The software industry has experienced a number of booms since the 60’s, is still having them today, and will continue to see them tomorrow. If the last 50 years are any indication, there is still a lot to come within the realm of software. From Mainframes to Microcomputers; from the Personal Computer to the amazing Internet and its dot-com boom, to Mobile and Apps and the Cloud; from Relational Databases to non-SQL Databases, we have come a long way and there is a lot more yet to come.
Throughout all of these software booms there has always been a thirst for the latest and greatest knowledge. Companies on the bleeding edge of these technologies need it to thrive, and software engineers, developers, designers and architects quickly adapt and learn, happy to provide the required expertise. What has allowed this adaptation and evolution of the software engineer is the fact that the progression of booms was a mostly linear evolution of technologies that built on previous accomplishments.
While the coming-to-be of AI is certainly piggybacking on existing technologies, it is still heavily dependant on academia, and it represents a change in paradigm for the software industry. AI, unlike those other wonderful technology booms, requires the software designer themself to make an evolutionary leap. AI is functionally in-line with the progression of technologies, but not in-line with the existing knowledge base of the industry.
With AI, even expert software architects have a lot of trouble grasping the new paradigm of the computer learning by itself and the fact that they are enablers of that learning and performance, not authors. There are new roles in AI software and they are not that well defined yet, which is a enormous challenge for AI startups.
Furthermore, math is much more important in AI coding than it has been in other more traditional software, and it has been neglected by software developers. There are a lot of hard skills other than logic which will have to be mastered by software engineers if timely expertise is to be provided to knowledge-thirsty AI startups.
This all naturally leads to the single most important constraint AI companies are experiencing: the crunch for AI talent. Everything else about AI startups feels very much that same as other IT startups, there is abundant sweat equity and plenty of ease towards distributed teams, but that top talent is in short supply and competition is fierce.