Bluesky's quest to be the next Twitter | The Vergecast
15 Nov 2024 (22 days ago)
- The Vergecast is the flagship podcast, and the hosts are discussing various tech-related topics, including a product that might set the tech industry on fire (16s).
- The hosts, Eli and David Pierce, will be joined by Kylie Robinson later in the show to discuss several news topics, including whether AI is real or just a fake idea (32s).
- The conversation also touches on the Trump Administration, with a humorous mention of how weeks might be counted, referencing George Washington (43s).
- Elon is already working on various projects, including a fake Department of government efficiency, which the hosts think is worth discussing (54s).
- There are also rumors about a fake wireless carrier called Boost Mobile, which is actually a real carrier, and the hosts are excited to talk about it (1m1s).
- Additionally, there are rumors that Apple might be making a wall tablet, and the hosts will discuss garage door-related topics as promised (1m7s).
- The conversation will start with what's happening on social media (1m13s).
BlueSky vs Threads (1m14s)
- The concept of the open Social Web and the Fediverse has been discussed, where users can move around and aren't locked into one platform, with protocols like ActivityPub and Protocol being used by various platforms (1m16s).
- Threads from Meta uses ActivityPub, but has been slow to support people, and users have issues with its moderation, leading to a decline in popularity (1m30s).
- After Elon Musk's changes to the X algorithm, users are fleeing the platform, but instead of going to Threads or Mastodon, they are moving to BlueSky, a different federated social platform that uses the Protocol protocol (2m0s).
- BlueSky has gained popularity, with users including AOC, who caused a mini uproar when she started posting again, and people trying to make "sking" happen instead of calling it posting (4m9s).
- BlueSky has hit mainstream critical mass, with over 275 million users, and is growing rapidly, with plans to turn on ads next year, making it a big-scale business (2m44s).
- BlueSky has done things right in building the platform, thinking about identity and discovery, and building products like lists and messages, making it dramatically better than competitors (3m23s).
- The platform was initially popular with people who wanted to be "weird" on purpose, but it has now gained mainstream appeal, with people who don't want any part of that coming onto the platform (3m48s).
- BlueSky is now considered the winner in social media, with the "vibe shift" really mattering, and it has gained the "juice" in terms of popularity and appeal (2m20s).
The Algorithmic Dilemma (4m38s)
- The algorithmic feed on platforms like X and Threads can create an echo chamber where users are only shown content that aligns with the platform's or owner's views, rather than a diverse range of perspectives (4m40s).
- This algorithmic collapse can lead to a platform becoming less appealing to mainstream users, as seen with X, which is now more similar to platforms like Truth Social and Rumble (5m16s).
- Elon Musk's approach to free speech on X has been hypocritical, as he initially presented himself as a free speech warrior but has since restricted the types of speech allowed on the platform (5m36s).
- Threads, like X, is an algorithmic platform that prioritizes engagement bait over meaningful content, making it difficult for users to find valuable information (5m47s).
- The default algorithmic feed on Threads can be overwhelming, and even when users opt for a reverse chronological feed, they may be switched back to the algorithmic feed at any time (6m2s).
- Platforms like X and Threads do not serve the same role as Twitter and are not suitable replacements, as they prioritize different types of content and user experiences (6m16s).
- Elon Musk's purchase of Twitter, despite being a questionable business decision, has given him significant influence and made him a powerful figure in the United States (6m58s).
- The impact of X on Elon Musk's influence is debatable, as it is unclear whether the platform's algorithmic feed or his own wealth and spending were more significant factors in his rise to prominence (7m21s).
- Creator platforms like Threads and X are closed ecosystems designed to promote engagement within their own platforms, suppressing links and using algorithms to prioritize content, with Threads being a closed platform despite its potential for growth and integration (7m41s).
- Threads has established its identity and purpose, and its features, such as algorithmic feeds and moderation, are unlikely to change, with users being expected to adapt to its closed ecosystem (8m30s).
- Bluesky, on the other hand, is an open platform with 15 million users, allowing links, offering customizable algorithms, and featuring composable moderation, with the goal of being a more open and flexible alternative to closed platforms (9m16s).
