The race to win AI — and hack iMessage

Transparency and simplicity are key values for Visible, a wireless company that prides itself on having no hidden fees and flexible contracts. Meanwhile, technology like Canva's AI tools and Samsung DEX can offer innovative solutions, but may require some adjustments and trust-building. Visible's commitment to transparency is evident in their straightforward pricing model, with unlimited 5G data costing just $25 a month, including taxes and fees. This contrasts with the mystery and unexplained appearances of Canva's AI tools, which can generate slides and words in seconds, leaving some users wondering why they hadn't discovered this feature earlier. In the world of technology, Samsung DEX is an example of an innovative solution, allowing users to turn their phones into a desktop setup. Dan Seifert, a guest on the Vern Chast podcast, showcased his unique setup, which included a Next Dock, a magnetic mouse, and a Samsung Galaxy Fold. However, Dan expressed reservations about recording through DEX, highlighting the need for trust and confidence in the technology. Overall, the discussion emphasizes the importance of transparency, simplicity, and trust in technology, whether it's a wireless company like Visible or innovative solutions like Canva and Samsung DEX.

The use of devices like the Galaxy Fold or Decks, which aim to provide a laptop-like experience through a smartphone, offers advantages such as portability and eliminating the need to sync data between multiple devices. However, these devices can also come with their own challenges, such as clunkiness and specific demands. The dream is to have a phone that seamlessly transforms into a laptop, but the experience can vary greatly between different devices. For instance, the Galaxy Fold may not be ideal for use as a phone due to its size and shape, while the Pixel Fold may have better aspect ratios when folded. Ultimately, these experiments continue, but it's essential to consider the pros and cons before diving in.

The Supreme Court is currently considering cases regarding social media moderation laws in Texas and Florida, which have been deemed unconstitutional. These cases question if social networks have the right to moderate their own platforms under the First Amendment. The outcome of these cases could significantly impact the future of online communities and moderation. In the tech world, Google has recently launched its new AI model, Gemini, which is intended to compete with models like GPT-4. According to Google, Gemini outperforms GPT-4 in 30 out of 32 benchmarks, but the significance and usefulness of these benchmarks for evaluating AI systems is debatable. Lastly, Gemini is multimodal, meaning it can process and understand both text, audio, and video inputs. These developments in law and technology underscore the ongoing debates around freedom of speech online and the role of AI in our daily lives.

The AI industry continues to evolve, with hot dog (a metaphor for AI) serving as a foundational framework. The most significant development this year has been the addition of multimodal capabilities, allowing AI to process and respond to various types of data beyond text. Google's new chips and AI models, such as Gemini Nano, Pro, and Ultra, are leading this advancement. While the text-based chatbot version of Google's AI is currently accessible, the multimodal version with vision and video capabilities, known as Ultra, is yet to be released. This multimodal aspect is what excites Google and the industry, as it brings us closer to a future where AI can effectively navigate and interact with the world. However, the reliability and authenticity of AI responses, such as those from Google's Bard, remain a concern, as they can sometimes confidently hallucinate incorrect information.

While Google's new AI model, Gemini, has impressive coding benchmarks, these tests don't address concerns about its ability to hallucinate or lie. The discussion highlighted that Gemini scored lower than GPT-4 in several reasoning and comprehension tests, but the significance of these scores is debatable due to the challenges in testing AI's trustworthiness. The speaker emphasized that benchmarks for gaming laptops can provide useful insights into their performance, but the same doesn't apply to AI models as easily due to the vastness and complexity of their capabilities. Ultimately, it's essential to remember that AI benchmarks should not be the sole factor in evaluating their potential impact and reliability.

