Welcome to Disconnect’s first weekly roundup! This will be something I’ll use to update you on some of the stories of the past week, but also topics I find interesting but might not be ready to write a full newsletter about. Once March rolls around, these will be for paying subscribers, but this month you’ll get a taste of what to expect!
In this week’s issue, we’re digging into why there are so many private space companies right now, how tech companies are trying to capture podcasts so they can squeeze profit out of them, and why ChatGPT isn’t as novel as we might think. Plus, a ton of other interesting reads you might enjoy and a fun video to close things off.
Obviously, we also had a bunch of tech companies report earnings this week, but I decided I didn’t want to wade into that here. The short of it: profit is down, they’re searching for a new model, more cuts are coming, and they’re doing all they can to keep shareholders happy and wealthy — regardless of what that means for workers. Bring on the “year of efficiency”? 😕
Private space companies are a policy choice
I finally read an FT piece from December about the private space industry and the growing race to capture the opportunities of outer space to extract profit. I thought it had an important set of details that you don’t often see discussed in articles about SpaceX and Blue Origin (among many others), but they’re important to understand why this whole billionaire space race is even happening in the first place.
While space exploration used to be about state-backed programmes focused on national security, national pride and scientific research, the US started pulling back on centralised government control of space after two fatal space shuttle accidents (the Challenger in 1986 and Columbia in 2003). This led a presidential commission on US space exploration policy to conclude in 2004 that “Nasa’s role must be limited to only those areas where there is irrefutable demonstration that only government can perform the proposed activity.”
While public-private satellite programmes had existed since the 1960s, it wasn’t until the shuttle programme began winding down (and was ultimately cancelled in 2011) that the new crop of commercial space companies began to take off. Congress shifted funding incentives and created a new policy (the Commercial Orbital Transportation Services programme) to encourage privatisation. Nasa and other government agencies became customers of private space contractors, rather than creators or even supervisors of new technologies.
Counter to the usual narrative that these companies are the result of the entrepreneurial zeal of tech founders moving into a new industry or because we’ve reached a technological threshold where it makes sense for the private sector to take the reins, the truth is much more simple: the decades-long push to further privatize the actions of government, regardless of the consequences, finally came for NASA and the “final frontier” (of capitalist expansion).
The government faced a choice: would it make a greater investment in NASA so it could set some ambitious new goals to pursue and speed up work on whatever technologies would be necessary to achieve them, or increasingly hand that role over to the private sector with NASA taking more of funding role? No surprise, they chose the latter — not because it’s best for space science and maximizing the public benefit from space activities, but because it will allow a small industry of companies to make some big profits and get a bunch of public subsidies.
As the piece further explains,
A handful of new space companies could piggyback on Nasa technologies that took decades to develop, while the established contractors which helped build them lost out. Taxpayers who funded the basic research got no stake in the wealth being created by billionaires in space, the largest public commons of all.
The argument you often see made that private space companies were essential to improve on rocket technology is completely false. The public sector could’ve done that if it was given the proper resources, but where was the urgency in that moment? China wasn’t as significant of a threat to US hegemony and technological supremacy as it is today; there was no Sputnik moment to force the government to step up its game. Instead, the colonization fantasies of some rich guys were the driving force, so policy was shaped to accommodate them.
I’ve long argued the billionaire space race is a distraction from real problems and just this week I wrote about why we should be placing Starlink and these efforts to launch tens of thousands of additional satellites into orbit under much greater scrutiny. The privatization of space just so happened to coincide with the explosion of cult of the tech founder. It’s no wonder billionaires seized the opportunity to extend it to a cosmic level, but we should be much more skeptical of the final outcome and whether it has any real public benefit.
Can Big Tech enclose podcasts?
Among the earnings reported this week was Spotify, which lost €430mn on revenue of €11.7 billion — but one of the stories to come out of those revelations was how poorly its bet on podcasts is doing. During the pandemic, some tech companies decided they were getting into podcasts in a big way, especially as people were stuck at home, starting their own (like yours truly), and hopefully also listening to them.
Spotify bought Gimlet, Anchor, Parcast, and Ringer to try to expand from music — where it has to give 70% of all revenue to rights holders — and gain a dominant foothold in podcasts. But after hundreds of millions of dollars spent, the company’s CFO says podcasting “was a big drag on our business in 2022.”
I’ve always found this interesting (after all, I am a podcaster) because podcasting has largely evaded the enclosure pressures of some other mediums, where one or a few platforms basically carve out a dominant position and they mediate the entire relationship. It’s true that many listeners go to Apple or Spotify to listen to podcasts, but they don’t control the shows (at least in most cases) and that’s what they were hoping to achieve by making a more concerted push into podcasting.
Podcasters tend to upload to a distributor, then the open RSS feeds of their shows get accessed by a whole range of companies — including those of Apple and Spotify — and that’s how they arrive on the various podcast apps. Few podcasters upload directly to Apple, Spotify, and the others. Some shows have ads, but Apple and Spotify don’t control those relationships, and while they’ve moved into offering premium tiers, many shows still use external platforms like Patreon to manage that relationship with listeners.
So, if the previously strategy failed and Spotify isn’t trying to create hit podcasts that are only on its platform, what are they planning now? Honestly, it sounds pretty bad. According to Bloomberg,
The company will still fund original shows, but it’s more interested in being the platform that every podcaster uses than the studio or exclusive rights holder. … It wants to build advertising technology that resembles Google’s AdSense or AdWords for audio. Or, at a minimum, a version of the YouTube advertising business. It hopes that one day it will capture billions of dollars in advertising from podcasts, money it won’t share with music rights holders.
