Thursday, August 8, 2019

Tesla's claims are not just unbelievable; they aren't even internally consistent

Over at Jalopnik, Aaron Gordon has an excellent rundown of the Boring Company's Las Vegas tunnel project, but the most important section focuses on another part of the Elon Musk organization. [emphasis added]

So, we’re supposed to believe Teslas will be capable of full-self driving in all conditions by next year even though, by the following year, a safety driver will be needed for a .8-mile tunnel with a dedicated right-of-way, the single simplest application of self-driving that could possibly exist.

Not only does this lend serious doubts to the Tesla robotaxi promise, but it is also a definitive step backwards from better, existing technology.

Airport people movers, close relatives to whatever the hell The Boring Company is building in Las Vegas, have been driverless for decades. Just last month, for example, I had a lovely, quick journey on Denver International Airport’s driverless people mover, which opened in 1995.

Yet here we are, in 2019, and The Boring Company says they’ll need a driver for their people-mover which moves fewer people over a shorter distance for “additional safety.”

But, hey, 1 million robotaxis on the road by next year. If you can’t believe Elon Musk, who can you believe?

For those who haven't been following this story, Musk's claim that a fleet of Tesla robotaxis is just around the corner is a perhaps essential part of the justification of the company's stock price. The audacity was remarkable, even for him. To run an Uber-type service without human backup drivers requires complete level-5 autonomy, and that appears to be years away.

Though not quite bullshit free, compared to a standard Musk spiel, the Las Vegas project has to be grounded in reality. There are actual contracts and deliverables to consider. In other words, this is what Tesla promises when they know they'll be held accountable.
Posted by at No comments:

Wednesday, August 7, 2019

As the consequences of the conservative movement's media strategy grows more costly...

Tuesday, May 16, 2017

How things got this bad -- part 4,675

I was digging through the archives researching an upcoming post and I came across a link from 2014. It led to a Talking Points Memo article that I had meant to write about at the time but had never gotten around to.

Since then, we have learned just how much the mainstream media was covering for Roger Ailes. Ideological differences proved trivial compared to social and professional ties and an often symbiotic relationship. We have also seen how unconcerned the mainstream press (and particularly the New York Times) can be a bout a genuinely chilling attack on journalism as long as that attack is directed at someone the establishment does not like.

It was a good read in 2014, but it has gained considerable resonance since then.

From Tom Kludt:

Janet Maslin didn’t much care for Gabriel Sherman’s critical biography of Roger Ailes. In her review of “The Loudest Voice in the Room” for the New York Times on Sunday, Maslin was sympathetic to Ailes and argued that Sherman’s tome was hollow. But what Maslin didn’t note is her decades-long friendship with an Ailes employee.

Gawker’s J.K. Trotter reported Wednesday on Maslin’s close bond with Peter Boyer, the former Newsweek reporter who joined Fox News as an editor in 2012. In a statement provided to Gawker, a Times spokeswoman dismissed the idea that the relationship posed a conflict of interest.

“Janet Maslin has been friends with Peter Boyer since the 1980’s when they worked together at The Times,” the spokeswoman said. “Her review of Gabe Sherman’s book was written independent of that fact.”
Posted by at No comments:

Tuesday, August 6, 2019

Tuesday Tweets











There were those like Josh Marshall who got most things right in 2016, those like Nate Silver who went from mostly wrong to mostly right, and those like Nate Cohn who were pretty consistently wrong. I'm not sure why we should still be listening to the third group.



This goes to the heart of the problem with Chait's education writings. His heart's in the right place and he cares deeply but his knowledge of the field is spotty at best. Among actual educators, this issue has been widely discussed for years.













Posted by at No comments:

Monday, August 5, 2019

Persistence of Narrative -- Netflix edition

One of the interesting things about the reaction to the recent bad news at Netflix is the way that the bulls are sticking not just with the company but with the narrative.
"The company appears to operate in a virtuous cycle, as the larger their subscriber base grows ... the more they can spend on original content, which increases the potential target market for their service," said Jeffrey Wlodarczak, an analyst with Pivotal Research Group, in a report after its recent earnings release.

Wlodarczak added that Netflix may remain a hit with consumers because it has continued to defy calls to launch a service with commercials, even though an ad-supported plan could be cheaper. He dubbed Netflix "an increasingly compelling unique entertainment experience on virtually any device."
Let's start with the rather bizarre claim that Netflix will "remain a hit with consumers" because it refused to launch a second ad-supported service. Taken on its own, the reasoning is hard to follow -- a second service might not be profitable but there's no reason it would turn off members -- but it also flies in the face of recent developments in the industry. The free-with-commercials model has had a big resurgence with major players like PlutoTV and Amazon's Freedive entering the streaming space and Terrestrial Super-Stations like MeTV posting extraordinary profits.