- Bluesky's protocol allows for user-level federation, enabling users to take their accounts to different servers, and its moderation is focused on individual user accounts rather than servers (10m9s).
- The user account is the core element that can be moved between servers, allowing for greater flexibility and control, which is seen as a key advantage of Bluesky's approach (10m41s).
- One of the main challenges facing Bluesky and the broader fediverse is the issue of identity and username management, with users struggling to establish a consistent and meaningful online presence (11m2s).
- Blue Sky's ActivityPub is a crucial part of the system, allowing users to type their username into a bar and load up their profile, which is a powerful feature (11m22s).
- Despite its potential, Blue Sky is currently not as open as it claims, and it needs to build the next thing to achieve its goals (11m32s).
- Blue Sky is considered more truthful about its intentions than Threads, with a focus on not doing news, which is seen as more trouble than it's worth (11m56s).
- The current social media landscape has an imbalance, with some platforms showcasing one view of America and being explicitly in favor of one candidate, while others have nothing (12m9s).
- People are leaving platforms that don't align with their views and are looking for alternatives where they can make the platform what they want, which Blue Sky allows with its built-in tools and APIs (12m39s).
- A platform like Deck Blue, similar to TweetDeck, already exists for Blue Sky, allowing users to build tools and customize their experience (12m55s).
- The future of social media may involve different platforms being able to talk to each other, with the protocol not mattering as much, but this is still a long way away (13m21s).
- The best outcome would be for platforms like Mastodon, Blue Sky, Threads, and others to be able to communicate with each other seamlessly, allowing users to choose the platform that best suits their needs (13m24s).
- The development of the fediverse and ActivityPub has been ongoing for years, and it's understood that this process will take a long time, but it's still frustrating to see progress being slow (14m7s).
- The gap between different platforms and protocols is getting wider, which is problematic, but it's also possible that this will eventually lead to a more open and interconnected social media landscape (14m31s).
- Threads, a platform, is building out strong activity Pub integration, but as it becomes more successful, building that integration becomes less compelling, especially when considering being a good steward of an open ecosystem (14m36s).
- The people working on Threads are serious about wanting to be a good steward of an open ecosystem, but that may not be enough to achieve their goals (15m7s).
- The Threads team is taking a different approach to building their platform by adding one piece of federation at scale immediately, which is a different way of going about it compared to other platforms like Bluesky (15m48s).
- Bluesky built all their features and struggled to scale them to even 15 million users, which is why they had invites at the beginning and a feature called "hell thread" (15m29s).
- The Threads team is adding features one at a time, such as allowing users to see who follows them in the fediverse, but the loops haven't closed yet, which is driving people crazy (15m50s).
- The fact that users can't reply to people from the fediverse who reply to them is an actively bad experience and makes the platform less compelling (16m25s).
- The current state of the platform is bad product, and the longer it stays that way, the less compelling it becomes to users (17m2s).
- The emergence of Bluesky as a competitor to Threads is good news, as it brings competition and cooperation to the market, which is beneficial for the development of open platforms (17m15s).
- Competition is good, and it's what should be rooted for, even if the right answer is a combination of competition and cooperation (17m49s).
- A hypothetical social media platform could be created by starting a server and allowing users to follow its feeds, with the option to curate content from other places using ActivityPub, a decentralized social networking protocol (18m2s).
- This approach would allow the platform to own its distribution and avoid relying on algorithms from other companies like Google or YouTube (19m2s).
- The goal is to lean into federated distribution, which is the next phase of the internet, and to use protocols that enable decentralized social networking (19m23s).
- There are several protocols competing to become the standard for decentralized social networking, but ActivityPub is currently ahead in terms of adoption and support (19m46s).
- Bluesky, a social media platform, is currently winning in terms of user adoption, as it has made the bet that users want a platform like Twitter without the problems (20m1s).
- Twitter was a complex platform that accomplished many goals, but it was also poorly run and had many problems, and now other companies are trying to recreate parts of it (20m23s).
- However, no company is trying to recreate the full Twitter experience, and it's unclear whether that would be a good thing for the world (20m50s).
