Hacker News Reader

So for the example of an actuary, ideally their job would also be automated by an advanced program...If we took every actuary working 9 to 5, and instead threw all of their tens of thousands of daily working hours at coding a program to do their job...

What, exactly, do you think actuaries do during their 9-5? Do you think they manually compute actuarial tables with a TI-89?

Actuary here: only 5% of billable-hours go into the concept of the math and probability behind setting appropriate price.

95% of the work involves translating the messiness of real-world into the model-based world. This is the type of thing that's obliviously difficult to automate: e.g. more difficult to build a robot-nurse than a robot-surgeon.

Yes, as I was following this subthread, I imagined that an actuary's real job was to recognize and model things to be insured or otherwise predicted. The math is just the executive summary.

As a software engineer, you probably imagine that I spend most of my time programming.

I personally see a lot of potential for automation, -just tonight I was out working for a non-profit that saves a lot of hassle because I learned them a few Excel tricks back in 2012.

Another place I was paid for less than 20 hours and it shaved two hours of preparations for the run of a machine, meaning they could keep it for several years more.

Actuaries are exceptionally well trained professionals who use incredibly sophisticated automation. They do not need help with Excel macros.

Farming isn't a bullshit job, at least not if you love it. The same could be said of everything.

What society, and by secondary intent technology should aim to do is to enable everyone to be capable of performing the job for which they have passion for.

My parents are farmers, and love it. I love it too. It is far from a bullshit job. But for the average joe slaving away on a rice paddy in China, they may dream of something grander than the opportunities afforded them.

I have been fortunate enough to have more education and opportunities than even 0.001% of the population could wish for - excellent schools, 3 bachelors degrees, 10 years at uni, mostly having the time of my life. I'm a doctor performing a job I find fulfilling, and I have a startup that grew out of a hobby. And it's profitable. Society should aim to give everyone the opportunities afforded to me. Because then there would be no bullshit jobs.

Yes, but we have to automate all the "bullshit jobs" before that can happen. Otherwise people who don't want to be doing them get stuck doing them. "Bullshit job" isn't a judgement against people who enjoy something, it is a judgement against the system that forces people to do "work", even when they aren't interested in it and it doesn't really contribute much to society.

> "Bullshit job" isn't a judgement against people [...] it is a judgement against the system that forces people to do "work"[...] and it doesn't really contribute much to society.

I know this is probably not what you've meant, but: I do think that this can be seen as pretty judgemental. Contribution (without parameters/scale) to society (without further specification) feels very subjective to me.

You're right, that sounded pretty pretentious and high-horsey. I just meant that if a job can be automated cheaply, then we all generally benefit from automating it, assuming the people previously doing it are cared for.

> If we took every actuary working 9 to 5, and instead threw all of their tens of thousands of daily working hours at coding a program to do their job, how soon do you think we could automate an entire industry?

This is naive techno-optimism.[1]

[1] https://xkcd.com/1425; "The title text mentions The Summer Vision Project and Marvin Minsky of MIT. In the summer of 1966, he asked his undergraduate student Gerald Jay Sussman to 'spend the summer linking a camera to a computer and getting the computer to describe what it saw' ([1]).... The project schedule allocated one summer for the completion of this task. The required time was obviously significantly underestimated, since dozens of research groups around the world are still working on this topic today."

The DARPA Grand Challenge started in 2004 [1], and now we have autonomous vehicles on the road in select locations. While still not fully unmanned, we've gone from nothing to a great deal of automation in ~11 years.

Spacex was founded in 2002 [2], and is expected to have a successful first stage powered vehicle recovery this year. 13 years from incorporation to revolution.

IBM's Watson diagnoses cancer better than a human doctor [3].

Keep you skepticism sir. I'll bet on the future; its winning.

[1] http://en.wikipedia.org/wiki/DARPA_Grand_Challenge

[2] http://en.wikipedia.org/wiki/SpaceX

[3] http://www.wired.co.uk/news/archive/2013-02/11/ibm-watson-me...

The 80/20 rule will be in full force when it comes to driverless cars.

>When most people say "bullshit job", it seems to me like they really mean "job that could be solved by a well-built machine".

What Graeber seems to mean is jobs that serve the interests of some people at the expense of others and overall contribute very little, nothing, or actively work against the interests of society at large.

That is what Graeber aims to describe - he's even specifically written about how technology has undermined "meaningful" employment in agriculture and manufacturing - but the problem is his arguments against "bullshit jobs" conflate jobs which don't serve society with jobs whose employees feel they're worthless because they much preferred being poet-musicians. The latter argument, and Graeber's vision of "Hell is a collection of individuals who are spending the bulk of their time working on a task they don’t like and are not especially good at" are extremely susceptible to precisely the "lack of perspective' criticism raised by the OP. And unlike Graeber, I've met corporate lawyers who chose that line of work because they found it interesting and valuable (and I'd hazard the corporate lawyers that draft the worst contract terms and push the hardest for unnecessary litigation really enjoy their jobs)

Sitting in a corporate cubicle churning out CRUD apps on spec for customers you're isolated from probably feels like it's contributing less to society than co-founding a "change the world" startup that fails to achieve product market fit, but the former probably achieves more good for society, even if it's only the sort of good that involves people having to do less paperwork manually...