The claim makes no sense in terms of business, but it makes great sense in terms of story. Ads were the old way. Netflix is the disruptor. The success of the disruptor is an essential part of our modern mythology. You don't stop believing just because the evidence turns against you.

Even more central to the Netflix narrative was the role of original content. That was the competitive advantage, the secret of their success. "Secret" turned out to be something of an operative word. For years,  the company managed to keep actual numbers largely out of the discourse. Recently though, the picture has started  to fill in. We now know that, despite billions in production and billions more pushing these shows, originals account for only a quarter of the hours viewed and many of those shows don't exactly belong to Netflix.


Posted by at No comments:

Friday, August 2, 2019

11 songs that more people should know -- Special Guest Blogger

[I had to delay today's  post, so I asked an old friend (and pop culture maven) to suggest a few undeservedly obscure songs to kick off the weekend. Not actually what I meant by a "few" but I'm not complaining.]

Hello! I'm Brian Phillips and here are 11 songs that you may not know, but really should. If you like an eclectic mix of music, take a listen to my show, The Electro-Phonic Sound of Brian Phillips at rockinradio.com

1. The Crew was a positively dismal show, but fortunately, the producers had the good sense to hire Wendy Coleman and Lisa Melvoin, late of Prince and the Revolution, to do the very catchy theme. If you continue to watch...you have been warned.


2. "Treemonisha" was a triumph for the composer Scott Joplin, albeit posthumously. Sadly his opera, was never fully staged in his lifetime, but when "The Sting", which made extensive use of his music, it triggered a full-blown revival and the opera was finally staged in the 1970's. This is the finale and the oft-used Joplin quote applies here: "One should never play Ragtime fast". This is the "Real Slow Drag"


3.  From Houston, TX, it's Lee Alexander! This was recorded in 2006. This may be his only album.


4. Tribe was a band out of Boston and they recorded two albums before splitting up. They never got very popular outside of their hometown, even with the backing of Warner Brothers Records. There aren't as many great Rock songs in waltz time, but this is a good one! 


5. The Blue Ribbon Syncopators, a band out of Buffalo, NY who made six known recordings. Banjo is prominent here and, like the tuba, was mostly gone from Jazz by the 1930's


6. This is quite a lot of noise from this NY band, led by the chief songwriter, Joe Docko. Docko was 16 and one can only wonder what his music career would have been, if the Mystic Tide had any significant success. The highlight of this song are the two vocal lines. Docko destroyed all of his remaining Mystic Tide records, so what is out in the wild is all that is left.


7. Looking for the Beagles? For a time, some people were. This was a cartoon by Total Television, the same folks that made Underdog. It wasn't a hit and it was thought to be lost until recently. Surprisingly there was a soundtrack LP from the show, and it has become a collector's item, not only because of its rarity, but the music is quite good.



8. Joe Maphis was one of those fellows who went to the talent store and stayed longer than most. Don't believe me? Take a look at this and watch what he does at 2:00:


9. This is Kenneth "Thumbs" Carllile. You'll see why they called him that when you watch this. He was a sideman with Roger Miller for a number of years and released several albums under his own name.


10.  Oscar Coleman, aka Bo Dudley, is the leader of this session, but what makes this song great is the incendiary pedal steel guitar of Freddie Roulette, who was a sideman to Earl Hooker and Johnny "Big Moose" Walker, and he also recorded as a leader later in his career.


11. Have a good weekend everyone! I am not quite sure why this wasn't a hit. Here are the Loved Ones from California.
Posted by at No comments:

Thursday, August 1, 2019

What Nate silver has in common with Bob Dylan, Tom Wolfe, and Pauline Kael.

I don't remember if I ever got around to it, but for years I've been meaning to write an Andrew Gelman style class post on people whose work is great but whose imitators drag down the field.

I go back and forth on Bob Dylan here. There is no counting how many coffee house hacks have convinced themselves that the obscure lyrics they are croaking out are the next Tangled Up In Blue. On the other hand, a lot of brilliant artists have done some of their best work inspired by Dylan.

With Tom Wolfe and Pauline Kael, however, the divide in quality between the originals and the imitators is much more sharp. For more than 50 years now, we have suffered through journalists and critics who have annoyingly affected the tics and mannerisms of both writers, while having none of the talent or insight and being completely unwilling to put in the hard work of research and revision that made Wolfe and Kael so formidable.

