Matt Calkins (매트 칼킨스), CEO of business process automation company Appian, said AI will not be able to threaten the software industry, responding to some forecasts that AI will endanger enterprise software companies.

He pointed to the history of open-source software as a clue to gauging AI's future impact on the software landscape. There were worries that the spread of open-source software would drive down software prices, but that did not happen, and he said AI will be similar.

He recently told the Wall Street Journal, "Open-source software rose in the 1990s but failed to make even a dent in software companies' pricing power. Twenty-five years later, the software industry's size is five times larger than it was then." He added, "AI-generated code is more likely to play a bigger role in cutting internal costs at software companies than in lowering software prices."

He said that in software, elements beyond code are more important.

He said, "What software buyers want is not just code. They also want technical support, services, upgrades, expert communities, and the reassurance of buying from a proven company. Software buyers pay for trustworthy results. In areas like regulatory compliance or customer relationships, nothing can replace 100 percent accuracy." He added, "Red Hat built a large software company by selling the support, updates and services big companies need around open-source Linux code it does not even own. Software companies' competitive advantage comes not from code but from reputation and community."

In sum, he believes open source threatened pricing power across the broader software industry, but ultimately proved that code is not the core asset in enterprise software, and that AI will follow a similar path.

He said, "AI is a probabilistic technology. If you look inside, it is constantly guessing. That is an inherent characteristic of large language models. This is why AI hallucinates and makes errors. Because many applications require complete reliability, AI needs a deterministic software layer to provide direction and set boundaries. Code is getting cheaper, but the cost of mistakes is not. AI still lacks the reliability to operate fully on its own."