- A better approach might be to reconstruct the Twitter experience, but with government agencies and other organizations owning their own servers, so that they are not reliant on a single platform (21m23s).
- The idea of independent creators owning their own tools is gaining traction, with some considering alternatives to existing social media platforms (21m39s).
- The control of speech in America is largely in the hands of a few individuals, rather than the government, with three key figures having significant influence (22m0s).
- Twitter was able to strike a balance between moderation and free speech, but this balance has been disrupted by changes to moderation and algorithms (22m20s).
- The dismantling of Twitter's original structure has led to a decline in its effectiveness, with some comparing it to a "clown car that fell into a gold mine" (22m51s).
- The First Amendment in the US prevents the government from passing content moderation rules, making competition a necessary factor in shaping online discourse (23m5s).
- The potential appointment of Brendan Carr as FCC chairman could have significant implications for content moderation, as he believes the First Amendment should keep the FCC out of this area (23m18s).
- The current moment, with people considering alternative platforms like Bluesky, may point to a future where protocols and user attention drive better outcomes (23m46s).
- Twitter's decline and the rise of alternative platforms like Bluesky suggest that Twitter is not a monopoly, and that users have the power to vote with their attention (24m10s).
- The shift away from Twitter may be a temporary blip, but it highlights the challenges of leaving established platforms and the difficulties of building new ones (24m20s).
- The novelty of new platforms can wear off, making it hard to sustain user engagement, as seen with Mastodon and potentially Bluesky (24m36s).
- The appeal of Bluesky may be due to its perceived values and the influence of its creators, with some drawing comparisons to Steve Jobs' quote about Microsoft and "taste" (24m56s).
Introducing Show and Tell (35m55s)
- The Washington Post is a trustworthy source for news, offering substantial and considered coverage on various topics, including Capitol Hill, the economy, and climate change, and can be subscribed to for 50 cents per week for the first year at washingtonpost.com/verge (35m55s).
- Life 360 is an app that helps ease holiday stress by allowing users to keep track of their family members' locations in real-time, streamlining family communication and coordination, and providing safety features such as crash detection and roadside assistance (36m51s).
- AT&T offers a service called AT&T Next Up Anytime, which allows users to upgrade to the latest iPhone every year, and is currently promoting the new iPhone 16 Pro with Apple Intelligence (37m56s).
- Kylie Robinson is a guest on the show, and is participating in a new segment called "Show and Tell" (although the name is not final and listeners are encouraged to suggest alternative names) (39m1s).
- Apple is reportedly working on a smart home controller, which is essentially an iPad designed to live inside a house, with features such as a touch interface, camera, rechargeable battery, internal speakers, and Siri digital assistant integration (40m22s).
- The device is expected to be marketed as a way to control home appliances, chat with Siri, hold intercom sessions with FaceTime, and access various Apple apps, including web browsing, news updates, and music (41m13s).
- The device will be available in silver and black options, with a size comparable to two iPhones side by side, and a thick edge around the display (40m48s).
- Apple has designed different attachments for the device, including wall mounts and bases with additional speakers, which can be placed in various locations around the house (41m46s).
- The FaceTime feature is envisioned to be used for video conferencing during work meetings or while cooking (41m57s).
- The idea of an accessory ecosystem, where users can add new features to their iPad by docking it on different docks around the house, is seen as a promising concept, but the execution with a 6-inch iPad is considered flawed (42m17s).
- Apple's track record with accessory ecosystems is criticized, with the example of smart keyboard covers for the iPad being cited as an area where the company has struggled to provide a wide range of options (42m36s).
- An Apple designer visits the Foxcon factory, which sparks a number of ideas (42m56s)
Apple Smart Home rumors (43m0s)
- Apple is reportedly working on a new iPad that can be docked on a speaker, with a customizable home screen and widgets for checking information, but it will not have an App Store, at least initially (43m0s).
- The device is expected to have a newly designed operating system and a feature for quickly launching favorite software icons, but it is essentially an iPad without the App Store, making it a less desirable product (43m29s).
- Apple's plans for a smart home product may face challenges due to the complexity of integrating smart devices and the debate over whether the "smart" part should be in the device or the switch (44m38s).
