The Potato Chip Model
On this Memorial Day and one-year mark of starting this blog, I’m taking some time to recap how the tech landscape changed in the last year, explore what the “future of work” means for global economics, and highlight a few opportunity areas I feel are still underrepresented in media coverage.
Big players continue to compete at scale
This week Anthropic dropped fourth-generation Claude models and Google I/O almost exclusively focused on AI launches including a swing at Perplexity with AI Mode, which (from a brief side-by-by test I just conducted by asking both Google AI Mode and Perplexity to find me dinner in the Castro tonight) can provide more accurate and personalized results due to integration with things like Google Maps and your personal Google data. Shop with AI Mode, providing an “agentic checkout” experience and virtual try-on via Search Labs, has the potential to become a direct competitor to Amazon, although it will take time and a significant investment in quality to convince customers at scale.
These are crowning developments in the last year of the incorporation of AI and AI-adjacent technology into every and any vertical. The AI space is becoming exponentially more saturated, competitive, and bubble-y (see: OpenAI’s $3 billion acquisition of Windsurf) and yet multiple international conglomerates continue to funnel investment into expanding feature sets in a race to establish monopolistic control over the most valuable commodity globally: human attention. Similar milestones in the past where we saw this form of large-scale technological shift influencing human behavior and driving global manufacturing and software development trends were the development of televisions, personal computers, smartphones, and social media. The difference now is how much easier it is to spin up hardware and develop software to serve these customer applications. Iteration loops and talent are now orders of magnitude faster and more available with the proliferation of hardware and software infrastructure and developer tools, accelerated applications-oriented training and remote work.
Software engineering as a discipline and art
While we can all agree coding assistants and generative app developers aren’t currently capable of replacing experienced human developers entirely, the way we are studying and viewing software engineering as a career path is still changing significantly. Corporate developers are now far more focused on learning to apply AI tools to their work. Earlier stages of software development feature heavy use of assistants that can generate in a matter of days what previously might have taken weeks or months. Educators now have real concerns around student use of AI and assistants hindering their ability to actually understand concepts from first principles, which in my opinion is a valid concern, but also just another flavor of how technologies like Google Search and Wikipedia changed and democratized the ways we learn and process information from last-century era phone calls to information lines. The power still lies with the people to use this information in ways they wish.
To those that argue that the software engineering class will shrink significantly over the next century, that’s absolutely true in the traditional, 2000s sense of what being a “software engineer” means. Modern mobile and web app development, site reliability, devops, infrastructure; these can and will be increasingly abstracted away in the same way AWS and cloud service providers abstracted away actual data center and hardware maintenance serving software applications. At the same time, this creates a unique economic and social incentive for high-quality ideas and implementation to become noticeably better than competition. This opens up opportunity for engineers to become an increasingly creative class, where the purpose and skill required to build reliable, beautiful, human-serving software is amplified and even demanded by a continued and universal human desire for connection and simplicity.
The flip side
I’ve written before on the need to ramp up investment in sustainable hardware, environmental protection, and safety and security measures. These continue to be my personal areas to watch. Naturally and understandably, successful venture and investment funds tend to lean toward seeking out and therefore encouraging customer-facing product development to create and capture greater market share of the human attention economy. A pessimistic and capitalistic view of health and longevity investment might be that longer lifespans lead to more literal economic time. It is widely known that the food and pharmaceutical industries has twisted research incentives and public health outcomes, and critics draw parallels to social media and internet companies funding machine learning research.
My view is that it is impossible to avoid some level of “incentive impurity”, as any action taken will have some downstream effect that isn’t ideal. And, in fact, given the way our global economy has been built for centuries, perhaps we shouldn’t be trying to minimize our downstream negative impact but rather maximizing the positive influence of our actions and technology.
Here I want to break from the tech talk for a bit and reference two pieces of art that influenced me recently. (Some spoilers ahead.)
The first is John Wilson’s How to Cook the Perfect Risotto, which in classic John Wilson fashion documented his quest to surprise his landlord with her favorite dish and became an honest reflection on addiction, the inherent randomness and unfairness of much of human life, and thus the immeasurable importance of human connection, even the simplest of actions we take to show our love for each other. In one scene, John goes on a long drive to take himself out of an environment that reminds him of smoking, but is stuck behind a truck with a broken exhaust; upon questioning, the driver reveals that he is “addicted to diesel” and refuses to consider the environmental impact of his actions, citing (paraphrased) that bad chemicals and processes are impossible to avoid at this point, including the plastic that packages our potato chips. John comments to the viewer that it’s interesting how little regard some have for the impact of their actions on others and then continues his risotto quest. My takeaway from this in the context of this tech blog is that there will always be makers and competitors in any space with less regard for human and social outcomes (aware or unaware) that will continue to exist and even be rewarded for their actions, and our best course of action may be the peaceful stance of counteracting in ways we can. We can still learn to make risotto for our landlords in a world where diesel truckheads drive freely.
The second is an excerpt from Garth Greenwell’s Small Rain, in which a man teetering at the edge of death during the COVID pandemic reflects on the potato chips his sister (not knowing about his effort to reduce his sodium intake) has brought him at the hospital. I’ve termed this method of acknowledging and even appreciating the beautiful outcomes of a complex set of economic incentives the Potato Chip Model:
…Chips were exactly what I shouldn’t eat. But that would mean telling G she had brought the wrong things, I imagined her feeling hurt, and so I reached into the bag, taking just one chip. It was thin and almost weightless, an amazing object, really, if you think about it, a miracle of engineering… nothing in nature was like this, so perfectly tuned to our pleasure… I knew it was pernicious, one of the manipulations of capitalism… But maybe a Snickers bar [referencing another delicious product of capitalism] is a wonderful thing… as a product of science and experiment, and as the end point of a whole system of production and distribution, the ingredients sourced I’m sure from all over the world, which can only be abstract to me… whatever machines had been designed and built… systems of distribution to carry [and package these impossibly fragile things] all over the country, the world, so that you can walk into a store in the middle of America, in a college town in Iowa, and for a few dollars fire up those points in your brain that mean pleasure… it was at once amazing, proof of ingenuity and genius, and also the product of unimaginable suffering, of exploitation and violence and labor… and there was truth in that, too, the intrication of wonder and depravity, pleasure and violence.
I’ll leave it at that today. Thank you.
