The Rise of the New Groupthink

-Nikola Tesla

Here's a link to the top positive review in amazon, giving an insight as to its content http://www.amazon.com/review/R1R3N8JI29OPPE/ref=cm_cr_pr_vie...

Lots of startups want to recreate the cafe experience as a workspace, but I'd rather work in a real cafe, with more people and more noise than an open space office. In a cafe, you still have somewhat of an expectation of privacy, there's less the prospect of someone interrupting you just because you're within reach.

The young guys seem to thrive in this setup. It's driving me nuts.

I am most productive when I have a problem to solve and I can immerse myself in it for hours every day, days at a time, without interruption. But now I have to do timely code reviews for others. I also have to respond to reviews of my code. The encouraged way of working is to commit small, reviewable pieces, but then each commit involves several interruptions, and it's difficult to go on to part 2 while addressing everyone's concerns about part 1. The infrastructure is problematic, because no one actually owns it, so there are more interruptions to deal with problems there.

Everyone has an IRC window open, and while there isn't too much social chatter, there is a distraction every time that a message arrives. (I hate IRC. I was forced to it in my last job, when people would IRC my officemate to get hold of me.)

I recognize the benefits of some of these practices. Code reviews do catch problems. (On the other hand, a code review often acts as a substitute for a design review, so we actually end up reviewing and maintaining bad designs.) More people know more of the code, and we avoid problems created by a guy going into a cave for weeks at a time.

But there is no denying that this style of working really makes it difficult to concentrate, at least for me.

This modern management cargo cult has reached really gross levels of absurdity.

If open plan is a cacophony of voices amidst which a person is to find focus and deep thinking, then cubicles are the same cacophony without the peripheral vision part.

I find teamwork, collaboration and such bullshit words just a mask for true motive. Cheapness. Cheapness in landlords, Cheapness in architects and Cheapness in companies who ultimately make up reasons for why they don't want to spend money and time on building truly productive environments.

Its sort of sour grapes situation.

Creativity is often spoken of as an end. I do not believe that is true. Sitting in complete isolation being creative does good to no one. Instead, it is a goal to share and execute on your creativity and vision that you have thought of in your own mind.

Programming is interesting because it is a marriage of creativity and skill--it is not art nor engineering but is a craft. You forge creativity with skill (and experience) into lines of code.

However, that does not scale. If we have 50 programmers working in complete isolation then we will have 50 different programs. Collaboration becomes key to scaling the craft.

The only way to scale craft is through constant, effective communication. The tools we use every day as software developers serve to enhance, enforce or provide communication. Even in completely remote projects like Linux, there are tools used to ensure communication: code is submitted to a committer in small, readable patches; style guides focus on the readability of code rather than their cleverness or creativity; changes are discussed and debated on mailing lists and IRC; subcomponent Czars are assigned that share the overall vision of the project.

The so-called "Groupthink" isn't about Creativity--it is about effective real change in the real world by actually applying the creativity that we have instead of allowing it to evaporate.

There are some business advantages to not being cloistered, not the least being people know you exist and may be available for whatever kind of work you do. And it's good to have be around smart geeks who can answer questions or give you a sanity check on ideas.

But the big gain for me is the serendipity. I don't expect to be dreaming up amazing creative thoughts in any sort of communal or committee activity, but the exposure to semi-random ideas and events is important to my own thinking.

Left alone (or alone with the Internet) people may tend to only read what they're familiar with, or approve of, or in some way filter out the things that might challenge their beliefs and expectations. That's fairly natural, but it's bad. You need to take deliberate steps to overcome that.

If you pick your groups well it can help free you from the prison of self-reinforcing thinking.

It is an anthropological finding that the human brain decreased in volume by 10% — about the size of a tennis ball — around the dawn of civilization 20,000 years ago. It is believed that the decrease in volume coincided with the emergence of cooperative, prosocial behavior, which enabled the members of reduced-cranium groups to solve some problems that eluded their more amply-brained relatives. The larger brains were adapted to more independent modes of survival. Bonobos have smaller brains than chimpanzees, but can solve problems that larger-brained chimpanzees will not solve, unless they decide to cooperate, which is unusual. [There are videos online of cooperating chimpanzees in the lab.] Successive generations of domesticated animals exhibit reduced brain volume in comparison with their ancestors.

