Joshua Go

Today, I realized one of the key challenges facing small business: growing while preserving the quality of service. For a business selling goods, there's mass production: leave it to the machines. But in the service industries which still require high levels of skill, this is a little trickier. It's something we've still got to figure out. I thought of this today as I went to see Douglas, a mechanic friend of my dad. He's a good mechanic, and he set up shop for himself in Mountain View a couple of years ago. My family goes to him every time we want to fix our cars — not only because he's my dad's friend, but because he doesn't charge through the roof, because he won't charge for trivial things, and because he does a thorough job. His customers have been known to bring him pizza and other treats out of gratitude. Every time I go to his shop, all the bays are full, with several cars queued up waiting for each bay. I asked him if he ever thought about putti...

There are many reasons why I prefer not to argue with people. The two biggest reasons for me are that I don't want to waste my time, and I prefer not to rock the boat by creating personal unpleasantness. After thinking about it, though, I've decided that it's good to argue. As someone who has worked well on my own for a long time, I'm finding that working closely with a team and putting myself in a leadership role requires that I look beyond my personal productivity and look at how to build up the team. By avoiding arguments, I save myself time from explaining myself, but alienate the rest of the team by not involving them in my thought process. In economic terms, they are forced to operate on imperfect information--and the fault of this lies squarely with me because I have an aversion to heated arguments. Because I don't take the time to argue, I also force myself to operate on imperfect information. Yes, the argument can be unpleasant, but when people listen, it a...

The web is good for quick, throw-away information (time-sensitive things like news) and for reference. It's fast. It allows jumping. It can be updated in real-time. The web is not good for books or anything requiring deep thought. It's too distracting, and the display technology we have today makes it too straining on the eyes to read for too long.

So these days we've got Ubuntu , Ruby on Rails , and Aptana . I'm used to running Fedora, writing webapps with PHP, and doing my development in a text editor such as vim or emacs. It seems like there's something new coming out everyday, and it's a choice between checking out these new offerings or getting something done. I figure that if the tools I'm using are good enough to get things done, why switch? I'd only consider using a new tool if it allows me to do something that I can't do with my current toolset.

The trouble with aggregate measurements and policies is that they fail to take local variation and local circumstances into account. This is impossible even with the best-intentioned top-down policies. And this is why we've got local government. Aggregates simplify the complicated picture and most of the composite information is lost when presenting the aggregate.

Today I went to the post office to mail three envelopes to my brother. I found out that it would cost me over nine dollars to mail them separately, and the lady helping me advised me to just get it mailed in one package since they were all going to the same address. The total came to $4.05. What savings! If the post office were privately run, they would have an incentive to maximize profits, but since it's run by the government, I'm looked after as a consumer.

Macro-level views are useless without micro-level material to work with. Suggestions about how to run a company are useless if there is no company to work with. For all the attention that macro-level views get in the business journals, what I personally need to pay attention to now is producing the micro-level stuff that gives the macro-level theorizing a reason to exist.

The past couple of days, I've been working on putting together a software release and the manual to accompany it. I found the task pretty overwhelming, especially since I had to edit the document while putting the software together. One thing that I realized towards the end was that I could cut down the apparent scope of the job by ripping out pages that did not need to be edited anymore. I wouldn't have to worry about touching those pages, and could focus on the remaining pages. This helped me feel like I was making progress, and it actually helped me move forward by keeping my focus on the next page that I wanted to finish (and rip out).

This afternoon, I attended a lecture by Dr. Dean Simonton , professor of psychology at UC Davis. He talked about scientific genius and creativity. One of the common work habits of people who made major contributions to knowledge in their fields was to have a network of enterprises --that is, to have multiple projects going on at the same time. Instead of focusing on just one project until it's done, having a network of enterprises means that these projects feed off each other and contribute to each other, essentially making the ultimate outcome more than the sum of the parts.

When assigning people in a company to do knowledge- or research-intensive work, I can see the benefits of working in pairs. When I'm working as part of a pair on a difficult problem, I don't feel like I'm alone, incompetent, or somehow not measuring up whenever I run into unavoidable (but formidable) obstacles. I don't feel so bad because my partner is going through the same thing. When we report on our progress, it's less psychologically taxing because the burden of responsibility is halved. It might be an ideal way to do things in a lot of cases, because it doesn't completely alleviate the workers of responsibility, but distributes the load on each person to make it more bearable.

As an Asian-American, it might be alarming to read Bush's details on a foiled terror plot in Los Angeles. The article states that Southeast Asians were recruited to carry out the planned attack. I don't know about anyone else, but I'm figuring that Asians will be profiled from now on. I'm okay with that and I understand the concern for security, but I'm guessing that the whiny Asian-American activist types will not be happy to hear this. They'll make a stink about it. Eventually, the smart thing to do would be just to profile young people everywhere. Once Middle Easterners and Asians are profiled, al-Qaeda may turn to other ethnic groups to carry out their plans.