- The idea of a smart home product is fascinating, but it can be difficult to implement, as evidenced by the challenges of using smart light bulbs and switches (44m49s).
- Instead of creating a new device, Apple could consider adding a dock to an existing iPad to give it an always-on screen and smart home capabilities, but this might not be a viable option due to pricing considerations (45m36s).
- There are rumors that the next version of the device will have a robotic arm and cost $1,000, but it is unclear if this will be a successful product (46m6s).
- There is a case to be made for turning an iPad into a smart home device, as it is already often used as a homebound device, but it remains to be seen if Apple will pursue this idea (46m14s).
- Apple is planning to reinvent the wheel by creating a new product instead of improving existing ones, such as the iPad, which could be enhanced with an accessory ecosystem to make it more powerful and cool (46m23s).
- The new iPad is expected to have improved features, including Siri, which will allow users to perform automations, but it may not always work as intended (46m47s).
- Users may still need to Google answers to their problems, and the solution might be as simple as restarting their iPhone, which is a common issue with Apple's home products (46m55s).
- Apple is investing in AI technology, including Siri and Apple Intelligence, but it's unclear what practical uses these features will have (47m9s).
- Other companies, such as Google and Amazon, have also developed AI assistants, but their usefulness is limited to simple tasks, such as controlling lights (47m20s).
- Listeners responded to a previous conversation about using Siri to open garage doors, pointing out that it's not a practical or widely used feature (47m35s).
The Future of AGI: Hype vs. Reality (47m47s)
- The concept of scaling laws in large language models (LLMs) is being discussed, where it was previously thought that adding more compute and data would make the models bigger, better, and smarter, but recent reports suggest that this approach is no longer working (48m30s).
- Companies like Google, Open AI, and Anthropic are finding that their models are not improving with the addition of more data and compute, and are instead plateauing, causing frustration and concern among researchers (49m0s).
- The idea of using synthetic data to train larger models has also been explored, but this approach has not yielded the desired results, with models becoming confused by the low-quality data (50m24s).
- The failure of scaling laws has significant implications for the development of AI, as it suggests that simply adding more resources will not lead to significant improvements, and that a research breakthrough will be needed to achieve the next level of progress (50m38s).
- The situation has been described as a major challenge for AI companies, with significant financial investments at stake, and some researchers are feeling vindicated in their skepticism about the scalability of AI (50m50s).
- The ongoing narrative from AI companies that they will continue to improve their models until they reach a point of near-omniscience has been called into question, with some critics arguing that this approach is unrealistic and that a more nuanced understanding of AI development is needed (51m10s).
- Bluesky is a very capable platform, but its purpose and what people will do with it are unclear, and it's likely that the company will be stuck in the "what can you do with this" phase for a while (51m26s).
- Dario Mod from Anthropic claims that artificial general intelligence (AGI) will be achieved in two years, while Sam Altman thinks it's possible to achieve AGI with current hardware (51m52s).
- The idea of achieving AGI in two years or with current hardware is incompatible with the fact that bigger models aren't more capable than previous ones (52m32s).
- There's skepticism about trusting companies like Anthropic and their claims about AGI, as they need to sell products and make money (52m49s).
- Sam Altman recently did a Reddit AMA where he mentioned that it's hard to release new models in parallel due to limited compute, and that the next big thing might be agents (53m3s).
- It seems like companies are pivoting and trying to distract from their failures in making the next big model, and instead focusing on selling products and getting more money (53m20s).
- The goal of combining different models to achieve AGI might be more about selling products than actually achieving AGI (53m49s).
- Dario Mod's quote about achieving AGI in 2026 or 2027 is based on a straight-line extrapolation, which is unlikely to continue forever (54m4s).
Grubhub's New Owner: The Rise of Wunder (54m22s)
- A discussion about exponential growth and Moore's Law highlights that just because something has grown at a certain rate in the past, it doesn't mean it will continue to do so, and there's no guarantee that any technological advancements will work as expected (54m38s).
- Sam Altman's efforts to build trillion-dollar data centers to power digital advancements have led to competitors trying to figure out how he's cooling the chips, with one instance involving a crop plane flying over a data center to gather information (55m10s).
