Automated software is by definition shit
Apr. 13th, 2011 01:49 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
peterbirksConsidering the fact that I never write about amazing wizz-bang technology solutions that companies based near Slough have come up with for the insurance industry, it's surprising how many emails I continue to receive about wizz-bang technology solutions that, the senders claim, will change the way the entire industry works.
They tend to follow a well-trodden path. Some kind of computer program will do things automatically at 10% of the cost of deploying staff. I'm quite sure that 95% of the time these things work; the problem is, the 5% of the time that they don't work makes them look so ridiculous that you forget about the other 95%.
Amazon, for example. The algorithm followed by the wizz-bangers who put together Amazon's automatic email marketing follows the conventional wisdom that if you buy something (say, a shirt), then you are more likely to be receptive to shirt marketing (even though you have just bought one) than a random selection of the population. Unfortunately, this seems to run the entire gamut of the Amazon range, and I have to assume that this is because no-one in the coding side of the chain can be bothered to point out that if someone has just bought a television (i.e., in this case, me), there isn't much point in emailing them with television offers the following day.
Of course, the intelligent coding guy will say that so much is obvious, but that it isn't his job to do that. It's his job to do what the marketing whizzes tell him to do, and it's their job to spot that there's a certain "price limit" to this marketing technique.
This kind of attitude, in turn, explains why these coders are still coders, rather than something else with a rather higher level of pay, satisfaction, and status.
Facebook, meanwhile is constantly throwing insurance banner ads at me. Why? because I visit a lot of insurance-related sites, and "insurance" probably appears in lots of triggering places. Clearly, therefore, I am looking to buy insurance! Except, of course, I'm not, and the pattern of my insurance "visits" would indicate this straight away to a non-automated interface (viz, a human being).
Marketing people are delusional. This much we know. Most of humanity therefore avoids marketing people because dealing with them for too long renders you either furious or insane. But what about those people who get the insane diktats from marketers and who follow up those insane diktats without a word of dissent, even though they know it will end in tears? I'm afraid that such an attitude baffles me, steeped as it is in what looks like self-interest but which is in fact an unspoken collusion to fuck up the company that you work for, rather than help it succeeed.
+++++++++
In the land of economics, we are quickly getting into the land of where black equals white. Professor David Blanchflower was on Wake Up To Money this morning and I concluded that he was an idiot who seemed to think that inflation actually created wealth, rather than shifting it from one sector of society to another. I presume that he would deny that this is what he said, but the whole thrust of his argument in essence consisted of transferring wealth from those who have saved to those who have spent too stupidly in a past, pre-2008 life. If that's what you want, fine, but don't dress it up in different clothes.
However, Blanchflower was a paragon of sanity compared with the views expressed in an article in today's FT by Lorenzo Bini Smaghi, a member of the executive board of the European Central Bank. ( http://on.ft.com/igGAM8 ) If there's a logical thread to this article, it escaped me. He starts off with a para stating that "the principle of no taxation without representation should work both ways" (meaning, er, what? "no representation without taxation"?) No, what he means is that, if taxpayers ARE represented, then they have to live with the consequences of their decisions. He then says that "in Europe this relationship between taxpayers and the financial system is not working".
From this point, Smaghi heads off into a second para that attacks the moves by the UK and Ireland to have "favourable taxation for banks, and less stringent self-regulation".
This, he said, meant that "their banks grew larger and more profitable".
Well, er, yes, we can see how much larger and more profitable the UK and Ireland banks are at the moment as a result of "favourable taxation". The Irish banking system is bust, while Royal Bank Of Scotland and Bank Of Scotland have effectively ceased to exist in their previous form.
But, wait, says Smaghi, "when the crisis erupted, Irish taxpayers found themselves faced with the collapse of tax revenues and the losses incurred by the banks themselves". So, where's this "more profitable" argument gone? Either they were profitable or they weren't.
At this point Smaghi says that in principle it should be shareholders "and managers" who whould bear the costs of restructuring these banks (remember, those ones that were "more profitable") with losses subsequently felt by bondholders.
Why is this the case? Smaghi does not say. Instead he heads off into his fourth argument of the piece (rambling doesn't enter it) where he asks how you decide whether the crisis is truly systemic.
He doesn't answer the rhetorical question, stating instead that "here we encounter a problem" (translation = "I have asked a rhetorical question to which I do not know the answer"). Smaghi then spends half the article pondering whether the whole problem is based on national regulation vs euro-wide regulation.
