In the coming decades, as AI surpasses humans in building software, our ability to generate economic value may decline, or even disappear.
Original link:https://alexkolchinski.com/2023/12/11/the-end-of-the-software-industry/
Reproduction without permission is prohibited!
作者 | Alex Kolchinski 译者 | 弯月
责编 | 夏萌出品 | CSDN（ID：CSDNnews）
The AI revolution that is currently taking place has the potential to fundamentally change the software industry, leading to a reduction in the number of employees. Historically, there have been many smart people in the software industry who have gone to great lengths to tell computers how to perform specific tasks using obscure language that computers can understand. We call this process "writing software". And AI is disrupting that model.
Generative AI tools have been able to dramatically increase the productivity of programmers, and some of the more specialized or less valuable software that was previously unachievable due to economic factors can now be written. And even bigger changes are in the pipeline.
We are ushering in a new future: anyone with a certain level of intelligence who can understand that they want to solve a problem will be able to provide instructions in the form of natural language, and then AI will perform the information processing task. We don't need to manually instruct the computer to perform the task precisely. Essentially, the goal of the no-code movement is already emerging, albeit in a different way than it was expected a few years ago, with universal AI models making it obsolete for hand-writing Lego-style no-code tools.
This means that in the near future, the number of programmers may decrease significantly. However, more and more people will be doing what product managers and salespeople are doing today, which is to understand and solve customer problems, but the specifics of the solution will be implemented by the computer itself, not by a team of programmers.
This is similar to the trajectory of many other revolutionary industries, such as the railroad industry, where it is not uncommon for an industry to boom initially and then maintain a relatively stable economic position over a long period of time, during which time it has relied on less and less manual labor to maintain its position.
More and more work is being done by AI, and the experience in the software industry is broadly similar, a development that will have a significant impact on the startup ecosystem. Smart young people can start a business with little or no capital and make huge profits in a decade, and only the software industry is likely to have such success stories, after all, the software industry has only been in existence for about 50 years. Perhaps in the near future, this era will come to an end.
This is because a software startup is an organization of small teams of smart people who create new value in the market by building and selling new software, that is, they are able to manually tell computers how to solve specific human problems. As the need to build software manually diminishes, software startups will face a fundamental shift, or become extinct altogether. If this kind of "extinction" happens, then the startup ecosystem and the venture capital industry behind it will also be wiped out.
Of course, it's unclear when AI technology will advance to the point where manual software engineering will be phased out, just as high-level programming languages have eliminated today's rare machine coding. It could be next year, or it could be 50 years from now. But it's almost certain that all of this will happen within this century, and even if there is no complete "extinction" event, the nature and role of software will continue to change.
This puts software entrepreneurs, myself included, in a precarious position. Today, software is creating more value than at any time in recent times, even the highest ever since. Trillions of dollars of business processes can now be automated, creating huge benefits, and there may be more possibilities in the coming years. To meet these demands, building and selling software still requires high-tech teams to write code by hand, so there's been a huge boom in the startup ecosystem, and there are still a lot of entrepreneurs who are accumulating huge revenues from AI products this year, and I think that kind of revenue generation will grow significantly in the near future.
While startups are popping up one after another, many are quickly falling. If AI technology evolves to the point where it is possible to solve business needs using generic AI models with little or no human effort, some startups will quickly lose value in products that have been built manually on top of previous generation AI technologies, and if these companies fail to seize the opportunity to accumulate advantages in other areas, they will quickly lose revenue and wither.
Of course, some of the companies created in the current wave of AI are also entirely likely to build assets that are more durable than rapidly obsolete technologies, such as having valuable data or developing into intermediaries that other companies will be difficult to avoid. But many companies are likely to build and sell extremely valuable software, making huge revenues for a few years, and then those revenues quickly dwindle to zero as AI advances further, as such products can be easily replaced.
I think it's wise for today's AI entrepreneurs to consider this possibility. In particular, the large revenues brought in upfront can decline rapidly, and this dynamic must be taken into account when calculating the cost-effectiveness of raising funds. In the past, startups had to take on a lot of upfront engineering work in order to build a revenue-generating software product, so raising capital was often necessary. However, these revenues, once realized, are usually long-lasting, as distinctive and valuable business software usually doesn't become obsolete overnight. As a result, venture capital has always been a source of funding for software companies that require risky upfront investment, but can generate a large and lasting revenue stream if successful, and continue to provide returns for investors and founders if the company is successfully acquired.
For many of today's AI companies, this dynamic can flip. High-value needs that were previously unsolvable or impossible to be solved by business software can now be solved in just a few weeks ~ months by a good team, all thanks to the latest advancements in artificial intelligence, and these solutions can quickly generate hundreds of thousands or even millions of dollars in revenue. But these revenues can only last for a few years, and the underlying product can be easily substituted unless the company that sells the product creates a long-term competitive advantage by being stronger against the rapid development of AI.
Unfortunately, this dynamic carries significant risks for founders who embrace venture capital. Assuming that a founding team can generate tens of millions or even hundreds of millions of dollars in revenue through a product within 5~10 years, and then suddenly disappear, then whether to accept venture capital will have a very big impact on the final result. If venture capital is not accepted, then all of these revenues minus the costs go to the founders. If external investment is accepted, then the founders can only receive a meager salary and pin all their hopes on future growth, which is no longer there once the product has been steered. In this way, even if a small amount of venture capital is raised, the vast majority of the founder's income will go to zero.
My conclusion is that it might be wise for teams building products in the AI space today to abandon fundraising altogether, at least in the initial stages, not to accept outside investment, and to go all out to generate as much revenue as possible as early as possible. In many cases, a small, self-invested team can reach millions of dollars in revenue and then do their best to create a large and sustainable company. If that doesn't happen, then at least the company's founders will have a sizable, short-lived profit stream that can hedge against the risk of falling into a dead end.
Ensuring personal financial security may be more important than ever for software founders in this space, as our ability to generate economic value may decline or even disappear in the coming decades as AI surpasses humans in building software. There are flowers that can be broken and must be broken, and at the same time, it is necessary to reduce the risk of all the hard work going to waste.