The latest advancements in AI technology, such as Google's Bard and Bing's Deep Search, are revolutionizing the way we interact with information. These tools can decipher complex manuals, generate summaries, and even understand the nuances of human language, making them invaluable for various applications. However, evaluating their capabilities can be challenging due to the subjective nature of many tasks. For instance, determining the accuracy of settings suggested for specific songs on an old Roland synthesizer based on a manual decoded by an AI model is a complex task. Moreover, the idea of AI models having personalities or cultures adds another layer of complexity. These developments raise deep questions about the nature of intelligence and the future of search engines. While these tools are still in their infancy, they hold immense potential for transforming industries, from journalism to customer service. Google's Pixel phones, with their auto summarization and smart reply features, are early adopters of this technology. Microsoft's Bing, with its deep search capabilities, is taking a step further by attempting to understand the context and meaning behind search queries. Overall, these advancements are a significant step towards making information more accessible and understandable for users. It's an exciting time for AI and its applications, and it's essential to keep exploring and experimenting with these tools to unlock their full potential.

AI models like ChatGPT and Bing are increasingly using prompt engineering to expand and refine user queries, leading to more accurate and comprehensive responses. Microsoft's deep search feature is a prime example of this trend, as it takes a user's query and rewrites it as a longer, more detailed prompt for the Bing search engine. This automation of prompt engineering is a significant development in the field, as it allows users to access more information without having to manually expand their queries. Additionally, Apple has entered the AI news scene with the release of its new MLX model framework, which allows users to run various AI models on Apple chips using their memory architecture ideas. Apple's move towards open source and its unique approach to AI development will be interesting to watch in the coming months. Overall, these developments demonstrate the ongoing advancements in AI technology and its increasing integration into our daily lives.

The use of copyrighted information for AI training is a legal gray area with potential consequences for tech companies. Apple's chip performance and AI capabilities were compared to NVIDIA's, with the former being tested through games like Shadows of the Tomb Raider. The ongoing legal battle between stability AI and Getty Images in the UK regarding the use of copyrighted material for AI training is a significant issue in the AI industry. The unpredictability of court rulings and the potential for costly lawsuits make the situation uncertain for companies. A developer had reverse engineered iMessage protocol, allowing sending messages directly to Apple's servers from non-Apple devices, leading to the creation of a new app called Beeper Mini for Android. This development raises questions about the boundaries of technology innovation and intellectual property rights.

An app called Beeper Mini has been developed to allow Android users to send and receive iMessages, replicating the iMessage experience on Android devices. This is achieved by spoofing Apple device serial numbers, a technique that has been used in the past for running Apple software on non-approved platforms. The result is a near-seamless transition from iMessage on iOS to Android, with features like group messages, reactions, and RCS supported. However, users must currently use two messaging apps - Google Messages and Beeper Mini - with plans to integrate all features into Beeper Mini in the future. Apple's response to this development is uncertain, with some claiming it would require a complete rebuild of the iMessage protocol to stop it. This process, reminiscent of the Hackintosh community, raises questions about the boundaries of technology use and the role of companies in controlling their software.

The gray area between legal and illegal practices in technology, particularly when it comes to software and operating systems, can be complex and subjective. The speaker's experience with SciStar, a company that distributed pirated copies of Apple's OS X, illustrates this. The company's argument was that if they had legally obtained the operating system, they could distribute it as they saw fit. However, the speaker acknowledges that this is a moral argument and not a legally sound one. The situation becomes even more complicated when considering reverse engineering and interoperability. In the case of beeper, a third-party app that enables iMessage on Android devices, the app doesn't use Apple's code but rather sends Apple a serial number. The legality of this practice is debatable, and there's a risk of running afoul of laws like the Computer Fraud and Abuse Act. The speaker also notes that there are legitimate use cases for spoofing serial numbers, and turning off this functionality could disrupt the user experience for legitimate users. Ultimately, the line between right and wrong in technology is not always clear-cut, and it's important to consider the ethical implications of our actions as well as the legal ones.

The legal status of accessing public servers and services, even if it's against the company's wishes, can be ambiguous and subject to varying court rulings. This was highlighted in the discussion about Apple's potential legal action against Beeper for accessing iMessage servers. The EU's decision not to force Apple to make iMessage interoperable was also mentioned as a significant development. Another topic touched upon was the podcast industry's recent struggles, with Spotify undergoing layoffs and canceling popular shows. The industry's shift from an open ecosystem to a more closed one, as Spotify attempted to own the entire industry, was seen as a contributing factor to these challenges. Overall, the conversation emphasized the complexities and uncertainties surrounding technology companies' attempts to control their platforms and services, as well as the potential consequences for smaller players and consumers.