- The idea of building digital god or artificial general intelligence (AGI) is met with skepticism, with the argument that if it were possible to build AGI with current hardware, the tech industry should focus solely on that goal, but this claim is seen as unrealistic (56m51s).
- A humorous take on the idea of AGI designing the next iPhone or writing about Donald Trump for the next four years is presented, highlighting the absurdity of relying solely on AGI for such tasks (57m9s).
- The news of Just Eat selling GrubHub is discussed, with the observation that the words "Just Eat selling GrubHub" sound nonsensical, and the fact that GrubHub is a well-known delivery service that also owns Seamless (57m21s).
- The regional differences in food delivery services are noted, with GrubHub being popular in New York, while Postmates is more commonly used in LA (57m52s).
- In San Francisco, food delivery services like DoorDash are popular, but the business model is changing with companies like Wonder, which licenses recipes from famous chefs and sells them through a ghost kitchen and food hall model, allowing for industrial-scale production and delivery through services like GrubHub (57m54s).
- Wonder's business model involves licensing recipes from famous chefs, such as Bobby Flay, Marcus Samuelsson, and Detroit Brick Pizza, and then producing the food in a large commissary kitchen, which is then reheated and assembled in a food hall or delivered to customers (58m23s).
- The chefs do not actually make the food, but rather license their recipes to Wonder, which then produces and sells the food through its platform, creating a commoditized version of the chefs' recipes (59m10s).
- Wonder is owned by a part-owner of the Minnesota Timberwolves and a former Walmart VP, and is planning to buy GrubHub, which would allow it to expand its business model to an even larger delivery system (58m16s).
- The model is compared to a Vegas casino, where restaurants are run by famous chefs, but the food is not actually made by them, and the experience is more about the name and branding than the actual food (1h0m5s).
- The rise of this business model is seen as the end point of a trend that started with cloud kitchens, where food is produced in a central location and then delivered to customers through apps and services (1h1m13s).
- The concept of luxury in consumerism has reached a new level of absurdity, where people want every part of a process to feel fancy, even if it's not actually high-quality (1h1m37s).
- The idea is compared to ordering food from a high-end restaurant, but having it delivered in a luxurious vehicle, such as an Uber Black, rather than a more modest option like a Nissan Ultima (1h1m49s).
- This desire for luxury is exemplified by people joking about not wanting their food delivered by DoorDash, but rather by a more upscale service (1h1m45s).
- The comparison is made to Carbone, a high-end restaurant, and the idea that even a meal from a commissary could be made to feel luxurious if delivered in the right way (1h1m55s).
- The ultimate goal is to make every part of the process feel expensive and luxurious, even if the actual product or service is not (1h2m11s).
- The example of ordering steak from a delivery service is given as a way to illustrate this desire for luxury and high-end experience (1h2m17s).
- The concept of a "Golden Corral buffet" is used as a symbol of the endpoint of this system, where luxury and excess have reached a peak (1h1m39s).
- The idea is also linked to the excesses of "Donald Trump's America", where every restaurant is compared to Carbone, implying a culture of luxury and over-the-top consumerism (1h1m42s).
- The conversation starts with a comparison of food delivery services, with the mention of a $7.3 billion acquisition of GrubHub by Just Eat Takeaway, which is now being sold for $650 million, resulting in a significant loss over four years (1h2m46s).
- The discussion touches on the idea that food delivery companies like Uber, DoorDash, and others have claimed that shipping food is a higher-margin business than giving people rides, but the reality seems to be that these are reasonably successful yet low-margin businesses (1h4m1s).
- The business model of buying monopolies and raising prices, as seen in companies like Uber, is not working as expected, with VC dollars unable to sustain these companies (1h4m41s).
- The conversation also mentions the challenges of food delivery services, including low margins and the difficulty of sustaining a profitable business model (1h3m32s).
- The topic of smart home security products is briefly mentioned, with the Home Depot offering a wide selection of products for easy control and automation of homes (1h5m23s).
