Gigglebytes?

I don't think it would be difficult to get the IMV up to compliance with US regulations. If they're selling it in Mexico, it'll be required to have airbags. The hood looks long enough to have engineered in proper crumpling in a crash. Things like backup cameras might require a little bit of retooling, but that's not actually super expensive compared to the other expenses of officially bringing it in: the 25% import tax, a parts and service network, etc.

So it's a business decision not to even try to get it into the U.S., informed by those regulations.

In contrast, something like a kei truck wouldn't be easy to get street legal as a new car in the US: no crumple zone and higher center of gravity are more fundamental safety issues that can't easily be engineered around.

At the same time, expecting privacy in a room where a bunch of strangers hang out is already unreasonable. If everyone already in the channel can log the chats, for example by idling in the channel, then adding E2E on top of that is probably a false sense of security.

Userland

How much does a banana cost, $30,000?

YouTube serves probably dozens of formats/bitrates, and has spent years tweaking how it ingests, transcodes, and serves videos. Adding in-stream ads might have been a bigger engineering task in that environment. Depending on the percentage of users/viewers avoiding ads, it might not have been worth the return.

But I can't seem to find any data on actual Spotify viewership to pin these numbers down to be able make some apples to apples comparisons.

There are several third party research firms, most notably Edison, that compile their own data from the client side, akin to how Nielsen produces cross-platform numbers for television and streaming. And others compile data from different sources, including published ordinal rankings and publicly reported view counts, other metrics like mentions/shares on social media.

Podcasting historically predates always-online devices, so a lot of the interfaces count downloads rather than listens, as the tech companies try their best to push things towards data-tracking proprietary client apps. So it's a non-trivial effort (usually requiring lots and lots of paid survey participants, which also introduces some statistical skew), to get listener data.

So there is data out there, almost all behind business-to-business paywalls/subscriptions. Those closely guarded datasets are what advertisers use to price ad campaigns on these channels, for example, where they have to decide whether and how to spend on one particular type of medium versus another (e.g., product mentions on podcasts versus YouTube versus twitch versus traditional television or film).

Also do people not "watch" podcasts?

The NYT article I linked describes the issue somewhat, that there's a difference between viewing as a first screen or second screen, or listening as primary, or listening as background, and that advertisers and publishers are aware of how different shows perform at different functions. They describe the example of the political podcast, Pod Save America, having stats showing that about 20% of their audience is through video rather than audio-only.

As I understand it, through my own personal experience, podcasts are popular during commutes, as a primary audio experience for people driving alone in their cars or wearing headphones on the bus/train/sidewalk. Those are functions in which video is not desired, so "watching" would be a strange way to describe how people primarily consume the content, through listening only. And while I know official music videos get a lot of watches on YouTube, I don't think counting those watches are a good way to rank or analyze how people "listen" to music overall.

It's also why many of the most popular podcasts simply don't do video (NYT's popular The Daily), and stem from radio roots (the NPR podcasts, a bunch of iHeart or Sirius-published podcasts). Video isn't a good proxy, especially for the podcasts that are tied exclusively to a non-YT platform.

Maybe SiriusXM overpaid for their deal, at $40 million per year. Maybe Spotify overpaid for their previous deal, at $30 million per year. But those numbers alone aren't anything to scoff at, and in my opinion are the best proxy for cross-platform comparison: Smartless got about $33 million per year from SiriusXM, Dax Shepard got $27 million per year from Spotify, etc.

And maybe you're right that podcasting itself isn't ever going to be as big as video or streaming. The South Park guys got $900 million for 6 movies on Paramount+. Apple TV dropped more than 9 figures each on Killers of the Flower Moon and Napoleon.

But for advertisers, that specific podcast is a valuable property. I can't comment on the quality or cultural comparisons because I literally have never heard any part of it (or other huge genres like true crime podcasts or celebrity podcasts or video game reviews or whatever other things people consume), but I watch the business side of things because I'm interested in that.

And I'm not trying to give any commentary on the cultural relevance of this particular Harris interview any more than I'd give commentary on her appearance on Howard Stern (who I might not have heard about in like 20 years) or Walz's appearance in World of Warcraft or something I don't actually understand the logistics of. Politics isn't really something I care about, but tech and media businesses are.

Yeah, looks like a series of voluntary tags in the metadata. Which is important, and probably necessary, but won't actually do much to stop deceptive generation. Just helps highlight and index the use of some of these AI tools for people who don't particularly want or need to hide that fact.

Are you required to subscribe through Prime or can I add any subscription?

I think it only works with services subscribed through Amazon.

Yeah, Google does the same thing, but it launches the specific app in order to play from that screen (and not every app handles that gracefully).

What's interesting about Amazon is that they play the actual content in the same app, even when licensed through a subscription with another service.

I imagine the first image file would've come well after the first text transmission, and would've required the development of a binary file transfer protocol. Maybe they're talking about XMODEM and the BBS era, before those BBSes were eventually networked onto the internet.

I thought we were having a conversation about tech/entertainment/media companies and their audiences, but I guess you were ready to turn this into something about parasocial dynamics and American electoral politics, neither of which is a particularly interesting topic for me, and honestly wasn't even part of the conversation I thought we were having.

I guess that's on me for commenting in a politics community I'm not subscribed to, but your original comment at the top of the thread was about something I am interested in.

Either way, you're wrong on the audience metric stuff. I probably should've picked up on your grudge against podcasting as a medium when you kept using the verb "watch" to describe how people consume podcasts. Which also probably is why you insist on using YouTube as some kind of proxy for podcast reach or popularity. I disagree with your method of analysis, and I think I've made that pretty clear. But you do you, and go on believing that you've got your finger on the pulse of how all audiences consume content on the internet.

