* Posts by Jeremy Chatfield

6 publicly visible posts • joined 19 Mar 2008

Nest cloud storage for backup fun

Jeremy Chatfield

Add another 5GB on Wuala

They raised their free storage to 5G this week.

Himalayan glaciers actually gaining ice, space scans show

Jeremy Chatfield
FAIL

Re: Then lets take the financial incentive away

IIRC there is a strong body of evidence, inspected in court, that tobacco companies managed scientific evidence about nicotine and the damage that smoking causes. Scientists who found the data equivocal got funds from various "philanthropic" organisations funded discreetly by tobacco companies. Public pressure groups in favour of smoking were set up and indirectly funded by tobacco money.

Given a few billion, you can find ways to slow down peer-reviewed science, get scientists talking in the media about the lack of data, or questioning the basis on which the data was collected, etc. And you can stimulate public debate quoting the scientists in your camp, to generate pressure to remove funding from the various institutes of health and Universities that had scientists who didn't favour your views.

Despite which, most scientists engaged in the research found that smoking was damaging and nicotine addiction was part of the reason that smokers continued to smoke. The smoking lobbyists did slow the eventual impact, the business did make profits in the extra years it won from this tactic (which was itself documented by tobacco companies and revealed in court), and the courts eventually fined them loadsamoney. And you still find people arguing that smoking is non-addictive and that it is harmless and just a way for governments to raise taxes, and that no business would damage its' own consumers. This is case where shorter term interests in financial performance overwhelmed ethics and longer term interests, and deniers were influential in slowing the whole thing down.

And the whole "they're paid they must be lying" meme is probably because of some bad economics - you know that "science" which makes assumptions that it then uses as if it was data? The core of much modern economic thinking is that humans are rational actors. It's an assumption, not a truth. People are emotional creatures. Ask anyone in marketing and their main daily activity is to get people to use emotional judgements when that sells, or to to try and engage their rational selves when that sells better. Every day, marketeers disprove the idea that humans are rational and make rational decisions. But it is still the big idea that "everyone knows" from economics. The point of science is to create a process to subject flawed human opinion to some rules that make the opinions speak less than reality. It's slow, painful and argumentative, but its the best approach humans have developed so far.

'Robots can save America', says Obama

Jeremy Chatfield
Megaphone

TITF (This Is The F...)

Americans believe in free market economics (FME). An article of faith is that FME benefits everyone. Try telling that to a working mother who had an unexpected birth complication, and whose teeth rot in her head seven years afterwards, because as a house cleaner she couldn't afford medical insurance that would have paid the hospital bill, and because she is in debt to the hospital, now can't afford the dentist to fix her teeth - yes, this is a real example from the USA. I can give you names, dates, places. Just one example of the brutality of real-world FME.

Do I have confidence that the people eliminated from the workforce because they don't have the education, *or the accidental fortune to luck into a job at the right moment*, will be looked after? Not at all. Because FME also preaches that anyone who works hard will be successful; also an article of faith, denied by reality; and that article of faith allows ignoring those in poverty, because according to the faithful, poverty is their personal failure - it isn't.

FME ignores some basics of risk and wealth. While the US is a lot freer a society than the UK (where "who you know" can count a lot for more final wealth outcomes than anything else), there is still a vast benefit to having access to capital. If you come from a wealthier family, you can afford to take more risks. If you are poor, and poorly educated (which may not be the conscious decision of a 4 year old, more a reflection of their parents and the school system available to them) then you will have less access to capital and fewer opportunities to recover from risky activities - and the primary risky activity for the very poor will be crime. Which will not endear the poor to the rich and result in job offers ('hi, as a consequence of you mugging me, I see you have the makings of an entrepreneur, let me give you $100k to set up a company" - this doesn't happen, but "as a consequence of being my child and because i have been well recompensed through my life, let me give you $100k to set up a company" - this does happen, just ask Michael Dell where his initial capital came from, or the Kennedys.).

Mass scale accelerated industrialisation will lead to mass scale poverty unless the theology of Free Market Economics is challenged. I have lived in the US for about ten years, and the rest here. I've set up companies there and here, and health insurance in the USA, and here. The risks under an FME-believing system are much greater for the poor than in much of the rest of the world. And yes, that's relative risk - but you'd still get on average, better health care in poorer Britain than the richer USA.

And, as an entrepreneur, with experience of businesses in the UK and the US, I can also affirm from that the wealth creation of the USA is a function of another component of FME, a belief that here in the UK we'd do well to emulate. That is, that to be successful you are good are selling. In the UK, when a company hits trouble, it saves first and focuses on preserving what it has. In the US, when a company hits trouble, it goes out and sells more - and if it goes bust, then bankruptcy is regarded as a transient blip making you a *better* prospect for your next venture, not worse - because you now have experience. Here, unless you have connections, you've burned and crashed. The risks of failure in business are worse in the UK, and the VCs here are wizened accountants, not gamblers - and that makes a huge difference to the likelihood of success of the ventures they invest in. Attitude affects outcomes - outcomes are not a consequence of some value-free system of economics.

Megaphone because I'm as fed up with theology in economics as I am with deists.

Death to re-keying (but not yet?)

Jeremy Chatfield

Satire

Article on rekeying is rekeyed press release. Amusing.

Google's riches rely on ads, algorithms, and worldwide confusion

Jeremy Chatfield
Alert

Re; "Making Money Is Wrong"

@James: the problem is the implicit contract, and expectation of advertisers. From 2003-2007, AdWords has taken advantage of the second price auction to set a fair price for the return. Recently, to boost revenues, Google appears to have created a false market with uninterested bidders - people advertising completely unrelated searches are conscripted by Google, without permission of the advertisers, so that Google can, it seems, bump up the bid in the auction. This is completely unlike the behaviour from 2003 to the end of 2007.

That is scandalous, if true. It's all down to this "Search History Permutation" that they've started using, without notifying advertisers that the system is fundamentally different.

What makes this scary is that Google is large enough to affect the international economy - billions of dollars of ad revenue translate to trillions of purchases. Advertisers see the changes in impression rates, CTR and conversion and assume that this is a consequence of the economy tanking, instead of Google's manipulation of the auction.

FWIW, I *am* an AdWords expert - I'm the most posted contributor to the AdWords Help Forum, outside Google's own staff, and I've been a Google Advertising Professional since shortly after the program started. I'm appalled at the recent unannounced changes.

Jeremy Chatfield

It's The Search History

There's another hidden knob, which Google has been using, and which accounts for the CTR drop - "AdWords Search History Permutation". Take the current search, and the previous search, combine each word in each query to make a new search and then find the highest paying advert for each new search. This means that consecutive searches for, for example, "cheap holiday" and "us vacation" will lead to the top advert being "Cheap US Online Shopping" - completely unrelated to the searches being performed. Bad for advertisers, bad for users, but increases Google's short term opportunity to reap cash from the highest paying advertisers. However - it will also reduce advertiser conversion rates, so it is a short term strategy, as advertiser rational response will be to drop bids on a less effective channel.