knockoff nige

I just read an article about Uber partnering with Fiat Chrysler to build self driving cars. Obviously this has been a Google experiment for sometime and other companies are looking into it.

This is getting a little more worrisome though, as we seem to be moving much faster towards automation replacing the human in the workplace. I work in IT and can appreciate the benefit that comes from automation and can see how some people become almost useless to a company due to their lack of upskilling. But, even in IT there is still plenty of room for anyone to learn scripting and get a job with it. Right now, the people left behind are those not willing to learn new skills. I interviewed quite a few people towards the end of last year for a role but was surprised just how out of date some people are skill wise.

But, with corporations lobbying for tax cuts (and often getting them in some form), this tax hand out can too easily get put towards replacing a person with a machine. Technically it is investing in their business but at the cost of jobs. Maybe there should be a signed agreement on how this investment gets spent. That's for another topic. Are any of you concerned about the possibility of machines taking your place or are you excited about the potential of technology doing the tasks we don't want to do and doing them better than we ever can? Also, what do you see as the next big thing in automation?

Interestingly, Sweden recently voted against a guaranteed minimum wage for all citizens which would make way for machines to do most of the work.

GarryP

Beoz and me have been having a back and forth on another thread. He keeps saying "something will turn up to replace the automated jobs", and I keep pointing to 10s to 100s of thousands of jobs that look to be on the short term chopping block and saying things like "there will probably be a million unemployed by the next election", it's going to happen fast.

At the same time I've been researching 'deep learning' and getting my head around what it can, and cannot, do. Rough rule of thumb - if you have a process defined job (eg what you do is written in a binder) and particularly if your job inputs and eventual outputs have been captured in a database; worry. It is relatively easy to throw the algorithms at the data and see what sticks. If it does, they can easily sack large numbers of people in small amounts of time, certainly faster than it would take a politician to realise what's happening.

If you make a list of the jobs that look to have serious work going on to automate them, and the number of people currently employed in those jobs, you end up with some big numbers.

Also, automating things like driving fundamentally rewires society over a timeframe of 10-15 years.

Sure you don't mean swiss?

Swiss Voters Reject Guaranteed-Income Proposal : The Two-Way : NPR

I don't think we are going to get guaranteed minimum incomes past the Beoz et al of this world - which means we might well end up with riots and bifurcation of society. By the time they realise what's happening, it'll be too late.

knockoff nige

Yes I meant Switzerland.

I don't think it matters what your job is as its all on the line. Many things will be packaged up and sold in modular form such as cloud services. I would say it would be easy to replace tax accountants with a 'click-the-box' website and no need for an accountant that can be 'creative'. Just tick the 'milk it for me' box.

Beoz

Yes Garry and I differ on this.

I think you may have raised a very good point in your first post when you said people are unwilling to change and re-skill. That is going to be the key. As you said there's plenty of jobs and places available where skills are lacking. And us humans and our unwillingness to re-skill, upskill and change will slow automation adoption.

I work for a tech company so I see this stuff coming and going all the time. You mention the cloud. We have a server based product that typically has always been built on a server in someones office somewhere. This requires our techo's to configure it, a local guy to manage it, someone to create the hardware, etc. This is all moved or moving to the cloud now. The benefit for the client is that someone else provides and maintains the hardware BUT because it becomes more efficient to roll out and less hassle to grow the only men down is the people who provide the hardware (if they were smart they'd shift the building and selling to data centres).

So some of our techo's were worried they might not be needed. The opposite has happened. The "packaged cloud" comes packaged with techo's who need to constantly expand the service for the clients because it is so easy to expand it now. We can't hire enough. Clients are also hiring more techo's at the end to do the same.

So same with Ubers cars. Someone will have to write the software, and do all the other bits that go into driverless cars. Most of those companies don't exist yet.

So that's the lesson history has been telling us. The blacksmith went off and became a mechanic.

Automation will only grow as fast as we humans can upskill, adapt and change.

Oh by the way. How many jobs would the world NOT have today if the World Wide Web wasn't invented? ..... billions

Automation and tech creates work.

GarryP

There are a bunch of driving forces.

• if you are a McDs burger flipper then there's a lot of you, all round the world. If McDs can come up with a $35k machine that does your job, then it become a no brainer for the franchisees and very quickly all the burger flippers are replaced. That's one type of automation replacement - minimum wage jobs with lack of skill, but price sensitive.
• if you are a driver then there is a real complex skill there, but there's also a lot of you, and once again, once one machine that does your job exists, the economics say that lots of the job disappear quickly (actually it might well be only a few $k to replace them). These are skilled jobs, but still price sensitive.
• if you are a postie, and are currently doing a slow, cr*p, job, then there's lots of benefit for an Amazon to disrupt that industry, and in the process pull the rug from the whole lot. Even if it can't deliver 10kg packages, with no 2-5kg packages the economics kill the existing jobs. This is unskilled, but badly performed jobs where there is a demand for much better performance.
• if you are an office process bound paper pusher - let's say 'insurance broker', then not only is the process by which the job is supposed to be done defined, but also any 'flair' can be copied and automated via deep learning. Your program replacement will get trained up under a contract you don't even know exists, and once it's better than you, you get sacked. Maybe 1 in 10 are kept on to oversee and deal with the complex cases - but you won't even see it coming. This is skilled, but process bound, where training is easy on existing data.

What's left are the jobs where real innovative thought is required (not process bound), where physical input is key (eg massage), and where price is not an issue (err, I'm sure they exist).

