2030 is just around the corner.
It is less than 5 years away. As a comparison, COVID hit roughly 5 years ago. The point is time certainly does fly.
That means that we will hit this targeted milestone before we know it.
Over that time, we are going to see a massive transformation of the economy. This is due to the progress of AI along with robotics. We are embarking upon the era of AI agents, which will take over in the next 24 months.
As for the physical world, the ChatGPT moment for robots could be within 18 months. At that point, we will see the utility to the point where the implementation markets explode.
This, of course, means jobs are in trouble.
Image generated by Grok
There is a good chance that, by 2030, 80% of the jobs that are around today could be eliminated, at least within knowledge work. The physical world could be a bit slower due to the fact that robotic scaling is required.
The robots are coming.
Before they get here, the AI agents will take over. Consider most services that are available online. These will be automated into agents.
It is likely that we hit some form of AGI in the next 24 months. Also, as Deepseek showed, the amount of compute required to run these models is dropping a great deal. It is likely that a decent size model will run on a smartphone by the end of this year.
Hence, the question becomes what happens when everyone has AGI in his or her pocket? Why would people pay others for a service when the model can build as agent to do it just as well and not pay for it?
That is a point we will get to. In future articles we will discuss what that means for the economy.
As companies give way to agents, jobs are lost. We are going to see bankruptcies, probably on a scale never seen before (at least not since the Great Depression). This is going to obliterate incomes all over the world.
Are we already seeing this?
Have you noticed how tech companies are laying off thousands of workers? This strikes me as odd, considering the massive expansion into AI. Perhaps other companies are experiencing what Salesforce uncovered.
According to CEO Marc Benioff, Salesforce is not hiring any software engineers this year. In spite of that, their productivity has already increased 30% due to AI writing code.
Did Meta, Google, and some of the other developers of LLMs find a way to simply replace these workers, cutting the number of people needed?
It is my guess that they know something the rest of us do not. After all, wouldn't they be the first to implement this technology in a live environment?
What happens when that starts spreading to other industries. Benioff is on record as saying that he wants 1 billion agents on Salesforce, creating a virtual workforce. Think about that...Salesforce will offer companies a workforce of agents to replace its human workers. Of course, Salesforce will charge a fee for this, something that is going to be lower than what a human employee costs.
What happens when companies are churning out cities?
By this, I do not mean physical cities, although that will happen at some point. Instead, I am talking population.
Tesla is one of those companies that is going to have a line of humanoid robots. We will have to see how quickly they can scale but it is likely they willl produce more than 1 million robots in 2028 or 2029.
This is a workforce that is equivalent to a decent sized city. It is also one that comes completely trained on everything that was fed into the system to that point. There is no aging process, waiting for one to get through school to be capable of entering the workforce.
Then we have the ability to work 7,000 hours per year instead of 1,800 or 2,000. And this is just the physical realm.
Our global population is 8 or 9 billion. There will come a time, likely in the early 2030s, when this will be the yearly output of the entire robotic industry. Imagine building the entire global population each year.
That is what we are looking at.
All of this feeds into the automated economy.