Now, the step-by-step non-logic of all of this is too tedious to unravel at length. However, it's clear (to me) that at no point does Smaghi justify why Ireland's taxpayers should take their medicine uncomplainingly. His whole diversion into the national regulation vs EC-wide regulation is a complete red herring. Thus his final paragraph:
What?
Well, what IS Smaghi's claim? If you want an EC-wide regulatory system, fine. There are strong arguments for this. But this final paragaph implies, nay, effectively states, that removal of discretion in the implementation of national financial regulations would reduce the risk that the taxpayer would have to foot the bill (but, if they did have to foot the bill, they would at least have the moral right to moan about it).
Smaghi does not say that whether the risk would be reduced because SOMEONE ELSE would foot the bill (in this case, taxpayers throughout Europe), or whether an EC-wide system of regulation would make it less likely that that a systemic crisis would occur. Smaghi also says at one point that if there is a systemic crisis, taxpayers SHOULD foot the bill, but later he says that an EC-wide regulatory system would make governments more willing to accept that the crisis was systemic, thus shifting the burden to other Euro taxpayers.
Oh, bollocks, why try to understand stuff that's so badly thought out? The worrying thing is that this man is on the executive board of the ECB -- one of the institutions with which I have had most sympathy in the current crisis. He might have some good ideas in his head, but it would be better if hwe could express them in a cogent form comprehensible to other students of the current crisis.
++++++++++++++++++
In Moretugal-land, City AM had another misleading headline. "Lisbon bailout to be in place by June vote", it said. Really? I had my doubts, and read the shortish article. A different story quickly emerged.
What Portuguese finance minister Fernando Teixeira dos Santos actually said was that "a fully negotiated package" would be ready, giving voters the chance to accept or reject it at the polls.
This, of course, supports my initial premise that the current administration does not have the authority to put a bailout into action, although it would appear that it thinks it has the authority to negotiate it.
Actually, I don't think that the administration really has the authority to negotiate the bailout, but the speed at which Portugal is going broke is causing a "needs must" approach. The Finance Minister said that the bailout would be about €80bn and would be based on that austerity package that parliament had rejected. The election is is early June. In mid-June Portugal faces a €4.9bn bond redemption. Portugal's FM admitted that a deal had to be in place by then. So, the theory is I guess that in the June elections a party or coalition is elected that agrees to push through the deal that is on the table.
As we have seen (Irish referendum, Icelandic referendum), the electorate often hasn't read the script. What happens if a government is re-elected (and we're looking at PR here, so you might not even get a clear-cut result straightaway) that is opposed to the deal or, more likely, wants to "tweak" it? How close to the wire is everyone willing to go before the dreaded word "default" is uttered and a myriad of French and German banks have to write down money that they can't afford to lose?
Those days after the June election in Portugal will likely be more interesting than the days after the upcoming vote in Finland. Indeed, just about every election in Europe over the coming 12 months will be dominated by voters who either don't want to pay for their own country's profligacy or don't want to pay for other countries' profligacy.
Returning to Smaghi, only three things are certain:
1) The crisis IS systemic.
2) A euro-wide regulatory body wouldn't have made fuck-all difference in this case, so stop pretending that it would.
3) Someone is going to have to pay, either the borrowers or the lenders, but we still haven't decided for sure which it will be (Lenders are currently beating borrowers 1-0 after 20 minutes' play)
______________
They tend to follow a well-trodden path. Some kind of computer program will do things automatically at 10% of the cost of deploying staff. I'm quite sure that 95% of the time these things work; the problem is, the 5% of the time that they don't work makes them look so ridiculous that you forget about the other 95%.
Amazon, for example. The algorithm followed by the wizz-bangers who put together Amazon's automatic email marketing follows the conventional wisdom that if you buy something (say, a shirt), then you are more likely to be receptive to shirt marketing (even though you have just bought one) than a random selection of the population. Unfortunately, this seems to run the entire gamut of the Amazon range, and I have to assume that this is because no-one in the coding side of the chain can be bothered to point out that if someone has just bought a television (i.e., in this case, me), there isn't much point in emailing them with television offers the following day.
Of course, the intelligent coding guy will say that so much is obvious, but that it isn't his job to do that. It's his job to do what the marketing whizzes tell him to do, and it's their job to spot that there's a certain "price limit" to this marketing technique.
