1. 37
  1. 16

    “…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…”

    No, it does not. It means that as a professional, you do not associate your passion for a technical topic with your self-worth. It’s okay to feel strongly about something without being a jerk about it. It’s permission to speak honestly and get feedback from others. If you do this, others may very well call you an asshole! They may vote you off the island. So you learn what’s important and what’s not. What’s worth having public passion over and what’s just being a jerk.

    You should have strong opinions about error handling, or customer service. You should speak up about your opinions so we might learn from you. You might have strong opinions about the pasta down at Luigis, but nobody cares. And we’ll tell you so. The requirement with strong opinions lightly held is that you have to agree to suck it up and do things you might not agree with. You can’t have one without the other.

    It’s just the opposite of what this author thinks it is. It’s the thing that over time prevents people from being jerks, not encourages them to.

    ADD: Does our industry have a problem with tech bros and people making bombastic statements? Most definitely. But I don’t think that’s related to SOLH. In fact, for every one of these bozos, I can show you five tech people that should be speaking up and are not. Those are the ones who cause real damage to an organization.

    1. 21

      You should have strong opinions about error handling, or customer service. You should speak up about your opinions so we might learn from you.

      There is a famous Bertrand Russell quote: “The whole problem with the world is that fools and fanatics are always so certain of themselves, and wise people so full of doubts.”

      It matches my experience exactly. The more I learn about complex topics like error handling or customer service, the more I realize that there is no “One True Way”™, and that it’s actually really complex and a series of trade-offs.

      Does our industry have a problem with tech bros and people making bombastic statements? Most definitely. But I don’t think that’s related to SOLH. In fact, for every one of these bozos, I can show you five tech people that should be speaking up and are not. Those are the ones who cause real damage to an organization.

      At least some people are not speaking up out of fear or exasperation by being told that their viewpoints are “absurd” or “not sane”.

      I just tap out of discussions these days once someone starts acting like that, and I know that in a few instances it has caused real damage to the business. I’m not happy with that, but there is a limit to what I’m able to put up with before it starts affecting my mood to an unreasonable degree.

      1. 7

        Strong opinions does not necessarily mean obnoxious or loud. I don’t understand how we got to the point where people think that. I feel strongly about fishing in the rain. Doesn’t mean I ever yell about it. It’s an odd conflation of ideas, as if the demonstration is as important as the strength. I wouldn’t think that would be true at all.

        We have to self-correct. In order to self-correct, it is necessary to make a case for some path we recommend and then negotiate/argue/arm-wrestle as part of a decision-making process. Passion allows us to make our case. It does not have to involve yelling or being rude. We just have to care, to feel strongly. After all, we’re professionals. Why wouldn’t we feel strongly about various parts of our job? I’d argue we aren’t worth a bucket of warm spit if we don’t. I know I wouldn’t want to work with anybody who had no passion in our work.

        It’s such a weird confusion of ideas. Feeling strongly about something does not mean acting like a bozo. In fact, that’s pretty much acting childish.

        Yep, everything is complex and people who oversimplify and are too sure of themselves can be problems. It is also true that given incomplete and sometimes self-contradictory information, we are required to make choices. We should do what we can to make sure these are the best choices possible. Tapping out ain’t cutting it.

        1. 5

          What you’re describing doesn’t sound like a “strong opinion” to me, but rather just “an opinion”.

          Perhaps this is just a case of semantics, but the adjective “strong”, to me at least, means either “not likely to be convinced otherwise” or “obnoxious or loud” These are indeed two very different things, but generally neither of them are very constructive, and doesn’t really seem like what you’re describing.

          1. 1

            I thought the strong was an adjective on your belief in the opinion. As in, you believe with all your heart that a particular option/action/way is the right one.

            1. 1

              If you’ve got extremely strongly-held priors, you’re liable to be hard to convince (and require a lot of evidence to budge). This is a problem because such priors are not typically based on reality. As a general guideline, if you want to be more right about things, you ought to be less sure about them.

        2. 2

          I agree. I am not going to have a shouting match with my coworkers. It’s not worth the aggravation.

          1. 1

            There is a famous Bertrand Russell quote: “The whole problem with the world is that fools and fanatics are always so certain of themselves, and wise people so full of doubts.”

