A debate is heating up in the developer and investor community about the possibility that artificial intelligence could spiral out of control within the coming months. Some experts anticipate mass layoffs and cyberattacks, while others call for immediately disconnecting smart devices from the internet.

Amid these alarms, skeptical voices are also being heard, arguing that the development of AI models will soon simply slow down.

The Sudden Breakthrough Scenario

The gloomiest forecast came from Ed Leon Klinger, founder of FlockCover and an investor in Neuralink, Anduril, and Anthropic. He suggested that within just a few months, AI could begin creating its own updates, and so rapidly that humans would lose control over it.

This scenario has positive sides too: Klinger expects breakthroughs in mathematics, drug development, and manufacturing, as well as the creation of top-tier robots and flying cars. However, the price for this, in his view, will be high. Virtually all work currently done on a computer will be automated literally overnight. According to him, this will be followed by:

  • mass layoffs and the disappearance of entire professions;
  • rising inequality, protests, and looting;
  • cyberattacks on power grids, airports, biolabs, and transportation;
  • blackouts, virus outbreaks, and lockdowns.

According to Klinger, all of this will happen so rapidly that the average person will simply be overwhelmed. He compared such a scenario to a hundred COVID pandemics happening simultaneously, but assured that he is not trying to scare anyone.

Dangerous Smart Devices

An alarming warning was also published by Peter Szilagyi, founder of dark.bio, who led the development of Go Ethereum, one of the Ethereum network clients, for ten years. He advised everyone who owns smart devices to immediately disconnect them from the internet: pull the network cable and remove the gadget from Wi-Fi, and if that is not possible, throw it away.

"A storm is coming, and it will be scary," Szilagyi wrote.

He did not explain exactly what is supposed to happen. The developer only added that it is time to pay for decades of negligence in the Internet of Things — network-connected technology.

Skeptics Expect Stagnation

Developer Dmitry Kovanikov, on the contrary, suggested that AI models will soon reach a plateau. The reason lies in training data. Little useful knowledge has emerged in recent years, and if models are trained on texts written by AI itself, they only get worse. Moreover, because of AI, people have fewer reasons to create new knowledge.

Kovanikov does not believe that progress will stop entirely: AI will make fewer mistakes and better solve narrow tasks. But the main efforts, according to his forecast, will go not into the models themselves but into finding new ways to apply them. As an example, he cited Anthropic and OpenAI, which are already trying to challenge competitors in the corporate segment.

If the skeptics are right, betting on endless model improvement may not pay off, and companies that have learned to profit from existing AI will come out ahead. If Klinger is right, markets will have to endure infrastructure failures and the collapse of entire industries.

Cryptalist Analysis: The AI market is entering a phase where narratives about "technological singularity" and "the capability plateau" directly influence capital valuations. For the crypto sector, the key risk is not AI itself but the vulnerability of connected infrastructure and the rise in cyberattacks, which strengthens demand for decentralized and failure-resistant solutions. Investors should factor both scenarios into their strategies, prioritizing projects with real economics rather than speculative bets on endless model growth.