No Recent Automation Revolution

> wind up with a two-year career. You would work for about two years and that pays for the rest of your life, you can retire or do whatever.

True expertise in fields like medicine, engineering etc. generally takes at least a decade of experience to acquire. Societies need for, and ability to utilize, such highly skilled labor is only increasing as we automate simpler tasks.

People who never aspire to such professions might be spared long working years, but some portion of people will need to put in the long hours and years to keep the lights on.

> I want an economy where we can celebrate automation and distribute the profits evenly.

Distribution of profits evenly: This is called "company stocks".

I think there's an intrinsic human drive to be useful. It makes sense evolutionarily: the tribe can't kick you out if they rely on you for something. It's no guarantee of safety or success, but it's a bargaining chip.

Kevin Simler (via Dennett) goes as far as modeling this phenomenon on the level of individual neurons [0]: that conditions like depression and grief can be thought of as neurons and their networks becoming suddenly obsolete, and casting about trying to find new ways to be useful.

The employment model is hardly perfect; if anything, it has an uncomfortably feudal flavor in practice, where one is a thrall/vassal for 40(+) hours/week, in exchange for autonomy the other 128(-); and where one has the option to seek a new lord if sufficiently dissatisfied. There are countless other ways to be/feel useful outside that model: entrepreneurship, non-profit/community work, raising children, passion projects.

There are also potential perverse incentives: in particular, to maximize the problem one purports to solve. The most obvious parallels are the military- and prison-industrial complexes; planned obsolescence is another milder example. It's easy for unions to tip from collective bargaining into protectionism of industries that we'd like to become obsolete (prison guards, coal miners), out of a very understandable desire for those individuals to not lose their social bargaining chips.

So I do think we should celebrate the idea of "employment", as such, becoming optional; I strongly favor UBI. But at the same time, we need to be thinking about new institutions and proactive strategies for people to be, and feel, useful. I dare say, in some cases we want to be spending the economic surplus from the efficiency of automation, on intentionally inefficient pro-social economic activity. One example might be a community garden: it's a safe bet that industrial agriculture will always be an order of magnitude more efficient at producing food; but the community garden has other benefits, of connecting humans to nature and to each other (and perhaps a bit of resilience against supply chain disruption).

[0] https://meltingasphalt.com/neurons-gone-wild/

Administration has always been hit very hard. People don’t remember, but in the 1950s offices regularly had dozens of people who’s only job was calculating things and drafting charts and reports. In some domains they were literally called “computers”, since “compute” has a meaning prior to the electronic computers we now know.

All of these jobs were replaced by a spreadsheet decades ago.

I have sort of a first hand account of what the current problem with (At least one area of automation) is.

Up until the quarantine I was working in a cowork space; and it happens that - I'm the sole employee of the company I work for in my country - all the desks around me were occupied by a company that makes a lot of their income providing automation to other parties.

I even worked for them for a few months as an "on demand" consultant; the idea would be that I would provide assistance to their devs when they needed help with the "programs" they were developing, and that fastly turned into a process review that eventually cooled off because they weren't - the project manager, at least - very keen on taking my advice.

What I observed was that - bear in mind this is a 3rd world country - the salaries were rather low, so obviously they couldn't get experienced people.

This lead to they having to train them, but as they had no experience, the way in which they approached solving the issues where, a lot of times, "complicated".

Usually they suffered from the apps being interfaced having changes along the project (Or after). This meant they had to fix a lot of problems that were not there in the beginning.

AS the devs had only experience with this tool, and was their first experience coding, they couldn't make the most of the other tools at hand. This mostly lends to processes that take hours to complete, being the execution time only marginally better than a person. This coupled with the issue right above this, many of the projects had a time excess of around 100% (Effectively taking double the time that was allotted in the beginning). Last I heard from them, thought, is that they made a lot of improvements here.

In one particular project I practically had to threat the person responsible to use javascript to fetch data from the browser instead of the "native" way of doung that (Basically, querying the page through the tool with xpath expression). When they did change to javascript, the process execution time was reduced by >90%

All in all, the problem I see with automation is that the people doing that is not formally trained in coding, and the tools I looked at have rather weak integrations with the most used apps today (Namelly, browsers).

I think the idea for many of these tools is to allow people to automate themselves, but that will lead in my opinion into the "phpization" of the automation area.

All in all, there are a few caveats: I know from them of at least one company that had a rather strong and professional dept., and they where _very_ good at it.

I think that eventually automation is going to take the world by storm in the next 5- 10 years. And those that adopt early will have a clear advantage over those that not.

For the record, the tool they use is Blue Prism; according to the CEO of this company BP has some sort of marked edge over UIPath when it comes to enterprise settings. I know there are other tools in the market, but I don't know them except by name.