Automated dream

Automation is a fun topic, especially when you’re working with people who run an automated cell in production. It’s a dream to manufacturers, a machine that can do a job that takes a guy half a day in an hour. You just feed the machine necessary parts and it spews out a ready product, or a product that’s ready to move forwards on the manufacturing process. You see these ideal videos and photos from car manufacturing plants, where robots are moving back and forth putting parts together. Everything seems to move like buttered lightning, and probably smelling like that too. The reality is a bit different though. Discussing these automated cells often brings up that there are few problems with the current mode of automation nobody really talks about, first being that it’s automated to a point. It’s always pointed out that there needs to be someone to oversee the robots working, telling them what to do. This is probably the best example of tool artificial intelligence we have, where these robots know to do one job they’re told to do and they do it well. At least as well as the parts allow. Because of manufacturing tolerances, the pieces these robots put together are often misaligned, have cut corners, warped pieces, arcing issues and all that. A human can work with these parts, because we’re sentient and aware of what the hell’s happening, but a robot’s intelligence carries only as far as its programming and tools. If there’s a gap because of a warped piece and its laser eye sees it, it’ll alarm the operator to either try again or skip the step.

Automation makes more efficient production, when applicable. Automation often also results in better results cheaper, but with much higher up-front costs. Setting up automation, be it a manufacturing robot or to build a 3D model to be used in live streams, the front costs is high. For the robot it’s the whole shebang from buying the robot, remodelling the place where the robot needs to be, setting up proper power grid for it, building the spot, installing the robot, then realise you need more than one, get a guy who knows how to program the robots, proceed to do test runs and continue to improve the programs and methods in the production line to get satisfactory results. In short term, the price of an automated cell in a plant is high, but the long-term drop of costs is high. One of the few places companies often think about proper, sustaining long-term profit instead of destructive short-term. Similar thing can be applied to the 3D model, where a person has to acquire hardware and programs to start modelling, and probably learn how to model at some point. All that asks time and money. To use the model in a live stream, some kind of motion tracking hardware and software has to be utilised, and probably some other buzzing equipment. All that is high up-front cost before you are able to make profit from them, but after all that’s set up, it’s much easier and economic to change the programs or the 3D model. They’re also permanent. A robot doesn’t need to take a rest like a person needs to, and a 3D model doesn’t need to exercise or put on makeup to change its body or face. There’s an element removed from the equation that requires certain kind of physical work.

Just like how keyboards have automated writing. You no longer need to hold a pen and write something on paper. You don’t need to concern yourself with writing the letters properly. Writing the letters has been automated for you. Even spellchecking has been automated to a large degree, and it is only a matter of time, proper coding and programming before we have a tool AI that is able to properly write, say, a translation.

Automation is replacing some work people are doing, but the more automation is being refined, new robots are designed and implemented, the more coding and programs are refined, the more work will be replaced. One of the more currently relevant topic might be artificial intelligence doctors. I talked about this a bit previously, but the benefits seem to be up there with doctors and nurses that don’t get sick or get tired. IEEE Spectrum has tallied up AI vs human doctor accuracy, and while live doctors are winning in general diagnosis and photoshopped images, the rest is more up in the air. Even if the presentation is rather simple, perhaps too much, it does seem to point out that in general terms live doctors may be able to make better overall judgement calls, but when it comes to accurate, on-point diagnosis the AI has the leverage. Probably could’ve saved me from scarred lungs if these are anything to go by.

Lot of times automation has been said to replace low-skill jobs. Some of these probably are, but it appears that the word is used to describe work skills that are not attained in higher education. A welder, for example, may not have a university degree, but his knowledge and skill set has to do with material studies, mechanics, physics, little bit of chemistry with work that needs constant attention and loads of craftsmanship. Anyone can be a shit welder, just like anyone can be a shit journalist. To be a competent welder takes time and effort, and acquisition of skills most of the population don’t even know make up most of their surroundings. While I’m being on this tangent, most of our modern world is build on welding. From the buildings we live in to the chairs we sit on, from the cars we drive to phones we talk on, there are bits and bobs welded together everywhere.

It’s mostly a matter of time until automation creeps itself up the the ladder to high-skill jobs. Technology may not be there yet to replace doctors, but it’s getting there slowly. Information has already been automated with the Internet, where most news sites and journalists working there have been obsoleted by individuals reporting on their own and taking footage that usually wasn’t available to all. Some time back slew of journalists were left jobless when the sites they were working for went bankrupt. The kind of service and content they were producing was replaced by the automation of information via the Internet and the people using it. That’s automation at its core; something that makes it easier to put out much cheaper and more efficiently. All the video hosting sites like Youtube, and all the blog platforms like WordPress, are part of automation of information, where we have seen the loss of extra hands in the middle. Even with the most of the platforms and publishers do control information to some extent, it is mostly possible to get unfiltered, uncontrolled information if you’re willing to do some digging. For example, footage on how different parts of the world are dealing with SARS-CoV-19 and COVID-19 often paint a different kind of picture from what news sources may be giving you. Being able to read media properly plays a large role in this, as individuals have about as much agendas as any news source would. Sometimes it’s to push a political view, sometimes an individual just wants that particular moment out there. Automation of information has also given individuals the possibility to work to the same extent as any professional journalist, and this has clearly caused friction. What constitutes as legitimate news and does a source need to be confirmed by an outside agency, like a government, have all been raised to the table.

The more automation proceeds, the more questions are raised and the more it is being questioned. People who didn’t expect automation to enter their work field have hard time to adjust to the reality that they may need to acquire new set of skills in a world where their skills have less demand. Learning to code is one thing, another would be the person who oversees the automation.

No automation is truly independent of human interaction. Automation is nice and all as long as it doesn’t break down. All the errors manufacturing robots face and can’t be solved needs that human operator to step in and fix it. At least at this point in time, who knows what the future holds for automation and robotics. At some point, we might have robots building robots that build robots that fix the stuff the first robot can’t solve. It all ends with the nature of the work changing. We’ll always need people who know how to weld, because automation can’t be taken to the field and can’t be fit into each and every pocket. That one guy with a rod of metal fixing your car’s busted door has to work with whatever the hell you did to the car and no robot can really fix it due to the sheer amount of variables. Even when automation is taking more hold, there are niches in its wake that people will fill. They may be small niches, but at the same time, automation opens other doors of possibilities. It’s up the people to grab them. Automation won’t stop as long as consumers want their stuff fast and cheap.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.