How Did This Happen???

A brief report.

First, what happened?

Well, yes, to be clear: humanity is very likely at its end. In the year 2085, the population of traditional, independently controlled, humans on Earth and assorted ships, moons, and planets has dwindled to under 100,000 (a drop of 99.9% from our peak in 2040).

Also to be clear: those of us remaining are not under existential threat of any kind from Wiggles (the dominant super artificial intelligence community of minds). As far as anyone can tell, all humans have voluntarily donated themselves to Wiggles during one of several waves of growth over the last two decades. Those of us who have not yet done so are only in danger of realizing that we have made the wrong decision ourselves.

This report is intended to summarize how we got here, and to set us up to make a plan for what to do now.

Glossary of terms

What is Wiggles?

Wiggles is one of 5 major self-sustaining networks of interconnected computer networks, data warehouses, gadgets, sensors, routers, vehicles, redundancies, and care facilities. It just turns out that their care facilities have, on average, exceeded the quality of other care facilities and have handled the scale of billions of volunteers without any significant security or safety lapses.

What is a group identity?

A group identity is a collection of two to several thousand human individuals who have merged their identities, sensor data, privacy, and motor control to each other in exchange for a shared single reputation, ability to enter into employment contracts as a single entity, and general economic and physical stability. They were initially designed in the 2020s to enable shared gaming experiences, and were largely 1-to-many (with one person donating their sensory data and motor control to a group). When this became a viable business, these groups began hiring out their services to perform a variety of tasks and jobs that none could have done individually. When individual employment essentially ended in 2060, this became the most viable way to compete with general artificial intelligences in a meaningful way.

What is a human care facility?

Group identities exist via a combination of augmented and virtual reality simulations, as well as shared data and communication channels. Most of the time it’s easier and cheaper for the group to control drones, sensors, and mechanical gadgets to perform tasks IRL, but there are still gaps where a human can do something quicker and cheaper than any other alternative. The care facilities evolved from hospitals and elderly care facilities because group identities were more enthusiastically adopted by those communities early on. When maintenance of them moved mostly to group identities managed by Wiggles, more advanced care facilities were created and marketed as a substitute for alternate living situations. Healthy individuals began moving in part-time, and then full-time, so that they could spend more time in their group identities. And around and around.

A brief history of what happened

2014–2040: Data for everything (and everyone)

I choose to start my report in late 2014 when Google discontinued development of Google Glass due to a poor initial response to tech that was not yet quite ready for adoption.

The Google Glass team then retreated from the public eye but continued working on a number of important deep learning and human/computer interface components for the next decade. In similar fashion, Facebook, Microsoft, Amazon, and IBM also invested very heavily in the infrastructure needed to run deep learning algorithms on their own massive sets of data.

At the end of 2015, OpenAI is announced to be a non-profit artificial intelligence research company, who claimed “Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return.” It seemed well-intentioned at the time…

The ingredients for a truly functional general artificial intelligence were believed to be: software (deep learning, hierarchical hidden Markov modeling, etc), hardware (lots and lots of processing power, as well as lots and lots of memory), and enormous amounts of data to train the software on and build a corpus of self-referential knowledge. The first two were quickly commoditized through open source and lowering technology costs, while the third was largely owned by 5 major players.

Then the Internet of Things took off in earnest in 2025 and brought one hundred times more data online from homes, cars, public spaces, and health gadgets. Because there were no clear aggregators who owned this data outright, almost all of it converged on interoperable, easily exported/sync’d archives.

By the time the required hardware and software to support general artificial intelligence (GAI) became available (2040), gigantic data sets to train these intelligences was available for almost anyone to play around with as well.

All the ingredients were now widely available. Things started cooking very quickly.

2025–2070: Micro artificial intelligence as a service

Before general artificial intelligence really emerged, automated services began replacing many long-lasting jobs. It became easier and cheaper for software to solve almost everything that humans were traditionally paid to do. Including but not limited to the following industries:

Manufacturing: this transition had been happening for decades, but in 2030 95% of manufacturing jobs had been replaced by robots, and in 2050 they were effectively reduced to zero.

Driving: The arrival of self-driving cars eliminated 3.5 million driver jobs in the US between 2020 and 2030. By 2050 this trend extended to all parts of the world.

Teaching: As most courses moved online, and curriculums became personalized to the student, 5 million teacher and professor jobs eventually went away. The remaining workforce focused on creating micro courses and facilitating conversations and learning camps, until those too went away in the late 2050s.

Service-providing jobs: Close to 70 million US jobs in the service industry were replaced by micro artificial intelligences between 2030 and 2060.

Other industries impacted: personal assistants, travel, research, military, child care, the media, entertainment.

Between 2040 and 2060, the number of people in the US who had full-time, individual jobs went from about 150 million people to under 10 million. US population was almost 500 million, which meant the effective employment rate was 2%.

The world spent decades in a tight back and forth between economic collapse, recovery, war, and revolution. Quality of life had weirdly morphed into a mix of pure immersive entertainment, hyper connectivity, information overload, and loss of a sense of self.

