Short Takes: Tolkien, Clickbait, Tesla
The missing context behind Amazon's LOTR budget, a counterpoint to journalism's clickbait defense, and revisiting a rogue officer's framing of April's Model S fire.
Back again this morning with the new format, which I’ll be mixing in with my feature pieces here and there. I’m targeting 3x standalone stories per edition at 500-750 words each, where more time-pressed readers can choose freely from the day’s menu.
I’m also officially introducing footnotes, because some of these stories have side details that will fascinate some and bore others.
We reward corrections. See something wrong, misleading, or unfair? Say something. It helps. New readers can learn more about our mission here.
I.
What’s the most you’d expect a country to pay to subsidize the production of a single season of a TV show?
The Hollywood Reporter and others (including Reuters) have suggested that as much as ~0.13% of New Zealand’s entire national revenue [1] is headed to Amazon’s coffers to support just the first season of their Tolkien project (that’s often referred to as Lord of the Rings, even though plotwise, and perhaps legally, it’s only loosely related). [2]
If that percentage isn’t mad enough, the absolute numbers are also remarkable. We’re told that the total production budget (sans global marketing and most expenses incurred out-of-country) [3] is currently pegged at $465m USD, which New Zealand is now on the hook to rebate at a rate of 25%. [4] Put another way, they’d be rebating a greater total amount than 99%+ of TV seasons have ever cost.
This seemed, uh, unusual to me? But while it obviously couldn’t be correct in a conventional sense, the estimate came from somewhere and signified something.
We got a clue a month later from Amazon Studio head Jennifer Salke:
This is a full season of a huge world-building show. The number is a sexy headline or a crazy headline that’s fun to click on, but that is really building the infrastructure of what will sustain the whole series.
What I infer from this (and what seems in line with the program’s official rules) is that Amazon is frontloading season one with a lot of series-wide investments into sets, costumes, etc. And that’s fine. But it also seems that some production technologies are eligible, including the R&D behind them (see section 2.2 of appendix 2).
Rebating Amazon for stuff that’s going to be used for this and other productions in New Zealand is unremarkable in itself. But I were Amazon and someone made me this offer, I’d immediately throw $200m at R&D that can be repurposed elsewhere and say “yes, 25% back, thank you very much”.
If this indeed is what’s happening (it’s difficult to reckon how Amazon is otherwise spending 3x per episode what HBO did on Game of Thrones), I… admire it in a way? Leveraging an as-of-right incentive to the degree that a country’s treasury department has to issue an advisory about how they’re going to pay you is good corporate art really.
But should they be allowed? Is this a good deal for New Zealand? I don’t know! We seem to have the budget disclosure because Stuff, a local NZ publication, filed an Official Information Act request for the gov’s agreement with Amazon, leading the Economic Development Minister to go public with initial estimates. [5] While I applaud Stuff for their initial work here, they have more questions to ask still.
(Mind you, even if Amazon is loading up on portable R&D here, that’s not to say that New Zealand’s taxpayers are necessarily getting a raw deal. The government’s memo of understanding with Amazon seems largely well-considered, and there’s a lot of upside both to tourism and building out blockbuster-supporting capacity locally. Even so, it’s good for citizens to know what they’re paying for so that they can vote with eyes wide open, and these “big budget, much wow” articles don’t really aid that mission.) [6]
Anyway, as a PS, compare CNBC’s clickshare write-up with the prior coverage. We need a term for this functional plagiarism where no new information is added.
II.
(Caveat: My concern here is largely narrow to national/global papers and their top beats. It also elides the difficulties of headline-writing for print editions, which is thornier.) [7]
One argument you hear from some journalists is that “clicks” are a mythical motivation, and that those imagining otherwise just don’t grok journalism.
Of the various ways of rebutting this, I want to focus here on headlines (along with subheads and captions, which are typically done as a package):
Journalists generally aren’t free to write their own, despite having the strongest grasp of their story’s thesis, context, nuances, unique claims, etc.
This work usually falls to editors/subeditors whose job (sometimes with the support of split-testing technology) is optimizing the story for clickability, Google friendliness, and social sharability (which ultimately drives clicks)
If the resulting headline ends up being misleading or overly sensational, the reader is rarely told who the editor was, and defenders of the reporters involved will make sure to say “oh boy yeah that was off, but hey they didn’t write it!”
So while our protesting journalist may sometimes be right in some narrow and meaningless sense that the body of a given text was written with a pure and disinterested motive, that’s clearly trumped by the outlet’s interests. And those things are hardly severable to the journalist who’s being paid from the outlet’s winnings!
Now, of course there’s nothing wrong with leaning on the help of a headline specialist to spur creativity, or using split-tests to decide which of two or three equally fair and factual alternatives plays best. And for print there are always going to be compromises based on layout concerns. But those sorts of things aren’t our concern here.