I don't want to oversell Nate Silver at this point (he's no Bob Dylan) but he deserves great credit for his Innovations in the way we cover politics. He saw the fundamental problem with remarkable clarity and understood the appropriate level of complexity needed for a solution. As a statistician, he has considerable limitations. When he strays too far from his areas of expertise such as when he writes about climate models, the dunning-kruger effect has a tendency to come down hard, but he does have a natural talent and an all too rare capacity to learn from his mistakes.

His success unfortunately has inspired a wave of data journalists who are more often than not terrible at their jobs. It turns out that being enamored with data and yet being bad statistics is often worse than ignoring data entirely.
Posted by at No comments:

Wednesday, July 31, 2019

Any conversation about electability has got to start with the fact that a white male has not won the popular vote since 2004.

The electoral discussion among pundits and data journalists has been taking some especially silly turns of late and before the bullshit accumulates to the same dangerous level it did in 2016, we need to step back and address the bad definitions, absurd assumptions, and muddled thinking before it gets too deep.

We should probably start with the idea electability. While we can argue about the exact definition, it should not mean likely to be elected and it absolutely cannot mean will be elected.

Any productive definition of electability has got to be based on the notion of having reasonable prospect of winning. With this in mind, it is ridiculous to argue that Hillary Clinton was not electable. Lots of things had to break Trump's way for him to win the election and, while we can never say for certain what repeated runs of the simulation would show, there is no way to claim that we would have gotten the same outcome the vast majority of the time.

This leads us to a related dangerous and embarrassing trend, the unmooring of votes and outcomes. This is part of a larger genre of bad data journalism that tries to argue that relationships which are strongly correlated and even causal are unrelated because they are not deterministic and/or linear. In this world, profit or even potential profit is not relevant when discussing a startup's success. Diet and exercise have no effect on weight loss. With a little digging you can undoubtedly come up with numerous other examples.

The person who wins the popular vote may not win the electoral college, but unless you have a remarkably strong argument to the contrary, that is the way that smart money should bet.
Posted by at No comments:

Tuesday, July 30, 2019

Tuesday Tweets -- “rapid unplanned disassembly“ edition








I think Lee is underplaying the role that the huge, broad-based spikes in the late 19th Century and post-war era were in forming the idea of the exponential progress curve (and how resilient the belief has been in the face of conflicting data). 



I've been meaning to do a post on indicators that the standard political model may be heading for a rough patch.


Also want to come back to this.

Silver is having a good run.



Read this.



Posted by at 2 comments:

Monday, July 29, 2019

Taking on the votes-don't-matter meme -- the EC doesn't always follow the popular vote, but that's the way the smart money should bet

And once again we are in the silly season.

First off, a quick disclaimer: You can never reasonably dismiss an incumbent president's chances of being reelected. No matter how poor the prospects look, there are always plausible paths to victory. Not taking Trump seriously  would be irresponsible, but much of what we're hearing from the other extreme are just as foolish possibly more dangerous (the combination of defeatism, panic and bad reasoning seldom works out well).

 Popular variations on the all-is-lost theme include:

"Trump is unstoppable unless the Democrats move to the right/move to the left/embrace my pet issue."

"Trump is actually pursuing a cunning plan that will insure victory."

"The popular vote doesn't matter. The electoral college went the wrong way two out of the last five elections."

 Let's take that last one. The EC is a bad system that should have been scrapped long ago, but at least from a historical perspective, how likely are these undemocratic outcomes?

 For starters, we need to be very careful with our terms. There is a subtle but absolutely fundamental distinction to be made between winning the presidency despite losing the popular vote and winning the Electoral College despite losing the popular vote. In 2016, there is no question that the popular vote went one way and the Electoral College went another. In 2000, however, the picture is much murkier.

When we talk about a split between the popular vote and the Electoral College we mean that, given a reasonably accurate count, the all-or-nothing allotment of votes and the minimum delegate rule for small states will cause the Electoral College totals to go in a different direction that the popular vote.

There is always a certain amount of fuzziness when dealing with ballots. There might be a very slight chance that Al Gore did not win the popular vote. There is a very good chance that he did win the electoral college. Once again, I want to be very clear on this point. I'm not talking about who was awarded the delegates. I'm talking about who would have gotten them had there been a properly conducted  counting of the votes in Florida.

We've seen a recent wave of data journalist making bad distributional arguments about the implications of the Electoral College. Most of these arguments use Al Gore as an example despite the fact he does not at all illustrate their point, since who won Florida (and therefore the EC) is very much a disputed point).