The assumption that assimilation into the group is always good ignores the flip side of cooperation among prosocials. Prosocial behavior doesn’t imply feeble-minded docility. Fear of separation from the group, and antagonism toward larger-brained independent individuals is deeply ingrained. The reduced brain volume is compensated for somewhat by vindictiveness. Prosocials reward conformists and will punish transgressors at some cost to themselves. Road rage is an example. The capacity for revenge, even if this is costly, is the flip side of cooperation.

Groups are good at solving some kinds of problems, but not all problems (cf. Jaron Lanier, DIGITAL MAOISM: The Hazards of the New Online Collectivism http://edge.org/3rd_culture/lanier06/lanier06_index.html).

One should exercise caution when internalizing group values.

In a study published on July 15, 2011, in the Institute of Physics and German Physical Society’s New Journal of Physics, researchers have shown that swarming, a phenomenon that can be crucial to an animal’s survival, is created by the same kind of social networks that humans adopt. … Locusts rely heavily on swarming as they are in fact cannibalistic. As they march across barren deserts, locusts carefully keep track of each other so they can remain within striking distance to consume one another — a cruel, but very efficient, survival strategy. — Science Daily (July 15, 2011), Swarms of Locusts Use Social Networking to Communicate

It’s a locust-eat-locust world out there.

1. In my experience, private space and time is necessary for individuals to get real work done.

2. In my experience, small, focused teams which have as-needed discussions and well-partitioned responsibilities are massively more productive than individuals when it comes to solving large problems.

I've personally experienced benefits of that work style, but at the same time I've always wondered if in the process you lose the opportunity to be by yourself and deeply and creatively think about a problem from different angles, without having to share that with anybody while it's still brewing.

On the other hand, you have companies such as MySQL that were almost entirely remote, and still managed to get lots of great technology built, rarely, if ever, working in the same building. That did the trick for them, they were also quite successful.

I think at the end of the day you can have the cake and eat it too. How about this: if you need to do some deep thinking or some heavily creative work, you can do it on your own terms, by yourself, late at night, like the Woz would prescribe. If you need to do some work that you can easily wrap your head around, then working closely together with someone else or a larger team is a likely boost in effectiveness, as it provides you with safeguards and spare pairs of eyes.

Thoughts?

Pairing is the fastest way to ramp up in a new language, ecosystem, or codebase. And beyond just ramping up, it continues to teach you long after you reach mastery, because you're constantly exposed to new perspectives and ideas. Pairing is also good for the team... by rotating pairs frequently, you distribute knowledge through the team. It's like a RAID array for your codebase.

People talk about the cost of being interrupted. But what about the cost of having to read through someone's solo-written code when they're home sick and not understanding it? When your team is pairing, you can interrupt one person while the other person keeps coding, maintaining the focus for the pair. If you have a question, ask it. If the team needs to know something, tell them.

Working alone is good, too. I enjoy rocking out with my headphones or working quickly from a gut level understanding and not having to articulate it to a pair in words. But often I find that my solo code improves when I expose it to feedback from a pair. Not a slow asynchronous code review, but a real time back and forth conversation. For me, software is a social activity. Deadly silent offices give me the creeps.

Brainstorm sessions might surface creativity that has already happened. If you ask a group, "What are some product ideas we haven't considered?", there's a chance that some people have been independently thinking about new products, but wouldn't have thought to propose them.

Or they can serve to direct/request creativity. If you ask "What are some product ideas we haven't considered?", people who had not been thinking about new products, might start. It gives them something to think about on the drive home or in the shower. I've received some great ideas via email a few days after the actual session.

This is the essence of what's so destructive and horrifying about today's obsession with groups: the idea that everyone in the group is equal. Equal in talent, equal in work ethic, equal in drive, and equally entitled to being heard. And that the consensus of the group of purported equals is more important than the achievements of any of its constituent members.