            Well, I am personally full of doubts. Still I would give some of this type of decision makers more credit. It is easy to criticize everything since nobody knows anything of relevance with certainty (or you wouldn’t need to argue about it). In a tech context (and many others), it is often more effective to go with one reasonable choice, stick with it and control doubt with overconfidence.

            1. 1

              I like the quote and agree on the spirit ( https://quoteinvestigator.com/2015/03/04/self-doubt/ ) but you are confusing strong opinions with overconfidence / inflated egos.

            2. 2

              There’s something valuable about SOLH, so long as the LH part is taken to mean both qualifying statements based on an honest and informed estimate of confidence & actually updating your priors. And, some folks who stand by SOLH (including, presumably, some of its popularizers) take it this way. I’ve seen too many folks who apply it in the way OP describes to believe that it isn’t being used as a justification for ultimately destructive behaviors, though, even if that application is based on a misunderstanding of what the original popularizers intended.

              OP’s suggestion of annotating statements with confidence levels (which is popular in the rationalist community for blog posts, & seems to have come from Robert Anton Wilson, who recommended it along with e-prime for avoiding common patterns of miscommunication) is a good one, because it rewards accurate estimates of confidence, providing a road for careful folks to gain social status over pundits & blowhards by raising a useful metric above being loud and contrarian (which, unless it’s paired with careful thought, introspection, and a rigorous and accurate estimate of one’s own confidence levels, usually ends up being equivalent to being annoying and wrong).

              Of course, this runs contrary to norms. We live in an environment where qualifiers are called ‘weasel words’ & no matter how much you signal your level of confidence, all those signals will be stripped away as you are judged on your conclusions as though your confidence were 100%. Furthermore, confidence is held in esteem before it can even be proven to be justified, so we cheer on demagoges for being bold as they lead us full speed ahead into obvious traps. In such an environment, people who can get away with avoiding being held to account are incentivized to sound very sure, and everybody else is incentivized to keep their mouths shut.

              1. 3

                I think all of us are missing the point when it comes to passion, confidence, truth, and so on.

                These are language games. Most of what teams do are language games. Put another way, everybody wants to do a “good job”. The rest of what we do is trying to come to common agreement on what the phrase “good job” means.

                The certainty number isn’t awful. It just misses the point of what we’re trying to do. As arp242 pointed out, things are complex. What we’re looking for is the simplest question we can agree on that’s important, testable, and that we disagree on the answer. That question might be something like “Switching to SONAR will result in 10% fewer bugs to production” (I can’t give you a good example because it varies widely depending on the circumstances.)

                To get to that pivot question, we have to take strong opinions about fuzzy things, then work our way towards being more reasonable about absolute things. This is the job of being a smart person who creates technology. A user comes in and says “I hate this! Make it stop sucking so badly!” and we work towards testable chunks of code.

                It’s perfectly fine to respond with “What do you mean this sucks? This is awesome!” This is the beginning of that back-and-forth. Checking out is not an option. You could try to go the percentage route but then you’re not working towards better definitions of terms. Instead taking a strong position and then following it up with something like “Which parts are sucky?” takes the game forward a step.

                “I like X!” vs. “I hate X!” are fine places to start. There’s passion there. Now add some technique and flexibility. If everybody is appropriately apathetic, you are in stasis. Not a good place.

                1. 3

                  That makes sense when the origin of the decision-making process is passion, and when nobody comes to the situation with a nuanced understanding. In most professional situations, neither of these are the case: developers are working on things they don’t care about for users who see the application as a necessary evil, and one or two folks in the group have 30 years of professional experience to everybody else’s three months (along with complex nuanced and experience-backed takes that simply can’t be boiled down to a slogan). Three junior devs shouting irrelevant preferences doesn’t help in this situation, and because their nuance-free takes are low-information, they can be repeated over and over (and thus gain control over the thinking of everybody else). The person with the best chance of designing a usable system gets shut out of the discussion, because when takes are optimized for hotness nobody wants to read an essay.

                  This notional experienced dev has a greater justification for confidence in their general position, but will necessarily express lower confidence in any individual element, because they have experienced instances that provide doubt. Meanwhile, the junior devs will be more confident because they have never been contradicted. This confidence is not representative of justified confidence.