This kind of attitude, in turn, explains why these coders are still coders, rather than something else with a rather higher level of pay, satisfaction, and status.
Facebook, meanwhile is constantly throwing insurance banner ads at me. Why? because I visit a lot of insurance-related sites, and "insurance" probably appears in lots of triggering places. Clearly, therefore, I am looking to buy insurance! Except, of course, I'm not, and the pattern of my insurance "visits" would indicate this straight away to a non-automated interface (viz, a human being).
Marketing people are delusional. This much we know. Most of humanity therefore avoids marketing people because dealing with them for too long renders you either furious or insane. But what about those people who get the insane diktats from marketers and who follow up those insane diktats without a word of dissent, even though they know it will end in tears? I'm afraid that such an attitude baffles me, steeped as it is in what looks like self-interest but which is in fact an unspoken collusion to fuck up the company that you work for, rather than help it succeeed.
+++++++++
In the land of economics, we are quickly getting into the land of where black equals white. Professor David Blanchflower was on Wake Up To Money this morning and I concluded that he was an idiot who seemed to think that inflation actually created wealth, rather than shifting it from one sector of society to another. I presume that he would deny that this is what he said, but the whole thrust of his argument in essence consisted of transferring wealth from those who have saved to those who have spent too stupidly in a past, pre-2008 life. If that's what you want, fine, but don't dress it up in different clothes.
However, Blanchflower was a paragon of sanity compared with the views expressed in an article in today's FT by Lorenzo Bini Smaghi, a member of the executive board of the European Central Bank. ( http://on.ft.com/igGAM8 ) If there's a logical thread to this article, it escaped me. He starts off with a para stating that "the principle of no taxation without representation should work both ways" (meaning, er, what? "no representation without taxation"?) No, what he means is that, if taxpayers ARE represented, then they have to live with the consequences of their decisions. He then says that "in Europe this relationship between taxpayers and the financial system is not working".
From this point, Smaghi heads off into a second para that attacks the moves by the UK and Ireland to have "favourable taxation for banks, and less stringent self-regulation".
This, he said, meant that "their banks grew larger and more profitable".
Well, er, yes, we can see how much larger and more profitable the UK and Ireland banks are at the moment as a result of "favourable taxation". The Irish banking system is bust, while Royal Bank Of Scotland and Bank Of Scotland have effectively ceased to exist in their previous form.
But, wait, says Smaghi, "when the crisis erupted, Irish taxpayers found themselves faced with the collapse of tax revenues and the losses incurred by the banks themselves". So, where's this "more profitable" argument gone? Either they were profitable or they weren't.
At this point Smaghi says that in principle it should be shareholders "and managers" who whould bear the costs of restructuring these banks (remember, those ones that were "more profitable") with losses subsequently felt by bondholders.
"Only in a systemic crisis should taxpayers be involved."
Why is this the case? Smaghi does not say. Instead he heads off into his fourth argument of the piece (rambling doesn't enter it) where he asks how you decide whether the crisis is truly systemic.
He doesn't answer the rhetorical question, stating instead that "here we encounter a problem" (translation = "I have asked a rhetorical question to which I do not know the answer"). Smaghi then spends half the article pondering whether the whole problem is based on national regulation vs euro-wide regulation.
Now, the step-by-step non-logic of all of this is too tedious to unravel at length. However, it's clear (to me) that at no point does Smaghi justify why Ireland's taxpayers should take their medicine uncomplainingly. His whole diversion into the national regulation vs EC-wide regulation is a complete red herring. Thus his final paragraph:
"Recent events have shown that, as long as the accountability of supervisors to taxpayers is primarily a national affair, and discretion in the implementation of national financial regulations and supervision is allowed, then there is a high risk that taxpayers will foot most of the bill. They should not complain when it actually happens".
What?
Well, what IS Smaghi's claim? If you want an EC-wide regulatory system, fine. There are strong arguments for this. But this final paragaph implies, nay, effectively states, that removal of discretion in the implementation of national financial regulations would reduce the risk that the taxpayer would have to foot the bill (but, if they did have to foot the bill, they would at least have the moral right to moan about it).
