Getting It

I wanted to commend a post at Outside the Beltway to your attention. The post is about the impact of LLM AI on jobs in the United States. James’s observations are pretty quotidien (technology is improving rapidly, employment is increasing slowly, productivity rapidly, etc.). It’s the comments that are interesting. Most are the typical outright rejection and denial but there’s one, made by individual who comments here occasionally, that hits the nail squarely on the head:

It can generate code at the level of an intern, requiring large amounts of supervision. Typically, interns are useless on net, and the internship is just a months long job interview looking to see if the little rugrat can show signs of growing past that.

and

I worked at a startup decades ago, and quickly realized that what we were ostensibly doing was not the real goal — we weren’t really making a product, we were selling a dream to the investor class, telling them that we would be revolutionizing the internet and that they could get in on the ground floor. The actual users were just a means to an end, and their happiness was almost irrelevant, so long as a non-expert investor could look at it and think “yes, stupid peasants might like this.”

Here the dream is straightforward, and the dream of every C-suite executive: get rid of employees. It doesn’t have to be successful, it just has to be convincing to non-experts that long-term it will be successful. And even if all it does is make employees nervous and “grateful to have a job”… that’s part of the dream.

When I began working sixty years ago, big companies routinely and systematically trained new employees. That hasn’t been true for years. Today entry-level jobs exist primarily as training investments when they exist at all. Firms do not hire entry-level workers for productivity; they hire them to create future productive workers. If managers believe AI will replace that future worker, the rational action is to stop creating the positions now.

Other arguments are secondary at most. It doesn’t matter whether AI can do all jobs or, indeed, any jobs better than human beings in the near term. The point is what management believes and as long as managers believe that AI can do as good a job as a human employee, LLM AI will reduce the number of entry-level jobs created first, then higher level jobs.

That will be seen first where the largest number of entry-level jobs have been created, e.g. in South Asia. Entry-level technology jobs are already being disrupted in India, according to the India Times and Storyboard18:

Artificial Intelligence adoption is nudging India’s IT hiring in a new direction, slower at the entry level, steadier at the top, and sharper on skills.

A firm-level study by the Indian Council for Research on International Economic Relations (ICRIER), supported by OpenAI, finds that companies are moderating hiring, particularly for entry-level roles, even as mid- and senior-level employment remains largely stable.

This has happened multiple times in the past. Railroads massively overbuilt before profitability. MBA spreadsheet models drove layoffs in the 1980s before productivity gains existed.

The “title inflation” with which the technology sector will ensure that a significant number of jobs are eliminated. And it’s not just the technology sector. Any sector that has already experienced significant outsourcing including the financial sector will do the same thing. The number of associates being hired by large law firms will decline.

India is the canary in the coal mine.