            3. 11

              I never quite understood the “loosely held” part of this. As in, I genuinely didn’t understand what was intended with it. I never bothered to look it up so I always just abbreviated it to “strong opinions” in my mind.

              This:

              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.

              Made me laugh out loud because it’s such a contradiction. What it really means is “I have strong opinions, but also recognize I may not have all the information, so I’m willing to change my opinion”. In essence, it’s just admitting ignorance.

              Ignorance is not a bad word; everyone is ignorant of a lot of things; I sure as hell are. But this also means I’m more humble in dishing out my “wisdoms”. I think the habit of stating things such as “I’m 90% sure that …” is a great tricks, and something I’ve been doing a lot in the last few years (although not always equally effectively, but that’s another story).


              I find that a lot of the times these kind of discussions are just a failure in understanding viewpoints. Not everyone’s goals are the same, and not all brains work the same either. This is (part of) the reason that some people like Eminem and others like John Coltrane. Even though I personally strongly prefer Trane I don’t think he’s a “better” artist than Eminem, just a very different one, with a very different approach. The interesting part is that they both get the job done: providing a deeply enjoyable musical experience to millions of people.

              I’ll be honest, I don’t quite understand Eminem (or hip-hop music in general), just as I don’t quite understand why people are programming things a certain way, But on the other hand, I can’t fail to acknowledge that for the most part it does give results.

              1. 2

                I love both Coltrane and Hip-Hop, there’s so much to love in Hip-Hop even for jazz fans (matter of fact one of the biggest hip-hop fans I personally know is actually a jazz pianist). I’ll reply in length as soon as I can! cheers.

              2. 10

                Another approach I like is to start with exposing the elements that make up the argument rather than stating the argument itself.

                For example: since this project needs a standalone binary for distribution, is quite small and has some networking involved, Golang seems like a good choice. (possible expansion here).

                An opinion is just compressed information. If you want to engage in a discussion it’s better to lay down the decision tree and talk about it together rather than having a battle of ego.

                1. 9

                  I believe in weak opinions, strongly held. Sure the most I’m willing to commit to is “code review is probably more effective than pairing for finding bugs”, but by God am I gonna stick to that!

                  1. 3

                    with the code reviews I often see, I cannot imagine that this is even close to true…

                  2. 14

                    Thanks, I feel like a dick, now.

                    On a serious note, that hit me right in the feels. I wouldn’t go as far as saying that I am overconfident, but I’m definitely a loudmouth, and tend to argue to exhaustion when I think I’m right.

                    I’ve been working for a while in recognizing that I might be wrong and try to clearly communicate that I am open (and in fact, often hoping) to be proven wrong, but it’s a process, and I can’t help but wonder how often have I just been a massive dick and prevented other people from even throwing an idea =(

                    1. 8

                      I’m almost exactly the opposite. I realize that usually my preferences boil down to: “I like this way because I’m used to it”, or “I don’t think it matters”. I used to be the most experienced C# programmer on my team. My teammates used to ask me questions like “Should I assign a type explicitly or just use ‘var’?” any I would say that that’s a personal preference and it doesn’t really matter. Another programmer with a much more pronounced opinion joined the team, and I noticed that people liked his answers a lot more. Sometimes people just like to be told what to do, especially when the consequences of the choices aren’t so clear.

                      1. 5

                        Sometimes people just like to be told what to do, especially when the consequences of the choices aren’t so clear.

                        This, and also cases where the consequences on either side of a tradeoff are so minor, that the energy spent deliberating them out is costlier than just making a call and moving on. I spent far longer learning this, and fixing my behavior, than I wish I had.

                        On so many of these low-impact-low-cost details, where I did not strongly care, where I could expect another senior to reasonably disagree (even just for reasons of “this is what I am used to instead”), I would give teammates a full-depth, longwinded explanation, only to end in an equivocation and let them decide. Maybe one out of a dozen times they learned something, but often I was wasting people’s time when they just wanted a boolean result from the ask_swifthand function.

                        The first step in fixing this was realizing why I always led with longwinded explanations. It turned out to be misapplied empathy (which I assume-by-default that I lack, hence it took a while to realize).