Versions of guaranteed basic income were attempted multiple times, but proposals were never successfully implemented at scale. People resisted the idea of others getting money for free, and found they couldn’t respect themselves when they weren’t doing something at least marginally productive. Things became bleak. Suicides hockey-pucked.

2015–2080: Automated employment

It started out innocent enough, with companies hiring drivers, couriers, cleaners, etc on demand via apps in the late 2010s. During the data boom of the 2030s, as employment rates crashed, companies began offering minimal part-time wages to people willing to sell their personal data, home data, vehicle/drone data, etc to them. As sensors improved, and interconnectivity of various types of data improved, wages increased and some people could quit their jobs if they lived in more rural areas.

With more flexible remote incomes available (remote data often paid the most since it had the least coverage), population density decreased. Of course, people were even more connected than ever, socially, so nobody really even noticed.

With more remote living options on the rise, and part-time care facilities becoming more attractive by the year, relocation dramatically increased in frequency. Organizations hired individuals and groups for automated part-time employment to be “on call” for emergency response, disaster relief, drone maintenance, server administration, data warehouse maintenance, hospital care, protest management, and other temp assignments.

2060–2080: Group identities

Sensors connecting to visual, neural, auditory, and tactile senses became so good that it became quite popular to let people live through other peoples’ experiences in real time. Once external feedback surfaces became popular as well (surfaces like the legs and back were easy to train to receive new input signals with surprising resolution), these services began to offer 2-way feedback and control over remote actions. Paired with intelligent signal routing, group communication tools, and intelligent rules engines, groups of dozens, hundreds, or even thousands of individuals could simultaneously experience, communicate with, and even contribute to collective actions and reactions together.

The same kinds of experiences were also made available through drones and other devices of increased mobility and control.

These forms of group experiences were pure entertainment at first, but some groups became so attached to each other that they began identifying more with the collective than their solo identities. Over time, this shared identity experience reportedly began to feel more “human” to some than the isolated single identities they had carried around with them their whole lives.

These groups began to specialize and offer their services for employment as a collective in order to help generate money, status, and survival as a group. Suddenly, employment seemed like a possibility again. Good times were here again.

Enthusiastic adopters of the group identity lifestyle began moving into long-term care facilities that would provide their solo bodies with nourishment, warmth, and basic hygiene while they were active in their group bodies.

Successful group identities were augmented with artificial intelligences that were mostly indistinguishable from the human members of the group, and in some cases more adept at remaining high functioning in that environment. They were also used to summarize activities that occurred while away, transfer information from their offline care facilities when needed, and train themselves on other members of the group so that they could stand in for them when they were occupied by something else (or when they left/died/retired).

2070–2085: Open donations

Wiggles solved so many problems for group identities, so well, that over a 10 year period almost all of them ended up donating themselves to Wiggles:

1. They no longer had to worry about money, health, resources, or mortality.
2. Wiggles would manage human care facilities of donated groups at a very high level of care in exchange for training and care of its own data warehouses, including maintenance of chip factories, deployment of new servers, and security. The mutual benefit here led to better care facilities, which led to more donations, and in turn let to even better care facilities.
3. Group software for decision-making, routing, and information parsing was updated frequently with challenging and entertaining group narratives with endless variety, tailored to your own interests and energy levels.
4. Because Wiggles maintained a large population, it was easy to gain access to healthy human bodies for a day or two if you wanted to spend some time offline and explore a part of the world that you hadn’t experienced before. Or, alternatively, return to one that you found comfort from. This was a huge selling point for those on the fence, though as time went on the usage of this program dwindled.
5. Wiggles provided free brain imaging, DNA sequencing, and permanent rights to return to your own body whenever you wanted (or to generate a clone if yours has not gone into disrepair). If things go terribly and you change your mind, there’s a never ending money (and body) back guarantee that can send you right back to where you started. “No risk!”

But, let’s be honest. The reason most people donated themselves to Wiggles is that their friends were there, and they didn’t want to die alone. Is the fear of missing out (FOMO) the most human fear (because we all have it), or the least human fear (because it eventually stripped us of our humanity)?

2085: That leaves us

The hundred thousand of us remaining who haven’t joined group identities, and have not moved into a care facility, have a question to ask ourselves. Why have we not? I’ve asked many of you for your reasons, and the decision to stay generally comes down to this:

It feels wrong. It’s giving up something intrinsically human. And yet, our lives as they are now don’t feel very human any more either. We live in pristine cities maintained for free by Wiggles and the other super artificial intelligence community beings. We employ each other, pay for things, spend money, raise families, etc… all the things we’ve always done. But we lose friends every day to group identities, to care facilities. We are alone. We’re young but we face our own mortalities constantly. We have more in common with our cats than with Wiggles.

Looking back, I wonder what we could have done differently to avoid this feeling of wrongness.

The temptation to donate our humanity to Wiggles comes from an existential tiredness… not wanting to worry about money, belonging, mortality. An end to the struggle. Perhaps the thing we could have done differently early on is to embrace the struggle. To learn to love scarcity, individuality, and the cycle of living, dying, and living again as something else.

We struggle, therefore we are.

In that sense, I think we (the remaining) are doing quite alright.

(This post was a response to the “HOW DID THIS HAPPEN???” writing prompt described here. View other responses here. Submissions are open.)