So
@grok, what do you think of my latest Substack article? :
The Office vs The Job
I’ve just realised that one of the oldest memes on the internet beautifully demonstrates why I’m not at all worried about AI taking my job: [see picture]
OK, so the meme is a joke about the state of the modern workplace, and how we rely on migrants to do the real work. But what happens when you go around all of these jobs, and ask yourself which will be the last to be automated by AI…? Well, it has to be the “Bloke From Poland”, doesn’t it? In which case, all of the oversight of the AI systems collapses into that one point.
@ericweinstein Eric Weinstein says, of AI, “it breaks the capitalist model”. Yes, but strictly speaking it collapses it to the single point of the necessary task. What is the necessary task, in that picture? It’s the digging of the hole. What tasks are being automated the quickest? Digital tasks are being automated en masse, as you read this. In other words, all of the tasks being performed by all of the others in the circle. By the time we get to humanoid robot that can dig the hole, all of the responsibility for overseeing the AI implementation of those tasks has fallen on “the bloke from Poland”. And he is also the one available to capture the value previously captured by those in the circle. And after the invention of a humanoid, hole-digging robot? Then the best place for a human with an interest in digging a hole - and in capturing the value of digging a hole - is next to the robot, and next to the hole.
The same goes with trucking. What is the one necessary task that cannot be got rid of, in trucking? The moving of the load. To what point does task automation collapse the model? Towards the thing that moves the load, towards the vehicle. Why have the human with an interest anywhere else?
Admittedly this will be a process, rather than a sudden change. But, in my opinion, it will be a process driven by economic forces. Which is not to say that it won’t be resisted. People in offices won’t want to lose their jobs! And the rise of office work has created a social and hierarchical dichotomy. Many of those in the office gained resources and status because the cognitive requirements needed gave them bargaining power. And then they were in the position, the physical position, to verbally fight for their status. Meetings are often as much about ensuring interests are met as getting the work done, and those on the road can’t go to meetings.
Yet when one thinks about the realities of autonomous driving a different picture emerges.
Full level 5 autonomy won’t be suddenly be introduced, it will be level 4 with a driver still in the cab, only needed to drive and work at certain times. So then you have a person sitting there, being paid, yet not doing anything. Sooner or later someone will say, “why are we paying this guy to surf Instagram? Let’s get him to do something”. Especially when AI is simplifying a lot of other tasks in the organisation. Let’s say an AI routing agent is sourcing potential backloads. The driver could be checking the agent isn’t doing something bizarre, while he’s sitting there. Or he could be skimming the emails an AI customer service agent is sending. Or a host of other things. What matters isn’t that AI is automating tasks, it’s the speed at which it is automating different tasks relative to each other. It pushes oversight of tasks to the point where progress is the slowest. Which, in the case of road transport, is towards driving. By the time level 5 autonomous driving is achieved, the ‘driver’ - by then a largely nominal term - is overseeing so many tasks he becomes indispensable.
And what if everything becomes automated…? Well then we look beyond human tasks to human interests.
This is an area where freight transport is fundamentally different to passenger transport. In passenger transport the ‘load’ is the point of human interest: passengers know whether or not their interests are being met, whether their money is being well spent. With freight transport, the human interest is remote, in fact it is the remoteness that creates the interest: the goods are in a place different from the humans that need them. Absent the interested party travelling to where the goods are, which kind of defeats the point, that means someone acting as a third party, someone who earns a living from interacting with the parties at both ends. That has been the function of a merchant, or trader, since the dawn of human civilisation. It is still true now, and will be true for the foreseeable future. When we look to profit from AI, we shouldn’t look to what changes, but to what stays the same.
The necessities of office based tasks have shifted our perceptions of ‘trader’. We now think of someone sitting behind a screen, talking on the ‘phone, as a ‘trader’. That has meant the creation of a series of human, technological, and legal parameters to ensure interests are being met. Meaning yet more tasks, more costs. As autonomous systems collapse tasks, they don’t just collapse them towards the necessary function, but towards the point of interest. Where better to ensure your customer’s load is as described than standing next to it?
This is where I see trucking going. The narrative of truck drivers being replaced is a fundamental misunderstanding of the nature of the change. It assumes the job remains exactly as it is now, which is highly unlikely. I think in twenty or thirty years time, people will still be earning a living by travelling in trucks, and probably a much better living, even if ‘truck driver’ becomes a purely nominal term, if it’s used at all. If there are still human jobs in the sense we understand, one of the most desirable jobs could well be that of Merchant Trucker.