Jul 26, 2024

How “Speculative Design” Can Revamp The Education System for the Age of AI

The future of education hangs in the balance as school systems struggle to keep pace with exponential technological growth in the age of AI. Speculative design, a methodology that uses design to explore and envision possible futures, may provide a powerful framework for reimagining education from the ground up. Adaptability as a survival skill: Speculative design methods can help foster adaptability, a crucial skill for thriving in an unpredictable future: By creating scenarios that represent potential future scenarios, students and teachers are challenged to navigate uncharted territories, developing the agility essential for success. Encouraging students to tackle ethical dilemmas posed...

read Jul 26, 2024

How AI Assistants Threaten to Disrupt Aggregator Platforms

The tech industry is on the cusp of a significant shift as AI-driven personal assistants like Siri threaten to disrupt the dominance of aggregator platforms such as Amazon, booking.com, and Uber. The rise of aggregators: Over the past decade, aggregator platforms have come to dominate various sectors by providing the most convenient access to services and maintaining a direct relationship with users: Amazon has become the go-to platform for product searches and purchases, with half of all users starting their product searches directly on the site. Aggregators in other sectors, such as booking.com for holidays, Netflix for entertainment, and Uber...

read Jul 26, 2024

Steve Jobs’ 1983 Speech Foreshadowed AI Revolution, Transformative Tech’s Promise and Perils

In a recently uncovered 1983 speech, a young Steve Jobs predicted the ubiquity of personal computers and their potential to revolutionize communication and society, foreshadowing the transformative impact of digital technology that we now take for granted. Key takeaways from Jobs' prescient speech: Jobs addressed a skeptical audience of designers, most of whom had never used or even seen a personal computer, and made bold predictions about the future: Jobs declared that in people's lifetimes, the computer would become the predominant medium of communication, even though few in the audience owned or had experience with computers at the time. He...

read Jul 26, 2024

Google’s Search Dominance Challenged as AI Reshapes the Web’s Future

The Google era is officially over as the search giant faces new challengers, but its dominance remains strong for now. This week's Vergecast explores the shifting landscape of search, reviews the latest foldable phones and gaming handhelds, and discusses key developments in streaming and entertainment. The future of search hangs in the balance: OpenAI's launch of SearchGPT marks a significant threat to Google's long-standing dominance in the search market, even as the tech giant posts massive earnings driven by its core search business and AI investments: OpenAI's SearchGPT leverages advanced AI capabilities to provide a new search experience, potentially disrupting...

read Jul 26, 2024

DeepMind’s JumpReLU Architecture Sheds Light on the Inner Workings of Language Models

DeepMind has made significant progress in interpreting large language models (LLMs) with the introduction of JumpReLU sparse autoencoder (SAE), a deep learning architecture that decomposes the complex activations of LLMs into smaller, more understandable components. The challenge of interpreting LLMs: Understanding how the billions of neurons in LLMs work together to process and generate language is extremely difficult due to the complex activation patterns across the network: Individual neurons don't necessarily correspond to specific concepts, with a single neuron potentially activating for thousands of different concepts, and a single concept activating a broad range of neurons. The massive scale of...

read Jul 26, 2024

Existential Risk from AI Too Unreliable to Base Policy Decisions on, Some Experts Say

In a recent blog post, Arvind Narayanan and Sayash Kapoor argue that forecasts of existential risk from AI are based on speculation and pseudo-quantification rather than sound evidence or methodology. Key issues with AI existential risk forecasting: The article identifies several reasons why current AI existential risk probability estimates are unreliable and unsuitable for guiding policy: Inductive probability estimation is unreliable due to the lack of a suitable reference class, as an AI-driven human extinction event would be unprecedented and dissimilar to any past events. Deductive probability estimation is unreliable due to the lack of a well-established theory or model...