                        I personally value knowing the detailed ins-and-outs wherever possible, and so I saw extending a full explanation to another person as a sign of respect: share all the learning, and help them make their own decision. But when that comes by wasting their time (on a low-impact-low-cost choice), was not being perceived that way.

                        Change was slow. I first changed my explanations to lead with the decision before launching into the explanation, giving a TL;DR to frame their expectation. After some time doing this, I began leading with the decision and then asking “There are reasons I could dive into, but they are subtle and inconsequential. So I’ll leave it there, unless you are curious?” followed by a long, awkward pause while I wait for an answer.

                        Due to some personality quirks that pause is the hardest part for me, but it gets the job done. It balances respecting their time by giving them the option to walk away with my desire to extend a courtesy that I would (personally) value of learning more.

                        Ironically, I do not always extend this sort of preemptive empathy to the rest of my life’s conversations. The last paragraph in the article really hit home, as I know friends and family often feel my SOLH is overconfidence layered with stubbornness.

                        1. 3

                          Sometimes people just like to be told what to do, especially when the consequences of the choices aren’t so clear.

                          Newbies need to be fed opinions because they don’t have enough experience to have any of their own, and the ones they might have at the moment are more likely to be wrong than right. Like, yes, you can write thousand-line subroutines in Python, that’s certainly something the language allows, but more mature opinion will state that writing smaller subroutines is better for reasons of ease of modification a neophyte hasn’t run into yet.

                          Gnomonic pronouncements make things go faster, but once the learner has some experience of their own, pushback is to be expected and should be encouraged. Except with things like styleguides. Those are completely arbitrary and set in stone for precisely that reason.

                          So teachers need to have strong opinions because they can’t go over every single thing when they answer a single question, but those opinions must be held loosely enough that students can successfully challenge them as they gain enough knowledge to exercise their own agency.

                          1. 2

                            This is one thing that’s true: people like to be told what to do. People should be encouraged to think for themselves. The best of both worlds is perhaps by explaining the difference and explaining why you lean the way you do. That way you teach thought and give a straight answer.

                          2. 5

                            tend to argue to exhaustion

                            that doesn’t sound “loosely held” :)

                            1. 1

                              I recently had an argument that went on for an hour until I conceded from frustration. Once we switched from the what to the how they immediately took to my original argument. Loosely held seems like it can be in the eye of the beholder.

                              1. 1

                                Well, for some definition of loosely =P

                              2. 3

                                I am open (and in fact, often hoping) to be proven wrong

                                Say that clearly and directly. Tell the other person that you are not feeling personally attacked and hope to be proven wrong.

                              3. 5

                                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.

                                Funnily enough, I usually experience the opposite if there are two bombastic engineers on a team. Endless back-and-forths that can be productive but usually aren’t.

                                1. 5

                                  Like much of nerd online reflection for the past decade, there are multiple cultures at play with, frankly, different value systems.

                                  There was and is certainly a culture where SOLH was the accepted norm for all in the room, and it works well in that scenario. When that mixes into, shall we say, less default-assertive cultures, then things really go off course. People get overridden, miscommunication becomes rife.

                                  1. 2

                                    Perhaps the author should take their own advice?

                                    My opinion of never use kdb+ untill hell freezes over has saved my latest company some few million dollars once we put the nosql stop gap in place and figured out actual usage.

                                    Here’s another strong opinion loosely held. When your company is in a death spiral to the bottom with Chinese manufactures beating your price point by an order of magnitude you have to hit that virtue button pretty hard to stay relavent.

                                    1. 1

                                      The value of strong opinions is they can be clearly stated and decisively acted on. The idea of SOLH then becomes if we go out and execute on the idea, we can quickly figure out if something goes wrong. In contrast, with a wishy-washy opinion, it can be really hard to evaluate if the opinion has merit or argue one way or another for it. The advocate of this opinion doesn’t even seem willing to defend their opinion. No motivation exists to make strong arguments for the idea or anything really involving the idea. The result is no clear action or evaluation.

                                      What often makes people bristle about SOLH is no such evaluation occurs. It is just blowhards spouting off at one another with no self-reflection. Or that the best ideas are hard to articulate. But the former is against the spirit of “loosely held”, and the latter conflates Strong Opinions with Simple Opinions.