keep your own mind much more open. As world-class poker player Annie Duke points out in Thinking in Bets, even if you start at 90%, your ego will have a much easier time with a reversal than if you have committed to absolute, eternal certainty. You can safely update your beliefs and say “Hmm, you make a good point. Now I’m only 60% sure.”

Even photons aren’t sure of themselves

Second, the folks around you will now feel invited to participate. When the opinion that you express has explicit room to evolve, they know there is going to be a respectful space to share their own ideas. If you are a leader, this is also a useful management tool, because it helps to calibrate your team as to how much effort they should put into alternatives. If you are already 90% sure, there is less reason to invest in counterarguments than if you are simply throwing a guess into the pot.

So, what about the situation where someone else goes first and makes an absolute statement? There is a simple ninja move! Just say “It sounds like you are 100% sure of that, is that right?” If the answer is yes, you can ask them to explain why they are certain and see if they have any data to back it up. If not, you’ll have prompted them to assess their actual level of conviction, sharpen their thinking, and open up the conversation. It is a simple, kind way of helping them develop a style of thinking and communication that will improve your organization.

There is a simple ninja move! Just say “It sounds like you are 100% sure of that, is that right?” Click To Tweet

One last point. If you have been an engineer for a while and are accustomed to the strong opinions, loosely held mindset, it can be easy and even more disastrous to bring it home with you. I may have had a significant other or three tell me that my certainty was obnoxious. Honestly, I thought I was being funny, and I took for granted that anyone else would feel free to put forward their own opinions. But hey, it has only taken me 30 years or so to get the message and start to work on it! I’m 80% sure it is helping.

Michael Natkin is Glowforge’s VP of Software Engineering. We are hiring for software and many other roles. We also happen to make an amazing 3D laser printer.

13 thoughts on “Strong Opinions Loosely Held Might be the Worst Idea in Tech

  1. Love your article, I think the principle behind Strong Options, Loosely Held is great, but you’re right if it’s just an idea and not at the core of every person on a team/org the strongest opinions will trample over those that are “Loosely Held”.

    It sounds like a good addendum to the mantra Strong Opinions, Loosely Held should be:
    If your opinions are at 100% (or over 90% most of the time) you’re *not* holding them loosely.

  2. I have never heard the phrase until I saw Dan’s retweet but this article makes so much sense.

    Our local volunteer makerspace has had a toxic-feeling atmosphere for the whole time it has existed, whereas the more expensive for-profit makerspace is just lovely. I’ve often wondered if it was because the former was founded by hackers and the latter by artists.

    It makes sense that the “I will say this loudly and confidently while calling your ideas crap” aspect of nerd culture has contributed.

  3. The course of action you recommend is exactly how the concept of “strong opinions, weakly held” was originally intended, at least as I have understood it. The behaviors you ascribe to people you’ve experienced who claim to hold strong opinions, loosely held sure don’t sound like the actions of someone open to new information.

    1. As I was reading the article I thought the same thing, but I believe that his point is that even if it’s possible for some people to use “strong opinions, weakly held” responsibly, that it’s not clear enough of a meme to be used sustainably across an industry.

  4. Unfortunately, your title is clickbait. What those engineers are practicing is Strong Opinions Strongly Held. If anything, you prove that Strong Opinions Loosley Held is what engineers ought to be doing and suggest a mechanism to allow for that to happen given many people lack the EQ for the practice.

  5. This requires a certain level of collective maturity to work as intended. It’s intended to produce discussions that improve the ideas being discussed. It requires that everyone understand the sole reason for partaking in the discussion is improving the idea and that everyone maintain intellectual humility.

    The second part is oftentimes problematic. I find that framing the situation as a quest for more data before making a decision is helpful. No human knows absolutely all of the data potentially relevant to a situation. It is emotionally much easier to be humble to this fact than to another person.

  6. In an industry that in many places explicitly allows and even encourages age discrimination, which is illegal, not to mention gender discrimination and other manifestations of bigoted mindsets that are founded on toxic certainty, how will people ever be mature enough to overcome toxic certainty? It is at the core of the industry.

  7. I concur and agree with the basic premise, that a strongly held opinion loosely held is toxic. I have experienced a team lead, supervisor, or manager doing just that, and a meeting is simply everyone giving a rubber stamp to concur. But the strongly held opinion is the conclusion, any facts in counter-argument only support the conclusion–the facts fit the conclusion.

    However I disagree with the suggested solution of giving percentages to buttress that opinion. The problem is that a percentage or even a probability gives the veneer of factual, but is another opinion. An opinion needs the logical reasoning, and for a percentage–the set of criterion used for the basis of the opinion.

    The logic is explained by quantifying a statement, and giving the rationale. From the article I’d ask:

    “Only an idiot would use MongoDB.”

    Quantify what an ‘idiot’ as a user is, and how this user will ‘use’ MongoDB. Also in contrast, what other non-SQL databases are less ‘idiot’ for users? Or is the ‘idiot’ using non-SQL, and non ‘idiot’ using SQL?

    “We should absolutely build this using GoLang.”

    Quantify what about GoLang makes it an absolute to use, and what other languages in contrast?

    Seven of Nine from Voyager said it best: “State your reasons for this decision.”

    One approach I have used is to list advantages and disadvantages, and take the simple ratio of advantages over the sum of both advantages and disadvantages; or list the criterion with a simple true of false. Then if for ten criterion, seven are true, I can then demonstrate 70% or 0.7 probable.

    As Homer Simpson said, “Statistics can prove anything, 80% of people know this.”

    For this comment I see the simple probability of 50% agreement, the law of averages, or more simply the choice to agree or disagree; thus one possibility out of two.

    There are other factors to consider such as the Peter Principle, groupthink, cognitive dissonance, etc. But unless an organizational culture avoids domineering, blustering, bombastic individuals that are making decisions, percentages and probabilities will only make the problem more toxic…more information that without explanation, quantification, or a rationale basis only makes toxic more toxically obfuscated.

  8. Love the article, and wish the expression was “thoughtful opinions, loosely held”. Design decisions in team environments are often dominated by the strongest/loudest personalities, and exacerbated by power differentials (as Michael explains well). Adding the uncertainty factor should certainly help the team be more thoughtful and less prone to “toxic certainty”. @Tad, the difference is that Saffo’s article is about individual thinking, not team dynamics. Translating to teams requires some allowance for those dynamics. Has given me some food for thought in how I can better help steer teams to avoid toxic decision making.

  9. Strongly held opinions are not always held just by engineers, sometimes it is also held by an individual who maybe a superior and may feel they know it all and will not allow the engineer’s experience & knowledge play a part in the decision making. The superior uses his muscle power & uses his tittle to impose their idea although it can be determent-al . how do you address egos or insecurity of leadership?? Thanks

Comments are closed.