read Jul 26, 2024

ChatGPT Voice Mode Expected to Launch Next Week for (Some) Plus Users

ChatGPT Voice Mode set to launch next week for Plus subscribers: OpenAI has announced that its highly anticipated Voice Mode feature for ChatGPT will begin rolling out to Plus subscribers starting next week, giving AI enthusiasts something new to look forward to. Limited alpha rollout to gather feedback: The initial release of Voice Mode will be limited to a small group of Plus users as part of an alpha test, allowing OpenAI to gather feedback and make improvements before expanding access based on what they learn: OpenAI CEO Sam Altman confirmed the alpha rollout timeline in response to a user...

read Jul 26, 2024

How Microsoft is Making Power Users out of Some of its Biggest Skeptics

Microsoft is turning AI skeptics into power users by actively integrating AI into the workplace and leveraging employee enthusiasm to drive business value. Adding critical new capabilities to Copilot: In response to customer feedback, Microsoft has enhanced its Copilot AI tool with several new features: Workplace-specific prompts, function-specific prompts, and the ability to share and reuse prompts Auto-complete and rewrite features to assist in the prompt-writing process Scheduling prompts to run at specific times Measuring the impact of generative AI: Microsoft has shifted its approach to quantifying the impact of AI in the workplace: Initial focus was on time savings...

read Jul 26, 2024

How to Prevent Africa from Being Left Out of the Global AI Transformation

Africa risks being left behind in the AI revolution due to its weak digital infrastructure, which is holding back the continent's potential to fully participate in and benefit from AI advancements. Key Takeaway: Africa's lack of connectivity is a major obstacle to harnessing the potential of AI: More than two decades ago, The Economist found that all of Africa had less international bandwidth than Brazil, a situation that persisted until 2023. This digital divide puts Africa at risk of missing out on the promises and benefits of AI technologies that are transforming other parts of the world. Broader Implications: Bridging...

read Jul 26, 2024

Data Centre Boom Strains Australian Power Grid, Threatens Price Hikes

The rapid growth of data centres in Australia is straining the electricity grid and could lead to higher power prices, according to energy experts. Soaring demand for data centres: Australia is a top hub for data centres globally, with 214 facilities nationwide. The explosive growth of cloud computing and AI is expected to drive exponential expansion in the coming years: One large data centre can consume as much electricity as 50,000 homes. Data centres currently use an estimated 5% of Australia's grid power, projected to rise to 8-15% by 2030. David Dzienciol of data centre operator NEXTDC believes electricity demand...

read Jul 26, 2024

OpenAI Reassigns Its Top Safety Exec Amid Mounting Scrutiny, Antitrust Probes

OpenAI has reassigned top AI safety executive Aleksandr Madry to role focused on AI reasoning. Key developments: Last week, OpenAI removed Aleksandr Madry, one of its top safety executives, from his role as head of preparedness and reassigned him to a job focused on AI reasoning: Madry's preparedness team was tasked with protecting against catastrophic risks related to frontier AI models. He will still work on core AI safety in his new role. The decision came less than a week before Democratic senators sent a letter to OpenAI CEO Sam Altman questioning how the company is addressing emerging safety concerns...

read Jul 26, 2024

AI Legal Startups Surge as VCs Bet on Measured Approach

AI-powered legal-tech startups gain traction as they take a measured approach to the technology's potential. Investors are rewarding law-focused AI companies like Harvey and Clio with rich new valuations as founders emphasize a hybrid approach that combines AI with existing legal services and acknowledges the technology's current limitations. Key developments driving investor interest: Several factors are fueling the unprecedented flow of capital into legal tech startups: Legal-tech startups are taking a more measured approach, emphasizing AI's fallibility upfront and using a hybrid model that integrates AI with proven legal services and products. There has been an estimated 100% growth in...

