After DeepSeek, AI developers are wrong that The Jevons Paradox will bail them out

In 1865 British economist William Stanley Jevons explained to the public that increased efficiencies in the use of resources per unit of production do not generally lead to lower consumption of those resources. Rather, these efficiencies lead to higher consumption as many more people can now afford the more efficiently produced goods which carry a lower price tag. Jevons was referring to coal, the cost of which was falling and demand for which was rising due to increased efficiencies in production. His idea became known as The Jevons Paradox.

When the Chinese-based artificial intelligence (AI) upstart DeepSeek demonstrated last week that complex and powerful AI can be delivered for a tiny fraction of the cost and resources of current AI tools, DeepSeek's competitors cited The Jevons Paradox and told investors not to worry. Demand for AI would now grow even more rapidly in response to greater efficiencies and thus lower costs.

What those competitors failed to mention is that DeepSeek's breakthrough is great news for buyers of AI tools, but very bad news for current developers who are sellers of those tools. DeepSeek is giving away free or at only 3 percent of competitors' prices (for those needing application programming interface services) something comparable to the very expensive products of its competitors. This suggests that the hundreds of billions of dollars spent developing those expensive tools may have just gone up in smoke. That investment may never be recouped.

Democratization of AI spurts ahead: More power to individuals for mischief

While the world was focused on the actions of the Trump administration in its first week, a little-known Chinese startup startled the tech world last week with the release of an open source artificial intelligence tool that for a fraction of cost and resources of American competitors such as ChatGPT matches and in some cases outperforms its American rivals.

The company, DeepSeek, has gone even further, making its eponymous tool available for download for free. Only those seeking to use the company's application programming interface―something which allows the tool to interface more easily with existing programs—must pay a fee that is only about 3 percent of that of other tools.

DeepSeek sheds light on two important issues. First, U.S. export controls on advanced computer technology, both hardware and software, are meaningless against a rival that has a highly educated population with the training and tools to work around such controls. The lack of resources essentially forced the DeepSeek team to solve problems related to efficiency that they would not have bothered to tackle if U.S. technology had been available. The team ended up producing a superior, much more cost-effective product as a result. In my December 2024 piece "U.S. China trade war: Is the latest battle really tit for tat?" I wrote:

Wishful thinking? Sweden building nuclear waste site to last 100,000 years

The sensible Swedes like planning ahead. This time its storage for nuclear waste from its own nuclear industry—storage that is supposed to last 100,000 years. Nuclear power currently provides 40 percent of Sweden's electricity from six operating reactors. The Swedes expect to fill the storage site—"60 km of tunnels buried 500 metres down in 1.9 billion year old bedrock"—sometime by 2080 at which time it will be closed.

For understanding whether the target of 100,000 years of successful storage is plausible, I suggest a trip back 100,000 years to understand what surprises might be in store over such an interval. One hundred thousand years ago the Bronze Age, the age when humans first started to refine and work with metal, was still 97,000 years in the future.

It might seem that not much happened in those 97,000 years, but actually a lot that could challenge such storage schemes did. For example, somewhere around 71,000 to 74,000 years ago Mount Toba, located in modern-day Indonesia, erupted in a supervolcano thought to be the largest in human history. The eruption was two orders of magnitude (100X) larger than another famous Indonesian volcanic eruption, Mount Tambora, which caused what is now referred to as "the year without a summer" in 1816.

Elevator 'crisis' as symptom of our infrastructure predicament

I recently noticed that the elevator at my favorite cinema has been out of order for weeks now. The less mobile patrons need that elevator to transport them downward to this underground theater. I then learned of America's "elevator crisis" and my mind wandered to the struggle to maintain the Roman Empire. I'll explain the connection below. But first a refresher on Rome's predicament:

The Roman emperor Trajan brought the Roman Empire to its greatest extent during his reign (98 to 117 A.D.) with his successful conquests in what we today would call the Middle East. Rome's many conquests had been financed by booty taken from the conquered.

But its hold on the sprawling empire—one that reached from northern England to southern Egypt, from Spain in the west to what today is called Iraq in the east—would henceforth have to be financed by rising taxes and inflated currency. The money was needed to pay for armies and naval forces to defend the empire's very long land and maritime borders. Building an empire turned out to be less expensive than maintaining one, including building and maintaining the infrastructure of roads and military and political outposts needed to protect and administer it.

What happened to Rome's maintenance bill happens in any system of infrastructure as it expands. The existing infrastructure must be maintained even as new infrastructure is built. Eventually, it becomes very expensive 1) to pay collectively for maintenance of all existing infrastructure and 2) to pay for increasingly aging infrastructure that requires extra expense.

Taking a break - no post this week

I am taking a break this week and plan to post again on Sunday, January 12.