There is a tsunami of the experts’ warnings on the eminent blast of the “AI bubble” referring to either negative profitability of the LLMs’ proprietary companies – or low final effectiveness of the AI-equipped utilities of the end-users.
This and that are true but the experts’ conclusions in general are not accurate. The experts address the situation by parts and do not recognize the pattern that the ascendance of the new revolutionary technologies historically demonstrates. The implementation of the breakthrough technologies – like gunpowder weapons and oceanic sailing ships in the 15th century, steampower and rifled firearms in the mid-19th century, and electricity, combustion engines, and high-explosives at the turn from the 19th to the 20th centuries – has one and the same model.
There are the suppliers of the new technology like the alchemists that mixed gunpowder in the 14th century, and natural scientists in the 19th century. There are the final users of technologies like the troops and navies in the 15th century, and industry and armies in the 19th century. And there are the developers who convert the technology into usable appliances. The gunpowder technologies required a century or two to overcome the knights’ armor and castles’ walls and take over the battlefields and sieges.
The development of the sailing ships took two centuries to transform offshore navigation into transcontinental discoveries. The issue was not the hardware – guns and handguns – that did not change much from the end-15th to end-17th century (when a flintlock came) and then to mid-19th century (when a modern rifle appeared).
The ship’s construction and rag were equally conservative. The issue was their utilization – an array and tactics of the troops and ships, deployment of guns, soldiers’ behavior and moral, and the commanders’ way of thinking. The technological characteristics of the industry did not change much from the end-19th to end-20th century – the same electricity, mechanical machines, and chemistry prevailed – however the economies of the advanced countries differed dramatically. The utilities and devices equipped with the new technology and pattern to utilize them caused the difference.
Today, LLMs may be in their technological deadlock while the much-longed general AI may be a fantasy. But the current potential of the LLMs is utilized in such tiny portion of it and in such unskilled manner that the true AI-revolution, based on it, is still ahead. The middle range between the owners of the LLMs and the AI end-users – developers of the AI-equipped utilities and devices – is responsible for it to come. This middle layer of developers – between the suppliers of the new technology and final users of devices, equipped with it, halves to two groups.
The first has abundant attention, they are the manufacturers and supplies of the material utilities to use the new technology. They were manufacturers of guns and handguns, and recruiters of soldiers in the period of the military revolution, and machine-builders in the period of industrial revolution, and today they are suppliers of data-processing computational equipment and AI-stuffed software like AI agents and AI platforms.
The second group is more obscure but more important. They are the people who developed tactic for the troops with firearms and armored steam-driven navies, technologies for industrial machines and electrical equipment, and develop the methodology to utilize AI today.
Methodology is the most important thing: victory belongs not to the troops that have more advanced arms but the troops that operate their arms with stronger impact on the enemy. It is the reason why we – in Strategy by AI – are focusing not on building computational capacity and building AI agents and AI platforms but on the solution how to utilize in possible maximum the huge capabilities of knowledge and reasoning that the current LLMs possess but that are critically underutilized. We do that in the narrow sphere of strategy for business and political organizations and our results – high-level strategy cases that we publish on our website – confirm: methodology is paramount, it is foundation of AI effectiveness. www.strategybyai.org