read Jul 26, 2024

Ex-Apple Engineers Develop Lidar-Free Autonomous Delivery Robots

A California-based startup called Vayu Robotics, founded by ex-Apple car engineers, has developed autonomous delivery robots that can navigate complex environments without using expensive lidar sensors. Key Takeaways: Pioneering a new approach to autonomous navigation Vayu's delivery robots leverage a transformer-based "mobility foundation model" and passive sensors to navigate autonomously, eliminating the need for costly lidar sensors or pre-mapping environments. This unique technology allows the robots to handle payloads up to 100 pounds and travel at speeds under 20mph, making them suitable for last-mile delivery applications. The company's founders, previously part of Apple's Special Projects Group working on the now-canceled...

read Jul 26, 2024

AI Unicorn Cohere Cuts Jobs Despite $5.5B Valuation, Highlighting Industry Challenges

Cohere, an AI startup founded by ex-Google researchers and backed by Nvidia, recently completed a $500 million funding round that more than doubled its valuation to $5.5 billion. Despite this, the company cut around 20 roles from its 400-person workforce on Tuesday. Focusing on enterprise AI: Unlike many competitors, Cohere has concentrated its efforts on developing generative AI solutions for businesses rather than consumer-facing chatbots. The company's clients include notable names such as Notion, Oracle, and Bamboo HR. Ongoing hiring in key areas: Even with the recent layoffs, Cohere is still actively hiring for positions in customer operations, partnerships, revenue,...

read Jul 26, 2024

OpenAI is Hoping Publisher Partnerships and GenAI Will Give SearchGPT an Edge in Competitive Market

OpenAI's new SearchGPT prototype aims to chip away at Google's dominance in web search by leveraging generative AI and partnerships with publishers to provide users with more efficient and conversational search results. Key details of SearchGPT: OpenAI has revealed limited information about its new search engine prototype, which is designed to help users find information more quickly and easily: SearchGPT uses generative AI to gather links and provide answers to user queries in a conversational manner, potentially integrating with the popular ChatGPT chatbot in the future. The search engine will draw upon information from the broader web as well as...

read Jul 26, 2024

TV Doctors Deepfaked, Exploited to Peddle Health Scams on Social Media

TV doctors "deepfaked" to promote health scams on social media: An investigation by the British Medical Journal has revealed that several well-known TV doctors, including Michael Mosley and Hilary Jones, have been "deepfaked" on social media platforms to promote fraudulent health products and scams. "Deepfaking" involves using artificial intelligence to create convincing videos of individuals by mapping their digital likeness onto someone else's body, essentially making it appear as if they are promoting or endorsing products they have no actual connection to. The investigation found videos on platforms like Facebook and YouTube, with one example showing a deepfaked Dr. Hilary...

read Jul 26, 2024

Meta’s Open-Source AI Sparks Debate Over Safety, Innovation, and Accountability

A debate over open-source versus closed AI models is emerging, as Meta releases an open-source model while OpenAI keeps its code private. This development raises important questions about the implications of these different approaches for AI safety, competition, and innovation. Meta's open-source approach sparks controversy: Meta CEO Mark Zuckerberg has called for open-source AI development and released an open-source model, Llama 3.1, which the company claims can compete with closed models like OpenAI's ChatGPT. Anthony Aguirre, executive director of the Future of Life Institute, suggests that open-source models are incompatible with safety regulation, as they lack the necessary guardrails to...

read Jul 26, 2024

“Copyright Traps” Helps Creators Detect AI Training on Their Work

Researchers have developed a new tool that could help content creators prove their work has been used to train AI models without their consent. The method, called "copyright traps," allows writers to subtly mark their text in a way that can later be detected in AI training data. Key points about the copyright trap tool: The copyright traps work by injecting long, gibberish sentences multiple times into a piece of text, which can then be used to determine if that text was used to train an AI model: The trap sentences are generated using a word generator and chosen randomly...