- The Home Depot is a one-stop shop for smart home products, offering a wide range of options for both experts and newcomers, and can be visited online or in-store to find the right products for one's home (1h5m53s).
- Legal Zoom is a platform that helps users create and register their own businesses, providing necessary tools and services to launch, run, and protect a business, and is offering a 10% discount with the promo code "Verge 10" (1h6m12s).
- Fundrise is a company that offers a more than $125 million Innovation Fund, which holds some of the most exciting pre-IPO tech companies in the world and is designed specifically for individual investors, allowing them to get in early on companies that were previously out of reach (1h7m26s).
- The lightning round is a segment of the podcast, and the host has been trying to come up with a journalistic reason to attend the Monaco yacht show, which is now a top priority (1h8m36s).
- The host jokingly mentions that they have mellowed with age and now just want to be on a boat, and that being sober on a boat seems great (1h9m19s).
- The host introduces the lightning round, which features three selected topics, but notes that it's not a typical lightning round (1h9m24s).
- Mark has made a sponsored video, which has sparked controversy and discussion about ethics and disclosure, with some people questioning whether it should be disclosed that Marquez's podcast is part of the VOX media podcast network (1h9m35s).
- The VOX media podcast network is an ad sales network that helps podcasts with scale to secure better advertising deals, but the conflict of interest does not exist in this case as the sales team is separate from the newsroom (1h11m5s).
- The newsroom has a wall between them and the sales team to maintain editorial independence and freedom, allowing them to cover companies critically without fear or favor (1h11m55s).
- The company's ethics policy prioritizes editorial independence, and disclosures are made where there's a plausible conflict of interest, but the question remains of when and why to disclose certain information (1h12m33s).
- The company is transparent about potential sources of conflicts and has no problem with disclosures, but the issue is determining the appropriate time and reason for disclosure (1h10m38s).
- The company's coverage often loses them money due to companies being mad at their coverage and walking away from deals, but this is seen as a necessary cost for maintaining editorial independence (1h12m0s).
- Comcast NBC Universal is an investor in the company and has a seat on the board of directors, which could potentially lead to a conflict of interest in the newsroom, although this has never happened in the 13 years of operation (1h12m39s).
- The company discloses conflicts of interest, such as the fact that the CEO made a Netflix show, which could influence coverage of Netflix (1h13m16s).
- There is a potential conflict of interest in covering Bluesky, as the CEO and David know many of the people involved, including Marquez, and are friendly with them, but this is not considered a business conflict (1h13m50s).
- Disclosing all commercial and corporate interactions could threaten editorial independence by requiring the newsroom to be aware of all company deals, which is not desirable (1h14m10s).
- The company has a media podcast network, including deals with Scott Galloway's Prof G shows, but does not own or produce them, and only discloses deals where the company is directly involved (1h14m35s).
- The company runs a programmatic ad tech stack called Concert, which serves ads on many publishers, but disclosing every partner would be impractical and could compromise editorial independence (1h14m55s).
- The separation between the business and editorial sides is important to maintain editorial independence and prevent reporters from being influenced by business considerations (1h15m22s).
- The media landscape has changed, with many creators and influencers blurring the lines between advertising, sales, and investments, making it necessary for them to disclose their commercial relationships to make a living (1h16m5s).
- In contrast, the independence of the newsroom is a key aspect of the product being sold, and disclosing every single potential conflict of interest could defeat the purpose of the rule and undercut the promise of independence (1h16m44s).
- The company discloses its relationship with Comcast, but extending the rule too far could lead to a situation where the newsroom is told who to be worried about every morning, which would be counterproductive (1h16m53s).
- The goal is to rebuild trust in journalism, and the audience's high expectations are welcome, but it's also important to point out that other media outlets and creators should be held to the same standards (1h17m46s).
- President-elect Donald Trump has announced that Elon Musk will be leading a department of government efficiency, which will aim to cut wasteful government spending and feature a leaderboard on X of the dumbest ways the government spends money (1h18m27s).
- The announcement of the department of government efficiency, led by Elon Musk and V Ramaswami, has raised questions about its legitimacy and effectiveness, with some wondering if it's just a meme or a real initiative (1h19m0s).