rok
2d ago
I didn't recognize "AIAYN" as such, but as soon as you mention attention or transformers...
I think anyone in math, sciences, or Econ probably knows that paper at this point. For CS, it is definitely up there with PageRank in terms of influence. Maybe one could argue that it is less influential than the fast Fourier transform because subsequent research has shown that other model architectures also achieve the ability to process natural language when scaled up into the hundreds of millions of parameters, but in the long run its social changes are going to be bigger than FFT.
Oh this is fun. I agree with your assessment: Africa will win any battles on their technological and knowledge advantages (decryption, radar!) until Europe crushes them in logistics and develops/reverse engineers tech. However, any European occuparion of modern Africa will fail for the same reason colonialism failed: malaria
One question would be how to motivate the conflict. Let's say a portal suddenly opens up across the Med, from the strait of Gibraltar to Suez.
-
European powers lose contact with their African colonies and send out scouting missions.
-
Several of these get wiped out after they appear on primary radar.
-
Africa, receiving European radio over the strait and communicating mostly by cell phone, becomes situationally aware first. Unites in a blaze of anticolonialism (ok, not realistic, but needed for the military competition to be anywhere near interesting) and decides to raid Moussellini's Spain while fortifying land chokepoints.
-
A few right-wing Europeans in modern Africa (Ceuta?) quickly defect with laptops and cell phones. They tell European leaders about the last 80 years of tech, but it takes time for them to be believed, news takes time to propagate and 1939 Europeans don't have the knowledge or industrial base to replicate modern tech.
-
Africa promptly genocides all the modern Europeans left on the continent, selectively broadcasts anticolonial history into Europe.
-
Europe slowly unites in response. Again, this is needed more for the military stoey than realism.
-
Europe plans a commando raid on university libraries to get engineering knowledge.
-
The raid is foiled by African radar and antiship missiles as well as by coordination over cell phones (Africa faces no fog of war), but it proves that the defectors were telling the truth.
-
A bunch of African refugees decide to try to make the trip to 1939 Europe (perhaps they were already on the way and were just confused when they lost cell signal).
-
They discover that eastern Europe is much more poor than they expected: rural communities lack electricity and are still cooking on wood stoves, smoking meat in the chimney and putting tubers into root cellars for winter.
-
Not sure how the Russians, Hungarians, or Poles handle African refugees. Might be with communist atrocity, rich Hungarian globalism, or beliefs based on Christian scripture.
-
Europe gears up industrial engines for war.
-
The rest of the story is like World War Z: fun if you are into slow-rolling logistical tales or HFY, but the conclusion is pretty much a given.
-
Surprise ending: in the ground campaign Europe conquers north Africa, but faces 50%+ losses when they reach sub-Saharan Africa due to the continued existence of malaria.
Edit: Actually, smallpox in 2026 Africa is going to wipe out 20% of military age men, although it will take some time. Covid is going to wipe out 0.5% of Europe, but Europe in 1939 was younger, so it is not a strategic concern for them.
This seems like one of those proposals of the form "We tried policy A and the problem got worse. This must be due to the external reason B, instead of the failings of the policy. We need more policy A!" It might make sense, but given the policy's failure one really needs to explain the observations within a mechanistic framework to be convincing. Those who advocated for policy A without predicting the failure should lose credibility, but since when has a politician or pundit ever been held accountable for poor policy?
In living memory, the only policy I can think of politicians and pundits eating crow on has been the legalization of online gambling, and that only happened because the pundits promoting gambling take pride in their ability to predict real-world phenomena rather than the partisan battles.
So without a mechanistic model which makes a contrary prediction, it seems illogical to say that the solution to sexual assault is "more intersexual mixing", especially when for the case in question the woman entered the male space of her own accord, asked the men by text message if she could do drugs off their genitals, and asked for it to happen again afterward. There are a bunch of different ways you could make that sequence of events less probable, but further integrating mens' and womens' private spaces is not one of them.
As another example of a policy that makes no sense to me, we might consider sex integrated units in the military. These units have high sexual assault and pregnancy rates. And yet it seems to be the tribe which is ostensibly pro-woman which both pushes for more integrated units and is horrified by the sexual assault. You put young men and young women together in a high stress environment with a strong social heirarchy and what do you expect to happen? EvoPsych predicts women fall for their superior officers.
My instinct says the best way to reduce the sexual assault rate at colleges would be to revert back to single-sex dorms without visitation. This provides the college perfect legal cover, and any woman who enters the male dorm is certainly consenting to something. College sexual relations will still happen, but at some point people do make decisions themselves and do consent. (There are a lot of stories from the 1970s and 1980s about lines of women from the local women's college entering and leaving the men's college dormitory via the basement windows on weekends.)
More options
Context Copy link