>he's even specifically written about how technology has undermined "meaningful" employment in agriculture and manufacturing

Wasn't how I read it. He mentioned that technology "freed up" people from working in the field, though, meaning they could then be put to work in bullshit jobs.

>his arguments against "bullshit jobs" conflate jobs which don't serve society with jobs whose employees feel they're worthless because they much preferred being poet-musicians

I derive a lot more benefit from poet-musicians than I do from corporate lawyers. Admittedly, corporate lawyers are never employed for my benefit, but I think was kind of the point.

One of the key differences between poet-musicians and lawyers is that if the former group do something that benefits you, you quite literally hear it. Corporate lawyers talking the board out of doing something damaging to the public because of potential legal ramifications do so behind closed doors. And lawyers for public companies are often quietly propping up the value of an awful lot of ordinary people's savings even when defending frankly indefensible acts. Similarly, the public is only indirectly picking up the tab for the production and distribution of unpopular music albums.

Shareholder return isn't an especially useful gauge of the redeeming features of sleazy corporate lawyers or artistic merit of moderately unsuccessful musicians either, but calculations of social benefit are a lot more complex than they first appear

>If we took every actuary working 9 to 5, and instead threw all of their tens of thousands of daily working hours at coding a program to do their job, how soon do you think we could automate an entire industry?

Not very soon? If that were possible, don't you think insurance companies would have long ago jumped at the chance to drastically reduce their labor costs?

>The moral of the story is we should automate jobs that people don't like doing.

As you say, you can see this happening today. This is already how the economy works. No company is going to pay workers if they can avoid it. So what's the problem?

> The moral of the story is we should automate jobs that people don't like doing.

who decides what jobs people dont like doing? Just because you might not like being a taxi driver, farmer, actuary, etc, doesnt mean you should demolish the job with automation -- plenty of people find these jobs provide a fulfilling workday and are passionate about them. any automation there would deprive them of doing them as a profession, which is somewhat sad.

ill take it a bit further: how would we feel if tomorrow, we woke up and some guys from stanford published new research on AI which can automagically produce working software given a set of requirements -- and they give it away for free. the role of software development is no longer required, since this AI can do the job in seconds without bugs (you know, assuming requirements can be written without bugs!). what then of software developers? our jobs no longer make economic sense. you'll need to find one of those jobs not solved by a computer.

Allow me to turn it back on you: what if your sentiment had prevailed at any time during our current revolution in automation? Say a group of factory workers protest automation not because it puts them out of a job, but because they like what they do (a totally unnecessary distinction - part of the reason I like my job is that it lets me eat and pay bills). Are you sympathetic to preventing automation then? Or is our current level of automation just right because it has presented you with the least amount of challenge to your job?

No, we should encourage automation full-tilt. Jobs are lost every day to automation, and the promise of a world where the workweek is short to nonexistent is one we should embrace, not reject because some people like their jobs. If we work hard enough to bring this to fruition, the actuary can crunch numbers on his own time if he wants -- or is free to explore thousands of other hobbies.

how would we feel if tomorrow, we woke up and some guys from stanford published new research on AI which can automagically produce working software given a set of requirements -- and they give it away for free. the role of software development is no longer required

It might be inconvenient for me in the short term, but the world would be much better off overall. Trying to forbid that technology so that I could keep my current job would be selfish and destructive.

I've programmed since i was age 11 and i have been waiting for that research to be published ever since.

Programming is often times fun, but at the end of the day I program to get a job done, back then to develop a game, not because i love solving Sudokus.

" how would we feel if tomorrow, we woke up and some guys from stanford published new research on AI which can automagically produce working software given a set of requirements -- and they give it away for free." I would personally feel great because that would mean I could focus on getting the requirements part right instead of slogging back and forth through build software that almost but doesn't quite get the requirements right and then perhaps I would have enough time to work on extending the system to being able to capture requirements from users as well.

How would we feel? In the grand scheme of things it wouldn't matter since the value in automating something like that would be absolutely enormous to productivity.

I don't write software because I like typing commands, I do it because I want to create something. We assume people do work because they want money, but real satisfaction is solving problems, and there are no shortage of those.

No when they say "bullshit" jobs they mean sort of non job that is satirized in the laundry files.

you know the sort "children's play area green sustainability consultant"

Automating actuaries should already have happened. Most actuaries will know some programming and nowadays you cannot get hired unless you can use VBA at a minimum or C++.