AI Industry Veteran Warns: Changes Bigger Than the Pandemic Are Coming

Few texts in the tech community manage to irritate readers and make them stop and think at the same time. Matt Shumer's essay "Something Big Is Happening" is one of them. It spread quickly across specialist forums, reached Habr as a translation with commentary, and has sparked a debate that has not died down for several days. Shumer's central claim is provocative: we are on the threshold of changes that will turn out to be bigger than the COVID-19 pandemic.

Shumer is not just another Twitter evangelist cashing in on hype. For six years he built AI startups, put in his own money, hired engineers and brought products to market. He knows the industry from the inside, not from press releases but from sleepless nights over prototypes and negotiations with investors. That is why his words carry weight: when a practitioner with this experience says he "sounds like a crazy person," it is worth listening.

The analogy with February 2020 is central to his argument, and it works better than it might seem at first glance. Recall that in late winter of that year, most people in Europe and America saw news from Wuhan as a distant curiosity. Some bought masks "just in case," but there was no mass reaction. Three weeks later, the world shut down. Shumer argues that the AI industry is in a similar phase of collective denial. The technology is developing exponentially, but society, including a significant part of the tech community itself, still thinks linearly.

What exactly does Shumer mean? He is not saying that a superintelligence will appear tomorrow and enslave humanity. He is talking about more down-to-earth things that are no less transformative. Over the past year and a half, language models have gone from impressive but unreliable demos to systems capable of doing real work. AI agents already write code, conduct research, process documents and manage workflows. Each new generation of models is not just slightly better than the last: it qualitatively widens the range of tasks that can be delegated to a machine. And the pace of these improvements is not slowing down but accelerating.

Critics, of course, point out that we hear predictions like this regularly. Blockchain was supposed to upend finance, the metaverse to replace reality, and every new iPhone was declared a revolution. Skepticism is healthy and necessary. But there is a fundamental difference: with AI, we already see real economic impact. The world's largest corporations are restructuring their business processes. Venture capital is flowing into AI startups in record volumes. The labor market in a number of sectors, from copywriting to entry-level programming, already feels the pressure. This is not theorizing about the future; it is statistics from the present.

The context in which Shumer's essay appeared matters particularly. We live in an era when new models come out almost weekly, when every major tech player is running in the AI race, and when regulators around the world are trying to build legal frameworks but cannot keep up with the technology. At the same time, hype fatigue is growing: people are tired of loud promises and have started to brush them off. This, Shumer argues, is the main danger: not the technology itself, but the chance that society sleeps through the moment when preparation is still possible.

Despite its emotional charge, the COVID comparison carries an important practical lesson. The pandemic showed that the countries, companies and people who reacted earlier got through the crisis much more easily. Shumer calls for the same approach: don't panic, but don't ignore it either. Study the tools, rethink career strategies, and think about which skills will remain in demand and which will be automated.

One can argue about the scale and timing. One can doubt specific forecasts. But brushing off a trend that is already changing the economy, the labor market and the very nature of intellectual work is a luxury few can afford. February 2020 taught us one thing: when practitioners start sounding the alarm, it is worth at least allowing that they may be right.