“Software will eat the world” is now a famous saying, from Marc Andreessen, which has dominated the ZIRP era: Hundreds of startups, backed by cheap VC money, gambled on building platforms and application to eat that world. Few spectacularly succeeded, some survived, many failed silently.
Thanks to those survivors, now the world is filled with rushed, half-baked, over-engineered, under-tested, tech-debt riddled, and under-delivered software. They power the application that made the world around us kind’a work. It is very clear that the next wave of AI powered software will be about eating that software world. The question is how this new wave will develop, and how it will impact different types of companies.
The current software world is built on the idea of “software as a service”. The next wave will be about “software as a product”.
VC backed, cheap-money infused, solve-one-problem SaaS companies will be the first to be eaten. Whatever their size, their reliance on software and human engineers, will force them to either get more out-of-their software, as in more features and more domains. Or try to reduce the cost of their software development. The hiring and wages will be the first to go. SEcond will be more off/near shoring. And inevitably, there will be a lot of mountebanks, consultants, who will sell AI assisted software development tools, that will promise to reduce the cost of software development. I wonder how many of them will be able to go through the trasnsformation, while relentless competition, who can and will build similar tools/services from scratch using the same AI.
Software Engineering and Software Development Life Cycle will be redefined
When the problem of how to best manage/organize large numbers of engineers disappear, all the solutions that were built around that problem will disappear too. This will include engineering management, agile, career ladders, scrum, devops, and all the other buzzwords that are currently dominating the industry. The coders and journeyman will linger around, working for the companies where shor-term cost of replacing them is higher than the long-term cost of keeping them. After-all, there are still successfull conpanies, that never made it to cloud or never cared about Agile. They were not THE solutions, neither will be AI-assisted software development.
The Future is very bright for Human Software Engineers
The best analogy I can think of about the future of software engineering will be similar to mathematics. Most people will have the ‘concepts’ to build the simple things, using very simple, intuitive toole. A big number of it is will be automated and hidden. There will be some, very skilled people working on the those automated tools and infrastucture, very few picking their brains to solve problems using lots of assistance.
So, my advice to myself, is to ‘learn’ as much as possible about different architecture/paradigms in the most hands-on possible way, so I can understand the problems that have ben isolated from me. That way, I hope, I could design and debug large systems.