Spotify's foray into buying up podcasting companies and producing celebrity content has not been a success, as evidenced by the departure of high-profile talent like Joe Rogan and Prince Harry and Meghan. The trend of celebrity-centric podcasts without meaningful content or tension may not be sustainable, and the focus on solo influencer shows could lead to a saturated market. The industry seems to be returning to a more sustainable growth model, focusing on interesting and meaningful content. Despite the financial success of Spotify and other podcasting giants, the long-term viability of this business strategy remains uncertain.

Digital ownership is an illusion, and it's essential to consider purchasing physical copies of media to ensure long-term access. This week, Sony accidentally banned multiple PlayStation users from their accounts, causing a stir and reminding everyone that digitally purchased content is not truly owned. Additionally, Sony also announced that some Discovery Plus content would no longer be accessible, further emphasizing this point. While it may be more expensive upfront, buying both physical and digital copies allows consumers to have a backup and maintain ownership over their media. In other news, Tesla's Elon Musk announced that he will be stepping down as CEO of Tesla once he finds a suitable replacement, and Twitter laid off a portion of its workforce. Apple also consolidated its iTunes movies and TV shows app into the Apple TV app. Lastly, Motorola's budget flip phone, the Razr, was reviewed by Allison, and it was found to be underwhelming despite its affordable price point. In summary, it's a reminder that digital ownership is not guaranteed, and purchasing physical copies of media can provide peace of mind and long-term access. Additionally, there were notable tech news stories regarding Tesla, Twitter, and Apple.

Technology advancements in foldable devices have led to a decrease in price, but the affordability of these devices still lags behind, especially for flip designs. The cameras and processors may not be top-notch, but these devices cater to a different market - those who value fashion, compactness, and discretion over high-end features. The recent incident involving a viral photo of a woman in a bridal gown and the subsequent theories of AI photo manipulation serves as a reminder of the importance of context and fact-checking in the age of social media. Despite initial suspicions, the photo was found to be authentic, but the misunderstanding arose due to the camera being in panorama mode without the label, leading to speculation and misinformation.

The line between reality and artificial intelligence-generated content is becoming increasingly blurred, leading to a growing acceptance of AI-faked images as authentic. In the discussed case, a woman accidentally took a vertically composed, panorama-like photo using her iPhone, which was misperceived as a manipulated image due to its perfect stitching. The incident highlights the cultural readiness to believe that AI can generate unbelievable photos, even if it's an accident. The expectation that AI can manipulate images is prevalent, and this belief can lead to debunking efforts, regardless of the truth. Microsoft's recent issue of randomly renaming printers to HP LaserJet further emphasizes the potential impact of AI on our daily lives, as it can subtly influence our experiences. The photo apocalypse may be upon us, and it's essential to recognize the role of AI in shaping our perceptions of reality.

In the ongoing legal battle between Epic Games and Google, the judge has expressed his anger towards Google for deleting records throughout the case, which he views as a deliberate attempt to obstruct justice. The judge has even threatened to investigate Google personally and has given the jury permission to infer wrongdoing if missing records would have answered important questions. This development could make the appeals process messy and may have significant implications for Google. Another notable event was Alison Johnson's article discussing the lackluster returns on investments in 5G infrastructure and the potential for 5G fixed wireless devices as competition for traditional internet service providers. The carriers seem excited about this possibility but face challenges in selling to enterprises and may resort to price increases to recoup their investments.

Despite major carriers technically reaching their coverage goals, they have not built a nationwide wireless network on their own. Instead, they have leased space on existing networks, leading to consolidation and increased debt. This situation has resulted in pricing actions, making it a good time for consumers to consider switching carriers. The speakers on the podcast recommend doing so and even offer a holiday contest where listeners can win Verge merchandise by calling in and sharing their favorite USB-C gadget. The podcast also announced the return of their annual USB-C holiday spectacular, a tradition where they showcase USB-C gadgets and compete for a listener's prize. The speakers encourage listeners to call in and participate. In essence, the discussion highlights the importance of consumer awareness and advocacy in the telecommunications industry.