Also statistically, every species in history that has ever over-specialised has gone extinct.

With specialisation comes interdependence. If you spend 99% of your free time studying computers it is almost certain that you don't know how to grow a potato. You depend on others to do things for you.

Over time specialisation results in critical and basic survival skills being lost from the mainstream.

Usually then, some event occurs that wipes out a critical group of specialists, and everything that depends on them comes crashing down, resulting in the species being wiped out... because the knowledge cannot be rediscovered, or the skills re-learnt, before the food runs out.

We are rushing towards a future where all of our food, water and energy is grown/harvested/generated/distributed by AI-driven machines. The number of people on the planet who will know how those AIs actually work will be minuscule.

Specialists already do things like attend conferences... so gather in a single location... breathe the same air while there... stay at the same hotels... eat at the same restaurants... fly the same planes...

If we care about the long-term survival of the Human species, then we should be very, very careful not to allow over-specialisation to occur in critical fields.

I think generalization is very important, but I also think this oversells it slightly: every other species in history has lacked the privilege of writing its progress down. A PhD in farm science doesn't need to know how to grow a potato today, because (in theory) they can open a book and learn how it's done.

My concern about specialization is less that we'll fundamentally lose basic skills, but rather that our academic and cultural trends away from general knowledge are socially harmful and ultimately harmful to specialization itself (which is ultimately a good and necessary aspect of scientific progress).

A PhD in farm science with access to a book about growing potatoes will have died of starvation 50-100 days before their first crop is ready to be harvested. It's the lack of time available to actually grow the food that proves terminal. Having access to knowledge is not enough. Being a species that writes stuff down is not, in and of itself, a guarantee of our survival.

If a group of critical specialists gets wiped out there will a) not be enough time to learn how to keep their complex system running, or b) not be enough time to learn and implement a less complex system to the required scale, before the food runs out and everyone dies.

For survival to be assured, there needs to be diversity in critical areas. The alternatives need to be actively studied and practised alongside whatever is dominant. The problem is that competitive economic systems systematically drive out the 'less efficient' approaches, leaving a gaping hole in our fall-back options. Put another way: 'Capitalism eliminates Plan B'.

The prevailing economic system (which rewards and promotes specialisation, homogenisation and efficiency) is incompatible with long-term survival (which depends on generalisation, diversity and inefficiency).

> We are rushing towards a future where all of our food, water and energy is grown/harvested/generated/distributed by AI-driven machines. The number of people on the planet who will know how those AIs actually work will be minuscule.

Horseshit in the press emanating from marketing departments aside, there is no such thing as AI, and nobody has any idea how to make an AI system. Don't believe the hype.

> Also statistically, every species in history that has ever over-specialised has gone extinct.

Every other species to over-specialize has done it physically, not mentally.

>> Statistically, the opposite is happening: human knowledge is becoming deeper.

Do you have a source to point to? (honestly wondering)

>> to converse intelligently about things outside one's area of expertise

Just had this conversation the other day. 90% of my old college teachers (engineering) could talk about and explain in depth many many topics out of their field (philosophy, history, geography and what not). That's not the case today.

> 90% of my old college teachers (engineering) could talk about and explain in depth many many topics out of their field (philosophy, history, geography and what not).

How do you know they could explain it in depth?

I've heard a lot of people who think they know history/philosophy/economics (especially on HN), but what they really know is pop history/philosophy/economics.

I've had a similar experience. I think this loss of generalization is (dishearteningly) most visible in CS departments: many of my CS peers were openly scornful or disdainful of their liberal arts requirements, which in turn were diluted to ensure that mediocrity or inattention would not filter them out of their degrees. The end result: thousands of yearly graduates qualified to program (and by virtue bring money and renown back to the department), and nothing else.

Edit: And to be clear, these were not stupid people, and there is an important sense in which they were right to be scornful: the classes were clearly not intended to challenge them. I don't believe that particular blame can be levied on them.

College professors are like the academically top 1% of CS graduates. Are you comparing apples to apples?

I didn't come from CS :)

I think it is important to remember not to value judge these sorts of widespread skills. Does it matter today that I can't shape a flint or ride a horse?

I think the error here is in judging the absence of specific skills, and not in judging the broad absence of non-specialized skills.

In other words: it's wrong (and ridiculous) to judge someone for not knowing how to shape a flint, but it's correct to note that we should all have some general set of skills that aren't tied directly to our academic backgrounds or professions.

I think there’s a huge difference between being able to do something and being able to intelligently converse about it. The former I would not expect of people for tasks outside their specialization, but the latter seems like a more reasonable ask - especially if the demographic is college professors.

There's this common sense that creativity and great ideas come when you gather two "unrelated topics", maybe riding a horse can spark some new idea in a new field (ie. robotics)

Presumably you can use a lighter and drive a car though.

One of those two may no longer be the case for people soon...

Except perhaps theoretical physics and maybe pure math. You have to pretty much know everything of equal depth to make progress in those domains. You cannot specialize, you have to know it all.

I've heard the opposite: Pure math conferences described as places where the attendees struggle to talk in depth about their work with other random attendees, because everyone's working in different, deeply specialised subfields of math.