I have realized there is an element to the previous post that I need to add. Right after the section about how The Current State of AI Is Like The Early Days of Automobiles, there needs to be a section talking about how The Competition Between AI Companies in 2026 is Like the Early Days of PC Operating Systems. It is as follows:
What I am saying is that the very fast and fractious development (and use) of AI in 2026 is very much like the early days of the development and fighting over the adoption of operating systems for the Personal Computer.
The early days of Personal Computers were complex and fractious. At first companies that made the hardware made their own operating systems that you used to make the computers do practical things. In the late 1970s and early 1980s, Microsoft had approached Digital Research about licensing their operating system, CP/M-86, but negotiations failed and Microsoft did not acquire the rights to Digital Research’s model.
By the mid-1980s, Microsoft quickly dominated the market for operating systems by acquiring a different operating system and calling it MS-DOS, adapting it to the popular IBM PC, which was the industry standard at the time, licensing its operating system widely among PC-clone manufacturers, and keeping the operating system as easy to port to as possible.
In the late 1980s, Digital Research struck back by creating its own operating system, DR-DOS, which often outperformed MS-DOS. DR-DOS was one of several DOS-compatible systems on the market, and all the systems competed heavily for market share and dominance. There was fierce competition, and bitter rivalries emerged.
Microsoft eventually won the battles for dominance when its Windows 95 came out in the mid-1990s and absorbed DOS into its boot process. As Windows came to dominate the PC market, almost all other operating systems fell to the wayside. This period of conflicting systems and market uncertainty is mirrored to some degree in what we see today in the Artificial Intelligence models market, with its competing models like Large Language Models (LLMs), with closed-source providers (OpenAI, Anthropic, Google) playing the role of Windows/macOS and open-source ecosystems (Llama) playing the role of Linux.
And that’s just to mention the Western Artificial Intelligence models. There are a wide range of Chinese models on the market being created and marketed in widely different ways than the Western models, and there is not only a philosophical but a cultural and a political dynamic to the relationship between the two different approaches that few people are talking about, and even fewer people seem to understand.
Also, the marketing approaches of the Western and Chinese models is drastically different. The Western models are mostly a pay-as-you-go economic model, often using monthly subscriptions for basic versions of the AI models. This is quickly becoming impractical, and prices are rising as well as other marketing techniques being explored.
The Chinese models are mostly open-source, and therefore either less expensive or free to use regularly, which gives the Chinese models a marked advantage in the marketplace. This is where issues of the control over and governance of work being created by the Chinese models comes into play. Some people fear that work being created by the Chinese models is subject to Chinese government interference, and some fear that Intellectual Property rights might be compromised if Chinese models are ever heavily regulated by the Chinese authoritarian government.
Like the early days of the PC operating system wars, there is not yet one (or even a few) dominant player(s) in the field of Artificial Intelligence, the field is fragmented, and the challengers seem to be updating their systems almost monthly (or more often!) with more and diverse advanced features. The competition between the different models and their makers is intense, and the unknown risks are basically untold as of yet.
And this does not even include discussion of the different kinds of tools created with AI: text-based writing and editing tools; image creation and editing tools; audio tools; and perhaps most disturbing, complete video creation and editing tools that can make video indistinguishable from professionally-created television and movies. Not to mention all the tools that can create new software itself. The possibilities seem to be endless, and there seem to be no guardrails of any kind at the moment to prevent a terrible accident from happening at any time.
To stretch the analogy into another sphere, this is not unlike the early days of the adoption and usage of nuclear power: AI systems, like nuclear power plants before them, are being designed and built as fast as humanly (and now inhumanly) possible, without safeguards being either fully realized or implemented, and we are in an unknown realm where everything is running at a breakneck speed. The danger is that we don’t even know what the dangers are yet. Essentially, we have not had our Chernobyl Disaster moment yet, and there is no idea of how to prevent it from happening, or even when it could happen to begin with. A terrifying thought to many who see the rampant, unchecked growth of AI as an existential threat to humankind as much as others see benefits in it.