To me, that reads as though they’re still going to try to force shows onto Spotify, but it seems unlikely to work out as they plan. YouTube holds a dominant position in video and has always been the platform where you directly upload the content. Podcasting doesn’t work that way, and I see little incentive for that to change. Some distribution platforms have even been adding their own advertising tools.
One final point here: just because Spotify’s failing in podcasting doesn’t mean all’s well on the music side. A new FT report suggests Universal Music wants to see big changes to the streaming model to curb problems that have festered for years, reward artists who bring in new listeners, and add a “superfan” tier. As streaming video undergoes its own messy reality check, music may not be far behind.
ChatGPT is ELIZA all over again
When ChatGPT started to take off and people were marveling at how well it could write and respond to prompts, Joseph Weizenbaum’s 1960s ELIZA experiment immediately came to my mind. I was thrilled to see that Zachary Loeb dug into that history on his Librarian Shipwreck blog.
ELIZA was effectively a chatbot that “took on the role of a Rogerian psychotherapist with the human interlocutor taking on the role of the patient.” When a human interacted with ELIZA, it rearranged their words to turn their statements into questions — making it seem as though it understood them, when it really did no such thing. But seeing how people interacted with ELIZA made Weizenbaum a lifelong critic of AI.
Weizenbaum was well aware of the ways that programs like ELIZA could “behave in wondrous ways” and even “dazzle” yet he had also initially believed that “once a particular program is unmasked, once its inner workings are explained…its magic crumbles away; it stands revealed as a mere collection of procedures, each quite comprehensible.” […]
Granted, what largely transformed Weizenbaum into an outspoken critic of AI and computers was his revelation that even once the processes were explained many people still bought into the “illusion.” And what’s more that even many people who understood the inner workings of computers quite well could still get swept away as well. Weizenbaum observed that ELIZA demonstrated “if nothing else, how easy it is to create and maintain the illusion of understanding, hence perhaps of judgement deserving of credibility” an observation he followed up by noting “A certain danger lurks there.”
There’s a lesson there, especially as Silicon Valley has quickly thrown off the notion that web3 will amount to anything and is fully embracing AI as its Next Big Thing. Meanwhile, the stories abound about how these AIs aren’t really intelligent at all. Sure, they’re trained on larger datasets in the past thanks to all the computing power they have access to and an open web they can endlessly scrape for content, but that doesn’t mean they’re really “learning” in a way we understand for humans. We’re already seeing how they’re spitting out plagiarized writing and images that just ape a real artists’ style or still contain a messed up Getty Images tag.
These tools just recombine things they’ve ingested; they’re not making anything new. We should heed Weizenbaum’s 50-year-old warning before it’s too late. As Loeb put it,
The surprising thing about these revelations is not the revelations themselves, but that we still find it so surprising when they come out. In 1962, Weizenbaum had argued that the “success” of an “artificially intelligent” program hinged largely on its ability to fool people. And if a program could do a suitably passable job of fooling people at first, many of those people would then project their own fantasies onto the program, deepening their own commitment to the fantasy.
As we know with tech, there are always people who want to believe.
Some other things to read
- Brian Merchant kicked off his new column at the Los Angeles Times with a great explanation of how tech companies are using layoffs to discipline their workers.
- On that point, the FT reports that Amazon, Meta, Alphabet, and Apple have been hit with $10 billion in charges related to laying off 50,000 workers and other cuts they’ve made — but it’s all good, because since those cuts their market caps have collectively soared by $800 billion.
- One of the Twitter Files you won’t hear from Elon Musk’s chosen “journalists” is a leaked internal message that shows the Chief Twit ordering a left-wing activist be suspended at the same time he was opening the floodgates to previously banned extreme right accounts.
- The European Parliament backed the Platform Work Directive this week. Despite heavy industry lobbying, it would see “the presumption of employment for platform workers strengthened and enhance workers’ algorithmic rights.”
- John Herrman writes about the “junkification” of Amazon as it further embraces third-party sales which bring in more fees for the ecommerce platform and reduce antitrust risks. It sounds to me like it’s become more like AliExpress (or what I understood it to be).
- Chris Gilliard and Pete Rorabaugh reflect on the irresponsible rollout of ChatGPT: “This is a genie out of the bottle, we are told—never mind that at the end of most genie stories, they return to the bottle, having inflicted some manner of damage.”
- Charlotte Alter explains how effective altruism has a deeply toxic culture where women face sexual harassment and abuse, despite its attempt to cultivate a do-good image.
- Former FTX CEO Sam Bankman-Fried has been restricted from contacting former FTX and Alameda employees and from using encrypted messaging apps after concerns about witness tampering while he remains on bail at his parents’ house.
- Elon Musk’s latest scheme to monetize Twitter is to have businesses pay $1,000 a month for gold checkmarks. We’ll see if they buy in.
- While not a read, I spoke to Anousha Sakoui on this week’s Tech Won’t Save Us about the possibility of a writer’s strike in Hollywood this year, and how streaming has transformed pay and working conditions in the film industry.
A fun video to end with
Local Seattle band Who is She? was invited to play for a few nights at the Climate Pledge Arena, named after an Amazon greenwashing effort, but after criticizing Executive Chairman Jeff Bezos in their lyrics, they were told not to return for their final two nights. I’m sure it was just a coincidence!