I was reading Eric Hobsbawm's Industry and Empire this weekend. One interesting argument in the book is that Britain suffered decline because it got complacent about its dominant position. If it was being out-produced or out-innovated, it could retreat back into the comfort of its world empire. Having a power base is useful to buffer against shock or sudden change, but the factors causing the shock or change shouldn't be ignored. Even though it's reassuring to have something to fall back on, this comfort is a temporary solution and something must be done about the root cause of the problem. Another book I read this weekend was Murdering McKinley: The Making of Theodore Roosevelt's America . Its thesis is that the man who shot McKinley was only one part of the "murder." The McKinley legacy was also wiped out by his successor, Theodore Roosevelt. (The author of the book, Eric Rauchway, is a professor at UC Davis. I took his course on the Gilded Age and Progress...

I was reading Eric Hobsbawm's Industry and Empire last night, and he wrote about how factories in Europe and the United States were often set up by Englishmen, or--more interestingly--through illegally copied designs. It was interesting to me because something that was illegal grew to something legitimate, even to the point that this initial "theft" went on to transform national economies. It reminded me of how pirating in Taiwan and South Korea used to be a big problem. Despite that short-term loss in revenues to software companies, their tools proliferated and they gained market share. People learned to use that stolen software, developed skill in using it, and even went on to create their own. As a software industry pops up, people learn to appreciate the effort that goes into writing software. With this, a respect for intellectual property is instilled. This just reminds me how messy the world is. Things end up working out in the end, though. It just seems funny when...

Those of us working with advanced technology, who push the frontiers of knowledge forward, take a lot of what we know for granted. We take our mathematical ability, our economic sensibility, or our personal responsibility as givens. It's always mind-boggling for me whenever I'm reminded that there are a lot of people worse off than I am. Sometimes it seems like the best gains that I could possibly make by working in advanced technology to better the world are marginal, at best. The higher yield activity would be to reach people, care for them, educate them, and convince them to take better care of themselves. If only Americans as a whole could be convinced to exercise consistently and to eat right, the impact on their health would be more substantial than if I were to discover the next wonder drug. Rather than spending absurd amounts of time developing myself to understand ever more esoteric realms of knowledge, I could instead equip people to make a good living for themselves ...

I've come to think again about what I said about government paying out dividends: that it's a bad idea because the money can be invested into big infrastructure projects. It's not so one-sided to me anymore because I've thought of three advantages to paying dividends rather than keeping the money, even if it can be used for big infrastructure projects. First, dividends modularize the allocation of money and make investing easier to manage. It splits up a large sum into more manageable chunks. It's hard to be careful when sitting on a billion-dollar pile of cash, but if you give it to people in smaller amounts, they'll be more frugal with how they spend the money. Amassing money is good and necessary for big projects (like infrastructure), but if we want to use money efficiently, a finer-grained approach makes more sense. This doesn't mean we abandon big projects completely; we'll always need roads. Second, dividends promote good will among the recipients...

There seems to be some magic in communicating well as a leader. You can do everything right in a leadership position, but some of your tougher moves tend to come off looking bad to those you are leading. This article at Fast Company tells the story of a pilot who communicated well with his disgruntled passengers. Yesterday, my company had a meeting when the Vice President came to our branch office and told us how the company was doing. There was one disgruntled person who boldly spoke up about how disgruntled people were that we had our company health plans changed in the middle of the year on short notice. The leadership at this company meant well, but people felt exploited and betrayed because they didn't know what was going on. The lesson here is to involve people at the lower level when decisions are made higher up. Show them the intermediate steps in your thought process, especially when you make difficult moves.

It seems like the key to success lies in publicizing something. There's a lot of good technology out there, and for it to be considered good, it's got to get everyone casting in their lot with it. That's what separates sites like Wikipedia from some Joe Schmoe who sets up his own Wiki site. That's what sets apart MySpace and Xanga from someone who downloads the LiveJournal software to run on his own server. So, building a good product seems to be the easy part. These good products themselves are often commoditized, and so the hard part--that which is most crucial to the success of the product--is telling people about it and convincing them to keep on using it.

I'm working on a project with IBM Rational ClearCase, and the builds take forever. It seems somehow wrong that the bottleneck is the computer's performance when I know it should not take that long. It's one thing if I'm the bottleneck as an engineer when I have to think about the problem, but with something like editing, compiling, and linking source code, I should not have to wait long to see if my changes were successful. In our quest to achieve efficiency, maybe a key indicator is to see where our IT bottlenecks are. In my case, the bottleneck is in the network filesystems. I think the compiler is also licensed only for a few machines, hence the centralized build system. If an unlimited-use compiler were available, the builds could be performed on each engineer's machine, now that workstations have caught up in computing power to yesterday's servers.