When is technology sufficient?
For instance, the definition of “good enough” for a mission-critical datacenter is “five nines,” which is 99.999% availability, or roughly five minutes of downtime per year. Does that seem a little excessive? If 99.5% seems satisfactory to you, that amounts to approximately 44 hours of downtime annually, or nearly two days. Then maybe not so good? Absolutely not if you were a passenger caught up in the recent chaos brought on by a power outage at NATS, the national air traffic control organization for the United Kingdom. However, there is a huge gap between “good,” or 99.5%, and “good enough,” or 99.999 percent. Some estimates suggest that every nine you add as a decimal point after 99% adds an extra zero to the end of the figure for total cost of ownership. Is it really worth striving for perfection? This is becoming an increasingly important question for the future of technology, because the Big Tech companies driving the digital revolution want to redefine “good enough”.
Take, for instance, driverless automobiles. It is remarkable what autonomous vehicles are capable of doing – an incredible achievement, brought about through billions of dollars of investment. However, would you entrust one with taking your kids to school each day? For the sake of argument, let’s say that autonomous vehicle technology is 99.5% good – and that you would trust one with your kids when it hits five nines. That is not an unreasonable expectation of what constitutes adequate performance. However, getting to “good enough” will most likely require the same amount of investment as before, if not more. It’s the same with generative AI (GenAI). It blows your mind the first time you use a service like ChatGPT or even Google’s AI Overviews. The researchers and developers of AI have accomplished something remarkable. We are all aware, however, that GenAI frequently makes mistakes. Are you prepared to stake your company’s future on the results of a GenAI service? It’s good, but not good enough.
And here again, the additional investment needed to make GenAI “good enough” is likely to outstrip the many billions already being ploughed into development. Because they need you to believe that “good enough” is imminent to justify all that additional investment, the executives of Big Tech are attempting to convince you that they are on the verge of some earth-shattering breakthrough. They can call it artificial general intelligence (AGI), superintelligence, or whatever fits your marketing strategy. But those investors want their returns. This means that driverless cars will become commonplace, and GenAI will run our lives and businesses. This implies that everyone, including customers, workers, and business leaders, believes these technologies are adequate. People in politics and policymaking talk about the Overton Window – a concept that defines the range of policies or ideas considered to be acceptable for mainstream discourse.
Politicians who are regarded as “extremist” attempt to expand or relocate the Overton Window over the course of years, sometimes even decades, so that previously unthinkable ideas become a part of everyday conversation. You can point to any number of examples in today’s geopolitics of issues that would once have been considered extreme that drive political discussion now.
If it is done in an ethical manner, we are entering a time of enormous potential and excitement in technology that promises enormous cultural, social, and economic benefits for all. But isn’t this also an opportunity for tech influencers to try to change the digital Overton Window—also known as the Altman Window? Zuckerberg Window? Musk Window? – to allow technologies that we would once have not considered good enough, to be sold and used as “good enough”.
And no one should consider that satisfactory.