The key thing is once one machine exists that can do your job, the whole global avalanche happens very fast. There will be a 'fashion' that pops up, with MBAs looking to look good by automating whole job families out of existence (cf outsourcing).

Oh, and another nice little thing. You see how politicians scream about multinational tech companies "not paying tax" - well what happens when the 'insurance broker' job is a deep learning algorithm sitting on the cloud? You can bet it won't be on a cloud in Australia, it will be in a tax haven. It won't just be web advertising revenue that's not taxed in australia, real jobs performed by virtual employees will go untaxed and shopping for 0% tax havens.

Amazulu

Aristotle discussed technological unemployment

'nuff said

Beoz

Microwaves, or the 3 of them which seems to be the done thing in McD's, don't cost 35k.

Already happened.

Moses2013

I have to agree with what Beoz said: Someone will have to write the software, and do all the other bits that go into driverless cars. Most of those companies don't exist yet.


And who will buy these cars if all jobs are gone??????


I would be more worried about population control, climate change etc.
Some jobs will go and others will come and as knockoff nige said "the people left behind are those not willing to learn new skills"


Well, it's already happening that many people have forgotten how to grow fruit & veg.

GarryP

Err, not quite.

Take Amazon drone delivery as an example. Currently they are at least 1000 people strong in Amazon, working to nail down the patents on drone deliveries (Google are doing similar). Basically its one of the techniques they use to not pay tax - better to employ R&D effort and take the tax advantages than give the government money.

Eventually they will crack it.

What happens then?

Well, that workforce will wind down. The automated delivery warehouses will probably employ less than the existing warehouses (it's not just the delivery drones that's automated) and with the patents sewn up, they can litigate competitors out of the market.

Maybe they will shift to some other market, but you shouldn't really expect those 1000 people to increase in number - that's that why the economics works.

Good question. Maybe a politician should be thinking about that now? You can be certain that the multinationals won't be. Not their concern, they think.

Nope. Kinda by definition, it's the jobs where "you are learning new skills for new jobs" type of roles that won't get created - they will be automated from the get go.

This automation works because less people are employed. If not, it's not worth going to the effort to create the automation in the first place.

GarryP

I was actually talking about most of the tasks in a fast food 'restaurant' getting automated, not just the burger cooking. They have them as pilots already - from front of house to cooking and delivery. And the $35k came from the price of the robots involved, as reported.

Walk up, with the McD app on your phone, with your favourite already selected. Tap and pay via the phone, and have your order delivered 10 seconds later via a chute.

Oh, and you notice that DVD shops died a death (and them only with one human involved) but the DVD rental machines are doing great business in seemingly every shopping mall....

Moses2013

Well if one warehouse shuts down, another will open elsewhere. You'll also need more people to collect the drones when they break down and you'll have more legal jobs (drone laws etc.)


Automation still only works if there's demand for a product, but you still need consumers to purchase products. I agree that it all looks like doom & gloom, but that was the same when DVD replaced VHS and now it's digital.

Moses2013

Exactly and automation does not mean less jobs. Funny enough, McD are creating even more jobs and somebody also has to build the new restaurants too and where does the food come from? More jobs in farming etc.

Beoz

There you go Amazon have employed a 1000 extra people to work on drone stuff. Automation creates jobs.

BTW the patenting bit, you can only patent portions. Not the whole idea. They may need to develop some software which manages the camera on the drone. The only thing they can patent there is the software, not how the GUI looks, not how it operates, but the code. Just means thr next person needs to write the next competitive software with different code that is not line for line the same.

And you say Automation is there to kill off human involvement and make it cheaper. A lot of Automation costs more than a few humans to implement, and maintain. Most of this stuff will follow a subscription model. Year on year costs.

Generally Automation will be used to increase supply and quality which subsequently will make the dosh. That's why McDonalds don't flip burgers anymore. It takes too long. They use thr Microwave to increase the volume of burgers they can pass over the counter.

By the way, that app you use to order the burger, it doesn't invent itself.

GarryP

Nah. Stop and stand back. In particular, thing of one time and recurring costs.

If you are OKing the business case for throwing out all the burger flippers and making the restaurants automated, you are going to want to see that the bottom line is better, in an obvious way. You look at the cost of making those restaurants automated (a 1 time cost), the maintenance (ongoing), the loss/gain in customers (from a pilot or two), and the change in wages and personnel costs (ongoing). It has to make economic sense if it's to happen.

Upshot is it will have to noticeably save money if they are to go ahead.

Which, practically, either means the land taken up is less (unlikely), the materials used are less (nope), or more likely, the human resources bill is much lower. So, in net terms there will not be more jobs, rather there will be many less.

A few facts. There are 760,000 FTE McD employees worldwide, or 1.9m people (part time most) - 90,000 in Australia alone. There are 36,000 restaurants globally (which means ~53 people per restaurant) In Australia they spend $40m annually just on training the burger flippers.

Here's what the cost structure looks like for a typical fast food franchisee :



So corporate and materials rakes off a big chunk, but wages and then other associated costs (other will include personnel items) is the next biggest cost. Look at the that profit line. Now imagine what happens if 53 people is reduced to 10...

As for more jobs in farming, nope, not a chance. In fact I'd probably bet that even now McDs are researching lab grown meat.

Amazulu

You can't have the mass production of consumer products and services if there are no consumers

Ain't gonna happen