In April, Mark Zuckerberg declared that the company would significantly increase capital expenditures in the coming year to enhance its position in the development of advanced artificial intelligence. Last week, he delivered on that commitment with the unveiling of a $15 billion “AI Superintelligence” team, boasting nine-figure salaries and a 49% stake in Scale AI. Meta has also brought aboard Alexandr Wang, the 28-year-old founder of Scale and former roommate of Sam Altman from OpenAI.
Prior to Meta’s investment, Scale was already servicing most major AI players, some of whom were not enthusiastic about the development. Bloomberg reports that the head of Scale AI now brings insight into the operations of other competitors. Google, Scale’s largest client, expressed concerns as the tech giant informed startups that their existing partnerships would be reconsidered Reuters reported on Friday.
My colleague, Robert Booth, has more insights:
One Silicon Valley analyst referred to the announcement as indicative of a “wartime CEO” mentality.
Super Intelligence is characterized as a form of AI that surpasses human capabilities in all areas—a current standard that remains unattained by existing AI systems, known as Artificial General Intelligence (AGI).
In light of effective advancements by rivals including Sam Altman’s OpenAI and Google, and substantial investments in Meta’s underperforming Metaverse concept, observers are keen to see if Meta can meet its ambitious targets in AI.
While the scale of this financial commitment may suggest a lead in the AI race, Meta is actually playing catch-up. The company’s latest models lack the sophistication of competitor offerings. Earlier this year, it revealed a model that was adjusted to succeed on popular benchmarks—an unconvincing strategy. Its enormous financial outlay and new team formations aim to enhance its stature among industry leaders.
Discover more about Meta’s Superintelligence initiative here.
AI News Week
Wikipedia, essential to AI, struggles to showcase human contributions
Last week, Wikipedia ceased its testing of AI-generated article summaries after backlash from editors and volunteers who contribute to the online encyclopedia. These AI-generated summaries were displayed above human-written introductions on the article pages. This experiment, which targeted around 10% of mobile users, lasted for two weeks.
Feedback was described as “strongly negative,” according to the nonprofit parent of Wikipedia. One editor expressed on the Public Forum: “A truly frightening idea.” Another stated, “Please exclude AI from Wikipedia.” Many joined in saying that this was uniquely detrimental. While some advocated for AI abstracts to simplify access to technical articles, the majority argued that errors and misinterpretations could harm Wikipedia’s reputation for reliability.
General sentiment among editors can be summarized as “Yuck.”
“We’re trying to balance AI integration with Wikipedia. There’s very minimal AI presence, just enough to aid editors and improve reader access,” explained one editor. The continual debate surrounds “What defines the line between human-generated and AI-generated content?”
In contrast, Reddit has begun charging for access to its application programming interfaces in 2023. This posting library is invaluable for AI companies, serving as a broad repository of contemporary human-written texts crucial for training large language models like ChatGPT. Reddit’s executives are leveraging this resource to monetize its value for AI firms. This decision has disrupted the user experience, particularly angering volunteer moderators and millions of community members who have protested the changes without success. They prioritized revenue from AI businesses over user satisfaction, starkly opposing Wikipedia’s approach.
A few days after Wikipedia rolled back its AI summary test, the Wikimedia Foundation hosted an in-person editing event at the United Nations. According to Rudder, the information manager, this gathering of a diverse group of editors was part of UN Open Source Week, coinciding with various open-source hackathons.
Editors interacted enthusiastically around laptops, collaborating on articles in multiple languages in a vibrant auditorium. Italian and Sri Lankan representatives acknowledged the vital work of editors in speeches, and Wikimedia Foundation executives expressed gratitude for their contributions, alongside UN Digital and Emerging Technology Governors. It was an energetic atmosphere.
Rudder emphasized the importance of in-person events, stating, “It fosters community. There’s a misconception that Wikipedia editors operate in isolation, when in reality, that’s far from the truth.” Wikimedia NYC recently organized the UN event, with Rudder observing it vital for connecting new editors to the community.