If you remove Gore as an example, you have to go all the way back to 1888 to find another. This does not mean that this won't happen again for another hundred plus years. It does not mean that we don't need to worry about this in 2020. It does not even mean that a split between popular and electoral votes isn't The New Normal. What it does mean is that historically there is an extremely high correlation between winning the popular vote and winning the Electoral College.

Just for the record, the undermining of democracy by the conservative movement, particularly through voter suppression and under-representation, is perhaps my number one issue, even more than climate change and income inequality. I am very worried about gerrymandering and voter ID laws and somewhat concerned about the Senate, but the Electoral College, while not defensible, is way down on my list.
Posted by at No comments:

Friday, July 26, 2019

Family Friendly?

This is Joseph.

This tweet makes an excellent point:



The context was the cost of summer camps when you have two working parents, without anywhere near enough holidays to cover the summer.  This has to be one of the odd contradictions of modern thinking; I first noticed in in Ayn Rand's book "Atlas Shrugged" where she glossed over the role of children in a hyper-capitalist society.  While not everyone is a fan of Ayn Rand, it was an early sign of a fault point in the individualist culture we have created. 

The problem is that we can't really decide on two things.  One, is the basic unit of humanity individuals or families?  This isn't meant to be exclusionary, but simply to point out that humanity, as a project, requires new humans so if there is a commitment to the species they have to come from somewhere.  I don't want to say how these come together -- family is a very diverse entity -- but they are real mechanisms for child rearing.

Two, is the social contract that Western Democracies have created is about previous generations being transferred wealth in their old age from upcoming generations.  This works best when there is a continuing flow of upcoming generations and that requires some investment in the future as well. 

I am not sure about the solutions, but I am pretty sure that the solution set does not include seeing children as expensive consumer goods. 
Posted by at 2 comments:

Wonder when they stopped running Thunderbirds reruns in South Africa.

Listening to this, I can't help but try  and reconstruct the conversations he had with the actual engineers who tried to explain these ideas to him. He gets most of the phrases right but there's no indication he understands the challenges involved in what he's talking about.






Personally, I want to see him add one of these pilot slides.










Posted by at No comments:

Thursday, July 25, 2019

Trickle-down innovation

This is Joseph


I think that this might be the single worst argument in health care today:
The slowdown in pharmaceutical innovation is widely acknowledged, well-documented, and deeply troubling. Most Americans have health insurance. Most Americans are able to get care if they need it. What matters at the point of crisis, then, isn’t just whether someone is covered, but what that coverage can buy. The best insurance in the world won’t save us if our antibiotics fall behind drug-resistant bacteria.
This is particularly pressing for Democrats because the best argument against centralized price setting is that it will slow innovation. So what plans do Democrats have to boost innovation in the health care space? Sanders, for his part, has an interesting idea to use prizes to generate new pathways for pharmaceutical development, but he’s one of the only Democrats with any kind of plan along these lines, and he does
First, even with the current system we are seeing a slow down in innovation.  There is an assumption that that isn't driven by it being easier to rent-seek with current than to come up with new ones.  Or by intrinsic limitations in what low hanging fruit might be left.

Second, the system is actually poorly designed for the example.  Antibiotics are always going to be less profitable than chronic use drugs.  Maybe the slower pace of development is because of the current system that focuses rewards elsewhere?

Third, isn't this the same argument as "trickle down economics"?  If we make health care CEO's rich then others will also seek to become rich and that will drive innovation.  How did that work with economics? 

The US spends twice as much GDP per person as the UK.  The difference between the two systems is about 8% of GDP.  US GDP is around 20,000 billion dollars, so 8% of that would be 1,600 billion.  The NIH budget is around 40 billion.  Let's spend only half of that on research -- I wonder what endless productive NIH could do with a 840 billion dollar budget to drive medical innovation? 

And this is with me not even really trying -- just thinking off of the top of my head.  I think we should be careful with the assumption that innovation by industry is all about improving care.  Some is and there are dedicated people in industry who work hard to help patients.  But some of this is clearly rent-seeking and profit taking.  It's not clear if we could redo the system that we couldn't save money AND be more innovative. 

Just a thought.
Posted by at 2 comments:

Wednesday, July 24, 2019

The Hype Economy runs on faith -- more on Netflix


I was going to run something on this Bloomberg piece, then life got busy and a few weeks passed and soon a number of other posts (including some about Netflix) were demanding to be written. Then this happened and the following seemed too relevant to put off. [emphasis added]
Just the same, Netflix has been producing more on its own. The company will release 1,000-plus pieces of original programming this year. By the time “The Office” deal ends, Netflix will have at least 3,000 new programs in its library and likely surpass 200 million subscribers worldwide.