Example one:
Here we have a case of a bad headline (top) coming from a bad article. Reuters came up with a very muddled viewpoint, which set up the sub/editor for sure failure. But part of their job (typically) is to save reporters from themselves. Why didn’t that happen here? Well, are they being paid more to make the journalism better or to drive readers to it? And if we admit the latter, which of those two headlines is going to drive interest?
[EDIT: Per a reader note, apparently most Reuters journalists write their own headlines, or at least have a seat at that table.]
Anyway, Reuters is about to look great by comparison:
The NYT published a web article on May 4th about a photoshoot that musician Billie Eilish did for British Vogue, where the gist was that Billie adopted a look somewhat out of character with her prior brand
The conflict-laden subhead: “The pop star known for defying gender stereotypes got a glamour makeover with a corset. Not everyone is happy about it.”
But the text itself made a very weak case for the latter point, only quoting a single example of anyone’s displeasure (a Twitter account with four followers!)
The Atlantic’s Derek Thompson pointed out the absurdity in a tweet at 1:44pm ET on May 5th (i.e., the next day)
At 3:05pm ET on the 5th, the story was updated (with a public timestamp) to correct a typo and add a second example to support the subhead’s argument [8] (also from a random Twitter account, which was subsequently deleted)
But the correction note left on the article only mentioned the typo and not the addition (as ninja-edits are in keeping with official NYT policy)
The print version on May 6th was then headlined “First, the Bombshell Cover. Then the Fallout”, which is a more aggressive version of the same problem
At some point the edit timestamp fell off the article altogether
Who’s to blame here? Everyone! Who’s being well-served here? Anyone?
These are just two examples. I have more. Many, many more. Sometimes it’s a good article betrayed; more often it’s a problemed article made worse. Sometimes it’s a whole suite of issues tangled together. But the underlying issue is often the same: the endless sacrifices made on the altar of traffic-maximization.
While I’m sympathetic to the broken models and incentives hampering the work of well-meaning journalists, those suggesting that reckless traffic-seeking isn’t profoundly warping the industry just aren’t doing justice to the true nature of the problem.
III.
Remember back in April when a Tesla Model S crashed and caught fire, killing its two occupants, after which it took firefighters four hours just to put the flames out?
That’s what the Houston Chronicle reported, as did the AP, as did the Washington Post, as did The New York Times.
But between these reports we mostly have just a single source, a Constable Herman, who made a rather strong set of claims:
That the deceased had been talking to their wives about Tesla’s Autopilot capabilities just prior to leaving
That the driver’s seat was empty on impact (“100% certain”), implying that the accident was an Autopilot test gone wrong
That deputies called Tesla for firefighting guidance and failed to get an answer
That it ultimately took four hours to extinguish the flames
If we take all this in an uncritical sense, it sounds really bad for Tesla! Two guys on a quiet residential street test out what Autopilot can do, and their deathmobile immediately careens into a tree (via a culvert and raised manhole). Someone phones in the emergency, but officials can’t get a hold of Tesla to instruct them how to put out this special fire, which forces them to invest unprecedented resources to rein it in.
Except, in reality:
Constable Herman is the only source I can find for the wives claim, and he’s (as we’ll see) a wildly unreliable narrator who was almost certainly basing his comments on secondhand info, including half-understandings and conjecture.
We still don’t know today if the 59 year-old driver was in his seat at impact. The NTSB’s preliminary report confirms that he was in it when they left the house, which was only ~550 feet from the crash site. Could he have moved into the back over that distance? Possibly. But why? He didn’t have Autopilot or Autosteer turned on (as the former requires the latter, and the latter only works on streets with marked lanes, which his street didn’t have). Perhaps they crashed normally and the doors jammed, locking them in? That happens a lot in accidents, lots of which turn into engine fires. It’s tragic. But also a purely local news story.
The fire chief whose crew responded to and managed the fire said they never called Tesla, and never needed to. [9]
Said chief said it took his crew “two to three minutes” to put the active flames out, after which they just occasionally poured water on the battery to keep it cool and smother recombustion. While they were on the scene doing that for roughly four hours, the actual battling of the blaze was over almost immediately.
So what exactly is Tesla guilty of here? What’s the actual story that’s justifying national coverage?
The famously-neutral (ha) Associated Press offers us a moral:
[Tesla] warns customers that its driver-assist system, called Autopilot, is not an autonomous-driving program, and that they must pay attention and be ready to take control of the vehicle. However, the National Transportation Safety Board said last year that the design of the system allows drivers to avoid paying attention and fails to limit where Autopilot can be used.
Fun fact: every car ever made allows drivers to not pay attention. This is why lots of people die! The point is whether Tesla ever set an expectation of “yeah it doesn’t matter if the magic-car-saves-your-ass feature is activated or not, it will still work”. But they, uh, obviously didn’t? And if I decide to stress-test that system in transit, WITHOUT MAKING SURE IT’S ON FIRST, I’m sorry that is 100% on me. And if I compound that by climbing into the back seat, I mean c’mon.