Like Reddit, Wikipedia’s vast text corpus serves as key training data for most, if not all, major language models. Wikipedia endorses a corporate approach to data access with Wikimedia Enterprise, which was launched in 2021. While Reddit operates for profit, the Wikimedia Foundation is a nonprofit entity. Although Wikipedia data remains available freely, utilization can be costly.
Unilike Reddit, Wikipedia emphasizes the human effort required to sustain the site and its information consistently by adjusting features in response to community feedback.
After the newsletter promotion
Technical Perspectives on the Anti-Deportation Protests in Los Angeles
Apple’s Walled Garden: From Eden to Pen
Recently, Apple hosted its annual WWDC developer conference, where the most significant announcement was a redesign of the iPhone operating system (iOS), referred to as “Liquid Glass.” The virtual buttons displayed on the iPhone screen appear transparent, allowing the background image to warp slightly, reminiscent of the colorful, translucent plastic used in older Macs.
The refreshed aesthetic is visually appealing but lacks excitement. Ultimately, it felt like a backdrop for news on Hurricane Prediction. How relevant is the color of my phone’s controls?
Of course, there were other updates, like live translations in FaceTime calls, enabling smoother communication with distant relatives and friends.
Yet overall, the updates lacked inspiration. Over the years, iPhone updates have seemed more forward-thinking than, for instance, the introduction of a second-generation 3G connection in 2008. Apple’s progress pales in comparison to the dynamic Android ecosystem. I once spoke with a venture capitalist in San Francisco in 2019, and I couldn’t even identify his iPhone model.
Unlike competitors, Apple has been slow to introduce new AI features. Google and Samsung have rapidly incorporated AI into their mobile offerings. For example, live translation has been a part of Android for years. Google’s Pixel Buds can convert speech in real-time, while Apple’s AirPods still lack this capability. Google consistently releases enticing, if sometimes quirky, AI updates.
Just last Friday, The Verge reported: Google is trialing audio summaries for specific search queries in its lab for English speakers in the US.
Following the demonstration, Gizmodo noted that while Apple recognizes that AI may not align with consumer desires, it can’t openly admit this. However, despite differing opinions on ChatGPT among monthly users, Apple’s announcements offered little substance. If Apple wishes to retain users within its ecosystem, it must deliver something captivating and valuable.
Last week, Google launched an updated version of its operating system, while Samsung unveiled features for its next foldable phone. Android Central. Have you heard about these developments? In tech journalism, it’s well-known that Apple and iPhone stories often receive more clicks than those about Android phones—despite a higher global market share for Android. The reasons behind this phenomenon warrant their own discussion. Many spent the week discussing Android updates, contrasting with Apple’s liquid glass interface. Apple’s updates seem primarily cosmetic, raising questions about their overall impact as they inspire humor instead of admiration.
Samsung’s seventh generation of foldable phones raises questions about the novelty in the smartphone market. When contemplating your new iPhone’s lack of excitement, what do you want it to achieve? There’s no inherent need for a folding phone. My Motorola Razr facilitated effective communication, but that wasn’t due to its physical design.
I find myself pondering the value of Google’s new AI feature. Although there may not be a consensus on what most users desire from an AI-enhanced smartphone, it’s clear that without innovation, stagnation occurs. I feel Apple fails to present remarkable features worth considering. While the audio summary for Google search results might seem unnecessary, it is indeed fascinating—particularly for the visually impaired. Meanwhile, Siri struggles with basic tasks like controlling Spotify and managing notifications.
Having relied on my iPhone as my primary device for roughly 15 years, I feel I’m at a turning point. It may soon be time to explore Android and discover the new shiny gadgets, especially as I navigate the Apple hardware and app ecosystem.
Wider Technology
Source: www.theguardian.com