Smaghi does not say that whether the risk would be reduced because SOMEONE ELSE would foot the bill (in this case, taxpayers throughout Europe), or whether an EC-wide system of regulation would make it less likely that that a systemic crisis would occur. Smaghi also says at one point that if there is a systemic crisis, taxpayers SHOULD foot the bill, but later he says that an EC-wide regulatory system would make governments more willing to accept that the crisis was systemic, thus shifting the burden to other Euro taxpayers.
Oh, bollocks, why try to understand stuff that's so badly thought out? The worrying thing is that this man is on the executive board of the ECB -- one of the institutions with which I have had most sympathy in the current crisis. He might have some good ideas in his head, but it would be better if hwe could express them in a cogent form comprehensible to other students of the current crisis.
++++++++++++++++++
In Moretugal-land, City AM had another misleading headline. "Lisbon bailout to be in place by June vote", it said. Really? I had my doubts, and read the shortish article. A different story quickly emerged.
What Portuguese finance minister Fernando Teixeira dos Santos actually said was that "a fully negotiated package" would be ready, giving voters the chance to accept or reject it at the polls.
This, of course, supports my initial premise that the current administration does not have the authority to put a bailout into action, although it would appear that it thinks it has the authority to negotiate it.
Actually, I don't think that the administration really has the authority to negotiate the bailout, but the speed at which Portugal is going broke is causing a "needs must" approach. The Finance Minister said that the bailout would be about €80bn and would be based on that austerity package that parliament had rejected. The election is is early June. In mid-June Portugal faces a €4.9bn bond redemption. Portugal's FM admitted that a deal had to be in place by then. So, the theory is I guess that in the June elections a party or coalition is elected that agrees to push through the deal that is on the table.
As we have seen (Irish referendum, Icelandic referendum), the electorate often hasn't read the script. What happens if a government is re-elected (and we're looking at PR here, so you might not even get a clear-cut result straightaway) that is opposed to the deal or, more likely, wants to "tweak" it? How close to the wire is everyone willing to go before the dreaded word "default" is uttered and a myriad of French and German banks have to write down money that they can't afford to lose?
Those days after the June election in Portugal will likely be more interesting than the days after the upcoming vote in Finland. Indeed, just about every election in Europe over the coming 12 months will be dominated by voters who either don't want to pay for their own country's profligacy or don't want to pay for other countries' profligacy.
Returning to Smaghi, only three things are certain:
1) The crisis IS systemic.
2) A euro-wide regulatory body wouldn't have made fuck-all difference in this case, so stop pretending that it would.
3) Someone is going to have to pay, either the borrowers or the lenders, but we still haven't decided for sure which it will be (Lenders are currently beating borrowers 1-0 after 20 minutes' play)
______________
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2011-04-13 06:53 pm (UTC)Bollocks
Date: 2011-04-16 01:02 am (UTC)Maybe I should give up comments on something I know nothing about and despise on a purely axiomatic basis, like, say, macro-economics; because you sure as shit know nothing at all about how the software world works.
Maybe I'm out of the loop these days, but I don't think I'm all that unusual in my utterly overpowering desire to ignore "marketing." We're either geeks or we ain't, y'know.
There are good (well, easily explained) reasons for the inadequacies you point out, and they have nothing to do with the poor sods at the chalk face. I could, of course, explain them here, but That Would Be Too Easy. In fact, I will leave it as An Exercise For The Reader.
Points will be awarded for outsourcing, cultural differences, management fads, and an unusual conjuction of Venus with Orion.
Basically, you are (and I am sorry to point this out to you) talking out of your arse here.
(And, btw, those of us actually inside the Walled Garden understand that none of this is to do with "coders." It's actually to do with "modellers." Modellers are like models, except far less sexually attractive, and even more pointless in the general scheme of things. We can build the engine for them, but will they come?
Answers on the back of a postcard from Nice, please.
-----------------------------------------------
On the other side of the fence, I'm deeply impressed by the Moretueguese stuff. My original theory about the Euro collapsing circa 2012 was based on an economic guess. My recent theory of 2014 was based on sane politics. The interesting thing here, in a way, is that the politics (as you so rightly point out -- and it appears to be leaking out into Economic Theory) is no longer sane.
Step aside for a moment, and consider exactly why any political leader would stake their (political) life on the continuation of the Eurozone. I mean, obviously, it makes no sense to the UK and never has done. I think it hits the cow-bells in Munich and Eurocrats the Hell out of the Ecole Politique ... but these two attitudes don't coalesce very well, and neither one makes much sense.