I'm hopeful that we'll eventually be able to separate the app itself from the license to view certain content. Some apps are shit with good content, and some apps are great with insufficient content.

I'd love to be able to have a one-stop shop for browsing or searching for shows and movies that I have access to, whether it's through my Netflix/Amazon/Disney/Hulu/Peacock/Apple/Paramount subscription or whatever. Amazon seems intent on trying to make it happen, but I'd personally prefer that the actual device (Apple TV, Chromecast or whatever it's replacement is called, Roku, XBox, Samsung TV) to just have it on the home screen when you turn it on.

I really, really don't get why you're digging your heels in on this.

There are no viewership numbers available from spotify. If we use the YT channel as a proxy, its a mid-teir audience podcast.

So you're going to ignore the published charts of the two most popular podcast apps (Spotify and Apple Podcasts), and the dominant market research firm in podcasting (Edison) to try to come up with your own proxy methodology based on a video platform? It's a bizarre approach of "do your own research."

And if you want to understand why video views are a poor substitute for actual audio stats, this article from 6 months ago does a deeper dive into the rise of video supplement to podcast audio, talks about the different approaches (and wildly different distribution/audience ratios depending on specific styles and audiences). It says that 16 out of the top 30 are publishing video now, an increase from before, but also obviously means that 14 out of the top 30 don't do video at all.

I went to YouTube Music to browse its podcast charts, where Call Me Daddy was nowhere to be found on the charts. Digging further, it's because the full podcast episodes aren't even available through YouTube. Instead, it's 7 minute preview clips. In other words, the SiriusXM deal bringing it out of Spotify Exclusive hasn't even trickled over to YouTube yet (or SiriusXM doesn't want to publish it on YouTube for whatever reason).

So digging into YouTube viewer counts is like trying to argue against published box office rankings based on trailer view counts on YouTube. It's a bad methodology, and looks like someone flailing to try to confirm their feelings against what the mainstream media is reporting.

all the noise about podcasts and Spotify trying to control that media was a couple years ago

The $125 million SiriusXM deal to bring it back out of Spotify exclusive was less than 2 months ago.

Like, I get if you're not the target audience, because I'm not either, but it's strange to try to argue that it isn't a huge audience or a heavy hitter in the podcasting space.

You can't say that, though, because it implies Chinese engineers and information technology scientists are trailblazers rather than plagarists and IP thieves.

I mean, I said what I said and I meant it. The Chinese are trailblazing a path nobody has tried before: DUV only for sub-10nm processes. It's not ideal, and the reason why nobody did it before is because they already had EUV by the time they got there.

But I wouldn't sleep on the ability of anyone to be able to solve problems using the tools at their disposal.

Especially since there's nothing stopping the mainland Chinese companies from hiring Taiwanese engineers.

Not a ton of people believed that Taiwan could surpass Japan, either, but that happened in the 90's. Not a ton of people believe that Japan can get back in the game, but Rapidus is making a play for 2nm.

Nothing is forever, and things are always changing. I'm somewhat optimistic that western sanctions will keep China from competing on the world stage at semiconductor fabrication, but I don't think it's a guarantee or in any way inevitable.

Meanwhile, you've got companies in Taiwan, Korea, China, and Japan breaking into the 3nm and 2nm scales.

The mainland Chinese SMIC is doing everything they can without access to ASML's EUV machines, and have gotten further than anyone else has on DUV. It remains to be seen just how far they can get without plateauing on the limits of that tech. Most doubted that they could get past 10nm, but some of their recent chips appear to be comparable to 7nm, and there are rumors that they have a low yield 5nm process that isn't economically feasible but can be a strong political statement.

TSMC is delaying the transition to Gate All Around, announcing that they won't be trying it on the 3nm processes, and waiting until 2nm to roll that out. They're the undisputed leader today, so they're milking their current finFET advantage for as long as it will sustain them.

Samsung has already switched to Gate All Around for their 3nm process, so they might get the jump on everyone else (even if they struggled with the previous paradigm of finFET). But they're not lining up external customers, as their yields still can't compete with TSMC's. It's entirely possible though that as the industry moves from finFETs to GAAFETs, Samsung could take a lead.

Intel basically couldn't get finFETs to work, and are already trying to skip ahead to GAAFETs (which they call RibbonFET). Plus Intel (like the others) is trying to introduce backside power delivery, which, if it can be commercialized and mass produced, would achieve huge gains in power efficiency. Intel did introduce both technologies in its 20A process (supposedly 2nm class), but then canceled it due to low yield. So they're basically betting the company on their 18A process, and hoping they can get that to market before TSMC and Samsung hit their stride on 2nm.

So your market research has been to ask two women who don't listen to podcasts, and then look at the view counts on a video platform not primarily used for audio podcasts. Real solid methodology there.

Just, like, look at any of the podcast metrics ranking podcasts by listens or subscribers from the US: #2 on Apple, #5 on Spotify, #4 on Edison's charts.

It's a top-5 podcast in terms of popularity, and probably literally the most popular podcast among women. It was, for a time, the #2 podcast after Rogan's.

Spotify paid $60 million to make Call Her Daddy a Spotify exclusive for 2 years, and last year SiriusXM paid $125 million for the ad/distribution rights for 3 years.

It's a heavy hitter in the podcast space, by pretty much every metric.

But they won't go bankrupt because the US gov./army need Intel to stay relevant against China, and Intel is basically the only American company that both designs and fabs their own processors and that is still relevant.

That's never stopped anyone from going bankrupt and wiping out shareholders. If the tech is that critical, the US government might engineer a bailout of the company, but will make the shares worthless in the process.

They did it with GM and Chrysler in 2009, they did it with Iridium in 2000.