The plain, and perhaps un-PC truth is that that's a load of hogwash. Some perspectives are better than others. Some people are smarter than others, or work harder, or are better read, or are more knowledgeable, or are more capable. Some creative visions are more correct than others, even if we imagine that it's not possible to be "correct" or "incorrect" in the praxis of creative expression.

Beyond these variances in ability or capability, we also encounter variances in style. As noted in the article, some people simply work best alone. Some disciplines -- writing, coding, art, etc. -- simply lend themselves well to conditions of quiet, relative isolation, and to long bouts of individual effort. That effort can be in service of a team, or as part of a team, but we shouldn't fool ourselves into thinking that creating by committee is in some way ideal.

The creative output of a committee tends to revert to the average creative capability of the least creative member on the committee. Social psychological studies have borne out that fact, and anyone who's ever been on a design committee knows it from brutal experience.

That's not to say, of course, that group brainstorms and the like can't be productive. They certainly can be. But the people in the group are everything. They need to be very sharp and very creative. They need to get along well enough not to clash over personal matters -- but not so well that they simply agree with one another and revert into groupthink. But such groups are rare, and I dare say that they're the exceptions to the general rule. High-functioning creative teams are things of beauty, but they're not for everybody or every situation.

Finally, the structure of a group is critical to its success (or failure). Groups with well established divisions of labor tend to be more productive, and far less painful, than groups in which everyone does everything together. Ever try to write a school paper as part of a team? It was either a horrendous experience or a fantastic one. The horrendous teams sat down together and drafted every sentence, laboriously and unceasingly, by committee. The fantastic teams either divided up the sections by areas of interest or expertise, or delegated the task of drafting the paper to the person who expressed interest and professed great ability in writing -- allowing everyone else to do research, discuss and debate the topics, and so forth.

Places which have nothing but cubes and sealed-off workers do miss some of the "creative" or "interactive" benefits of open, collaborative workspaces.

At the same time, places that strictly "open"...do indeed inhibit certain types of focused work. Isolation definitely helps progress in particular problems. There are examples throughout this thread of that.

I think the best work environments/attitudes are those that encourage a blend: the last few companies I've worked at have had open, collaborative workspaces, but at the same time really encouraged folks to go find their own nook (in the building or somewhere else entirely) when they needed to focus. The concept of "desk hours" or "continuous availability" was basically nonexistent.

It's all about balance. If your workspace lacks that in any aspect, you'll have something to complain about.

At my last job, all the programmers worked at a central table. I was criticized for working in my office (which I shared with one person) with my headphones on. This was called "siloing". They wanted people to work on code together.

I don't know about other programmers, but for me, when solving problems I find I often have many things in my head at once, many requirements to be met for the solution and sometimes a variety of variables whose meanings I have to keep straight.

I literally can't think when I'm constantly being interrupted. I can't think when I'm having to explain each symbol in a line of code to a co-worker.

Its impossible, until telepathy is perfected, for a co-worker to have the exact same understanding of what I'm doing and why. Sure simple obvious things are easy, but when you're constructing something original, you can't communicate it to someone completely because there are always too many details, even things you haven't resolved yet yourself.

At this groupthink place, every other programmer was in their 20s. They were all junior programmers (except for the founder of the company, who worked alone in his office.) They were not very productive. They adopted a whole lot of bad process, but they thought that by sitting around all at the same table they were more productive.

Maybe they were more productive than they would have been. But there was a huge gulf between the kinds of problems I was capable of solving (architecting major components of the product) and the kinds of problems they were solving (each unit test written was a task, everything was broken into tiny little independant, simple tasks.)

After I left, I decided I was done working for other people. I knew it was time to do a startup, because I'd reached my limit.

That company culture represents the view that people are cogs. Just like the east germans attempted to instill into their kids the culture of being cogs.

I think wozniak is right. Great programmers are artists, they live in their head. I wouldn't call myself a great programmer-- I'm competent, and I'm great at knowing my limitations.

But I make a really terrible cog.

That startup failed within a year after I left it. Somehow I doubt any of them think it was due to their process.