read Jul 26, 2024

IBM Says Quantum Computing May Solve AI’s Energy and Cost Crisis

The exponential growth in AI training costs is raising sustainability and equity concerns, but quantum computing may offer a solution. The AI training cost problem: The rapid increase in AI usage is driving up energy consumption and costs, which could surpass the US GDP by 2030 and limit AI's potential if left unchecked. ChatGPT queries consume at least 10 times more energy than Google searches, contributing to the unsustainable growth in AI's energy footprint. By 2030, the cost to train a single AI model once could exceed worldwide IT spending, creating a hard ceiling for the technology's development and accessibility....

read Jul 26, 2024

How to Get on the SearchGPT Waitlist

SearchGPT prototype launched by OpenAI, aiming to revolutionize AI-powered search: OpenAI has surprise-revealed SearchGPT, a new AI-powered search engine running on GPT-4 models that has real-time access to internet-based information, potentially challenging Google's search dominance. Key details about the SearchGPT prototype: SearchGPT is currently a prototype being tested by OpenAI with a limited waitlist for early access: SearchGPT provides a large textbox for user queries and attempts to summarize and organize the information it finds, with links available in a sidebar. The results can be followed up with relevant queries using a chatbox, and summaries are expected to have in-line...

read Jul 26, 2024

Samsung is Betting on AI to Expand Its Hardware Ecosystem and Drive Revenues

Samsung is betting on AI and new form factors to drive smartphone upgrades and expand its product ecosystem. Galaxy AI push to incentivize upgrades: Samsung is leveraging its Galaxy AI suite of artificial intelligence features to motivate users to purchase new products, emphasizing the convenience and benefits these capabilities bring: The company recently launched the Z Fold6, Z Flip6, and its first smart ring, the Galaxy Ring, with AI as a key selling point. Samsung mobile head TM Roh believes Galaxy AI will become more pervasive in daily life over time, similar to the trajectory of smartphones and the internet....

read Jul 26, 2024

Microsoft Launches Serverless Fine-Tuning for Its Phi-3 Model

Microsoft introduces serverless fine-tuning for its Phi-3 small language model, allowing developers to adapt the AI for specific use cases without managing their own servers, intensifying competition in the enterprise AI market. Key features of Phi-3 small language model: Microsoft's Phi-3, a 3 billion parameter model, offers affordable performance on coding, common sense reasoning, and general knowledge, now as part of a family of 6 models with varying parameters and context lengths: Phi-3 models range from 4,000 to 128,000 tokens in a single input, with costs from $0.0003 to $0.0005 USD per 1,000 input tokens. Designed with safety features for...

read Jul 26, 2024

OpenAI’s SearchGPT Demo Riddled with Inaccuracies, Raising Concerns About AI-Powered Search Reliability

OpenAI's SearchGPT demo has raised concerns about the accuracy and usefulness of its search results. Key issues with the demo: The prerecorded demonstration video of SearchGPT accompanying the announcement of OpenAI's new SearchGPT engine showcased results that were largely inaccurate or unhelpful: In response to a query about music festivals in Boone, North Carolina in August, SearchGPT provided a list of festivals with incorrect dates. For example, it stated that the An Appalachian Summer Festival would be hosting events from July 29 to August 16, when in reality, the festival started on June 29 and will have its final concert...

read Jul 25, 2024

Elon Musk’s xAI Claims “Memphis Supercluster,” Will Be World’s Most Powerful AI by 2024

Elon Musk announces launch of xAI's "Memphis Supercluster," claiming it will be the world's most powerful AI training system by the end of 2024, but questions remain about the facility's power consumption and local impact. Ambitious claims and technical details: Musk took to X (formerly Twitter) to share that xAI, in collaboration with X and Nvidia, has begun training on a massive supercomputer cluster in Memphis, Tennessee, with the goal of developing the world's most powerful AI by every metric by December 2024: The "Memphis Supercluster" features 100,000 liquid-cooled Nvidia H100 GPUs on a single RDMA fabric, which Musk claims...

read