The banks will go tits up and all Hell will break loose.
And it's very difficult to judge exactly when: because, as you say, these people are extraordinarily delusional.
--------------------------
(BTW the long term solution to not selling you a second television set, or whatever, is to remove humans out of the equation altogether and to go with what is basically an inverse anti-spam engine. Trust me. Humans are expensive. Also, unreliable. This is our goal at Bing.)
Re: Bollocks
Date: 2011-04-17 10:38 am (UTC)From Dominique Strauss-Kahn at the IMF:
"It's probably too much to say that it's a jobless recovery, but it's certainly a recovery with not enough jobs".
I'm no Luddite when it comes to progressing jobs out of existence. I understand the grand scheme of things when it comes to overall wealth. I just don't think that the automation we are seeing today increases the overall wealth (and we do have evidence that people are getting more miserable, rather than happier). It just appears to make it cheaper for individual companies. In the grand scheme, this automation DESTROYS overall wealth. Not only are jobs destroyed, but customers pay slightly less for a far worse experience. I got a bollocks automated email from Direct Line the other day. As with most of these, it said "please do not reply to this email". Why? Because there is no-one employed to read any reply I might make. It's "more efficient" that way. But it is frustrating for the consumer. Humans are, indeed, "unreliable". But they are HUMANS. That makes a big positive difference.
PJ
Re: Bollocks
Date: 2011-04-17 04:11 pm (UTC)One of the more worrying features about working in a computer environment is that, at bottom, your job is to put people out of work. Now, it so happens that they are generally doing tedious and repetitive work. But on the whole they prefer it to an absence thereof.
Indeed, one of the more worrying features about working in a modern capitalist environment is that, at bottom, the whole damn superstructure is geared towards putting people out of work. (If I didn't write this software, then someone else would, or so goes the argument.) It's the Prisoner's Dilemma all over again, isn't it? Except that, in this case, the only player in the game who is willing to take up the employment slack is the public sector (and Galbraith's Affluent Society suggests that there may, indeed, be slack to be taken up) ... and of course taxation is an axiomatic evil, so there you go and thanks very much for your ultimately doomed aspirations.
Where, perhaps, we diverge is that I have seen at first hand the relative merits of human-based scoring -- for want of a better word -- and machine-learning based scoring. Humans suck. Even trained and experienced and quite intelligent humans on a decent salary suck. The problem is that humans do not scale, whereas a computer algorithm (at least a moderately successful one) does.
I agree that Amazon's version is an abomination. I also spend more time than is healthy, either on Google or on Bing, pondering the bizarre "targeted" advertisements that pop up. ("Learn Norwegian in your spare time?" Good grief. Right now I am being offered cheap hotels in somewhere called "Londres" from booking.com, for some reason probably to do with my investigation of French news feeds at work, so I have a lot of sympathy for your predicament.)
However, this is a question of implementation; not of intent.
Re: Bollocks
Date: 2011-04-17 04:18 pm (UTC)How come, when it's a place named by the Romans in the neuter singular (Londinium), it is now known by the French in something I can only assume is the feminine plural?
I am a staunch advocate of respecting the Ancient Greek rule that a neuter plural (data) be associated with the singular conjugation, but this aberrant inverse of the rule baffles me no end.
stars ftp etc.. where is your blog entry, cant wait
Date: 2011-04-17 02:12 am (UTC)Re: stars ftp etc.. where is your blog entry, cant wait
Date: 2011-04-17 06:57 am (UTC)PJ
how to unlock iphone 4 bb
Date: 2011-06-22 03:47 pm (UTC)unlock iphone 4
how to unlock iphone 4 how to unlock iphone 4 (http://unlockiphone421.com) unlock iphone 4 how to unlock iphone 4
_________________
how to unlock iphone 4 [url=http://unlockiphone421.com]unlock iphone 4[/url] how to unlock iphone 4 unlock iphone 4
buy facebook page likes dw
Date: 2011-07-31 07:53 am (UTC)buy facebook likes
buy facebook likes cheap
buy facebook fans cheap buy facebook likes cheap
(http://www.natalie-williams.com/apps/profile/77098273/) buy likes on facebook how to buy facebook likes
_________________
buy targeted facebook likes [url=http://www.natalie-williams.com/apps/profile/77098273/]buy facebook page likes[/url] buy facebook page likes facebook likes buy