Tech Takes the Pareto Principle Far – Bobby Lockhart

There is a reason that video games have built in what is called a ‘vertical slice’. If you’re not familiar, a vertical slice is a place to play, with all the mechanics, final art, vfx, sfx, music, and more. The vertical slice is what game developers show to publishers and investors to show not only that the game itself will be good, but also that the game development team has all the skills needed to deliver the game to level of polish demanded by the market.

Contrast this with the technology industry, which submits for approval an ‘MVP’. A minimum viable product is the absolute least one can afford to pay someone. It seems that tech investors are very used to evaluating MVPs, and rightly so – they need to assess the potential of these prototypes so they can decide where their investment is worth. The problem is that MVPs don’t actually establish when the team /can/ reach a finished product, and in practice many don’t.

Recently, getting VC funding has become an end in itself as engineers in the technology industry, centered around Silicon Valley, optimize their skills for prototyping. There is an often quoted idea called the Pareto Principle, which states that 20% of the effort produces 80% of the results. So, if you can prioritize the right 20%, you can get most of the way to the desired result. The entire sector has done well in this, almost without participation at all. And who can blame them? Look at the reverse — doing 80% of the work only to complete 20% of the work isn’t much fun.

What the Pareto Principle misses, and what its followers seem to forget, is that you HAVE to DO that last 20%. End users usually don’t enjoy using 80% of a website, or driving 80% of a car. Unfortunately, with many digital products abandoned at the funding stage or forced to release early, engineers and designers often don’t have any practice in the last 80% of the effort needed to get something done. I’ve worked with many such engineers, and it can be frustrating to know that you can’t reach the level of polish the concept deserves.

Because there is such a culture around the early adoption of new technology, there is a large population willing to forget that the things they use are not finished. It’s good to have people willing to try new products, but we don’t need products lionized, or the people who use them. Because of their approval, the wider population began to protect the amputees. This also applies to games. Day 1 patches are the usual, like DLC that feels like it should be part of the core game. When products remain incomplete it is usually because all potential customers have already paid and there is no financial incentive to finish. How many of the products you use every day feel like they need a few more iterations to work properly?

However, there is another, more frustrating reason why a product may remain unfinished. It may be literally impossible to complete. I think that’s the situation we find ourselves in for some AI applications, like self-driving cars, image processing, and text generation. Even the people promoting these technologies rarely state that the results are useful, especially in a world where people are used to higher, human-level quality. At best it is useful as a starting point for someone to quickly finish the image, or the cover letter, or take the wheel. The problem is that I don’t think the current approach is going to get us the other 80/20% of the way.

I will leave my main point for a moment to justify that statement. It’s funny to think about now, but in the 70s, AI researchers believed that they were most of the way towards achieving AGI (artificial general intelligence, aka AI from the movies). They think that if an expert system, or a perceptron, or a set of predicates is sufficiently developed they will eventually achieve sentience, or at least eliminate the tedious work. Many believe that the hardware explosion promised by Moore’s law will be enough to create AI, and the software will take care of itself. Some of this is true – expert systems manage things like WebMD, and constraint solvers manage the unique logistics of modern freight. The limitations of the techniques of the time were not felt until later. We are also in Generative AI.

I think that the Pareto Principle is technically true in many fields, but I also feel that our society would be better off if we didn’t know about it. As I said, doing the last 80% of the work to produce only the last 20% of the result is morally difficult. No wonder the job is often left behind, or outsourced. Perhaps more investors should demand to see a vertical slice, instead.

If we take a more artisan view of the software, we will realize that a chair is not 80% done if you can sit on it. It’s the details and the polish that make something worth using. So while from a utilitarian perspective there is something that has most of the features a person could ask for, from a humanist perspective 20% of the work still produces only 20% of the results.

Bobby Lockhart an award-winning designer of learning games.

Keep in touch with Bluesky or to LinkedIn