In contrast to this style of “reporting”, consider a piece like this. Rather than sensationalize a local news story full of unknowns, it looks at how important these new technologies are, and what the future could look like “if we let it happen”.
Part of that letting, I think, is a collective willingness to not write or reward this kind of journalism. It doesn’t spark inspiration or hone any critical faculties that might lend to growth. It doesn’t advance us upward or forward. It’s just cheap angerbait that brings in traffic, and we should all hate it for what it’s going to cost our children.
Oh, and none of the four articles linked have seen a correction in the two months since they were first published.
[1] I got the 0.13% by dividing their pre-COVID revenues of NZ$119.3 billion by the estimated max size of the credit (NZ$157.5 million). While the NZ gov’s official statement says NZ$162.5m, that’s wrong for reasons outlined in footnote 4.
[2a] One curious subplot of Amazon’s $250m rights deal with the Tolkien Estate is that said estate doesn’t actually hold most of the rights to the Lord of the Rings trilogy itself. Those largely belong to Middle-earth Enterprises (a division of The Saul Zaentz Company), who bought then from United Artists back in 1976. While TV adaptation rights fall into a weird grey area, it wasn’t/isn’t clear that Amazon was/is clear to use any characters or names that might violate the trademarks held by Middle-earth. While it’s possible that Amazon and/or the Tolkien Estate struck a private deal with them to get around this (possibly as part of the big 2017 multi-party settlement about the extent of Middle-earth’s rights), I’m still stuck on the point of “if they own the right to Lord of the Rings, why are they starting with stories from a much less valuable part of the Tolkien legendarium?” I guess the answer could just be “to be new”, where they expect the Tolkien brand and the show’s budget to bring in casual fans who don’t otherwise have any familiarity with the characters on offer. But I’m a bit skeptical here. If they had a deal all along, why not just say that? And why not mention Middle-earth among their co-producers in their initial press release? And why not do a second show starring Gandalf etc? The relevant info is also redacted in their memo of understanding with NZ (pages 2-4, 10).
[2b] There are claims out there that MGM has a longstanding deal with Middle-earth, which perhaps allows them to license out some shared rights. If true, this might give more context to why Amazon was so motivated to acquire them. While they paid a pretty penny, if the move does indeed remove meaningful restrictions rightswise, I can see that being worth a chunk of the purchase price. But this is hard to verify, so I take this as a matter of intrigue more than fact.
[3] The official list of eligible expenditures also suggests that Amazon isn’t shoving in the $250m cost of their initial rights deal. The rule at 17.4(b)(i-ii) seems pretty clear that story rights are only eligible if the original copyright owner is or was a NZ resident, which Tolkien wasn’t. There are also restrictions on cast wages and other above-the-line expenses, per rule (o) in that same section. So either this $465m really is mostly below-the-line production costs (including R&D) or everyone has confused total estimated spending with total eligible spending and just never corrected it.
[4] New Zealand offers a 20% rebate as their standard international offer for qualifying projects, with an extra 5% for larger ones that meet special criteria. For weird one-off reasons, Amazon (via a special-purpose vehicle) is only getting 20% for the first NZ$100m of qualified spending on season 1, then 25% for everything above that. See note at 3.1(c) on page 12 here.
[5] I suppose those things could have happened in reverse order? It’s the less likely explanation, but plausible I guess.
[6] One could argue that global/US outlets don’t have any obligation to citizens of New Zealand, and are thus free to just run the budget highlights without caring much about what any of it means. But I dunno, that seems like a particularly low view of journalism to me? Amazon is a subject of global interest, and should be written about seriously and sanely everywhere.
[7] From a legacy perspective, reporters didn’t write their own headlines for logistical reasons. The editor laying out the paper in the evening would write them based on emerging spatial constraints. The reporter had long gone home and had no premonition of available character-count. But this is less true now, as lots of articles are web-first/only, and web/print headlines generally don’t have to match.
[8] Theoretically, the 3:05pm edit could have been just correcting the typo, with the extra tweet being added in at some other point in the five-ish hours between Derek’s screenshot and the next cached copy of the article (6:29pm). And, theoretically, that separate edit could have been made independent of seeing / hearing about Derek’s tweet. But even in this very generous scenario, the same basic problem remains: the article was updated with unmarked edits to support a bad/misleading subhead.
[9] The attending fire chief, Palmer Buck, head of the Woodlands Township Fire Department, allowed that someone working for the County Fire Marshall might have made such a call. But this is all academic as Buck’s team were the first responders, and based on how fast they got the fire out it’s not clear what occasion anyone could have had to make such a call. The best I can come up with is a proactive dispatcher who wasn’t sure what Buck’s crew may or may not have known. But it feels unlikely.