“People are missing it,” said [Michael Nathanson, an analyst at MoffettNathanson LLC]. “The loss of back titles will not kill Netflix or slow subscriber growth. It just forces them to make more original content.”
There is no rational argument for this position. Netflix is in the process of losing the shows behind roughly three quarters of its viewer-hours. It is maxed out on production without clear ownership of many of its most popular originals (such as She-Ra). It is spending at an unsustainable pace. It's about to fall into the ultimate competitive wood chipper.

Faith-based investing can keep things going a long time, but eventually reality-based results make doubter of us all.
Posted by at No comments:

Tuesday, July 23, 2019

Tuesday Tweets -- One Small Step



















 


Posted by at No comments:

Monday, July 22, 2019

Netflix represents the overlap of the two great traditions of creative accounting-- Hollywood and Silicon Valley

[I scheduled this a few days ago. Eventful days.]

Since writing this, I've run the story past a number of people who have worked with this kind of consumer data at large companies, and every one of them has had the same reaction. All of them suggested that Netflix was tracking these numbers; it just didn't want anyone to know what they said. I'm also coming around to that conclusion.

I have a feeling that Netflix's transparency is about to become a bigger part of this story.

Monday, October 20, 2014

The myth of a data-driven Netflix

There's been a lot of talk about Netflix stock this week (usually with words like "plummet"), but a big part of the story has largely gone unnoticed, probably in part because it involves statistics.

As mentioned before, some aspects of the Netflix narrative such as the company building and HBO type content library, are simply, factually incorrect. Others, while not blatantly wrong, are difficult to reconcile with the facts.

One of the accepted truths of the Netflix narrative is that CEO Reed Hastings is obsessed with data and everything the company does is data driven (for example "What little Netflix has also shared about its programming strategy is that its every decision is guided by data."). The evidence in support of this belief is largely limited to a model that Netflix crowd sourced a few years ago and to endless assertions from executives at the company that they do know what they are doing despite evidence to the contrary.

Of course, all 21st century corporations are relatively data-driven. The fact that Netflix has large data sets on customer behavior does not set it apart, nor does the fact that it has occasionally made use of that data. Furthermore, we have extensive evidence that the company often makes less use of certain data then do most other competitors.

On pertinent case in point, particularly for the SEC, is churn rates.
But Netflix disagrees. “With respect to various operational metrics, management has evolved its use of these metrics as the business has evolved,” it wrote the SEC in response. Because it is so easy to quit and then restart a Netflix subscription, it said, “the churn metric is a less reliable measure of business performance, specifically consumer acceptance of the service.”
This is problematic on any number of levels. In terms of marketing, pricing and long-term corporate strategy, having a complete picture of how long people stay and why they leave is huge. The only excuse for not reporting churn would be if you had such a detailed picture of who was leaving and why that this additional metric was redundant.

In other words, Hasting should have a good, data-supported explanation for a recent sudden loss of subscribers.
Netflix CEO Reed Hastings blamed the subscriber drop-off on a $1 price increase the company instituted back this spring.

"Our best sense is it's an effect of our price increase back in May," Hastings said Wednesday night in an interview with CNBC. "With a little bit higher prices, you get a little bit fewer subscribers. So that's our sense of it. But we can't be 100% sure. We had so much benefit from Orange in Q2 and the early Q3, but that's what we think."
Phrases you don't want to hear in these circumstances include "our best sense" and "that's what we think." They convey the impression of a CEO who was blindsided by a bad day at the NASDAQ.


When contemplating a price increase, well-run companies look at the impact on retention and on acquisition. When Netflix management said
[M]anagement believes that in a largely fixed-cost streaming world with ease of cancellation and subsequent rejoin, net additions provides the most meaningful insight into our business performance and consumer acceptance of our service. The churn metric is a less relevant and reliable measure of business performance, and does not accurately reflect consumer acceptance of our service.
They were basically saying that losing one customer and gaining another is the same as keeping the same customer. That's a dangerous approach under the best of circumstances but it can be deadly when trying to gauge the impact of a pricing change.

Just to be clear, for years analysts and the SEC have been asking for more data, or at least more detailed statistics and Netflix has been saying "trust us, the aggregate number are good enough." Now the company appears to have screwed up badly, and they've done it in pretty much exactly the way you would expect a company to screw up when it doesn't drill down into the data.
Posted by at No comments:
Subscribe to: Posts (Atom)