AI/News
Turning chaos to clarity: Our intelligent curation system analyzes thousands of sources daily — from blogs and social media chatter to podcasts and research hubs — synthesizing essential insights into plain English you can actually use.
AI Startup Not Diamond Sends Your Chat Queries to Different LLMs
The startup Not Diamond has emerged with a novel solution for enterprises looking to deploy the best large language models (LLMs) for their applications. By using smart routing, the company aims to enable a multi-model future where queries are directed to the most suitable model based on accuracy, latency, and cost considerations. Key innovation: an LLM router for a multi-model future; Not Diamond's router uses a 'meta-model' to understand incoming queries and automatically route them to the most appropriate model, optimizing for accuracy, latency, and cost: The company believes that the future will involve many foundation models, fine-tuned variants, and...
read Jul 31, 2024OpenAI Opens Limited Access to Advanced Voice Mode for ChatGPT
OpenAI introduces Advanced Voice Mode for ChatGPT Plus users, enabling more naturalistic conversations with the AI chatbot on mobile apps. Key Details of the Alpha Rollout: OpenAI has begun the gradual release of Advanced Voice Mode to a select group of ChatGPT Plus subscribers, with plans to expand access to all Plus users by fall: The feature allows users to engage in real-time conversations with four AI-generated voices on the ChatGPT mobile app for iOS and Android. ChatGPT will respond naturalistically, handling interruptions and conveying emotions through its utterances and intonations. Users selected for the alpha phase will receive an...
read Jul 31, 2024Anthropic’s Aggressive Data Scraping Is Causing Problems for Sites It Targets
Anthropic's aggressive data scraping practices raise ethical concerns as the company disregards website permissions in its quest to train its Claude language model. Anthropic's data scraping tactics: The artificial intelligence company Anthropic has been engaging in aggressive data scraping practices to gather training data for its Claude language model, often disregarding website permissions and robots.txt files: Ifixit.com reported that Anthropic's ClaudeBot hit their site a million times in a single day, highlighting the intensity of the company's data scraping efforts. Freelancer experienced 3.5 million hits from ClaudeBot in just four hours, with the bot's activities triggering alarms and waking up...
read Jul 31, 2024Microsoft Cloud Revenue Misses Expectations, Shares Tumble
Key financial results: Microsoft reported strong overall performance for its fiscal fourth quarter, but missed expectations on a key cloud revenue metric: The company posted earnings per share (EPS) of $2.95 on revenue of $64.7 billion, beating analyst expectations of $2.94 EPS and $64.5 billion in revenue. However, Microsoft's Intelligent Cloud revenue, which includes its Azure services, came in at $28.5 billion, falling short of the expected $28.7 billion. Impact on Microsoft's stock and the broader market: The disappointment in Microsoft's cloud revenue, particularly in the context of the current AI hype, led to a significant drop in the company's...
read Jul 31, 2024Meta AI Inadvertently Spreads Misinformation on Trump Assassination
Trump assassination attempt misinformation highlights challenges with AI chatbots, as Meta's AI assistant and other generative AI models struggle to handle real-time events accurately. AI hallucinations and misinformation: Meta's AI assistant and other generative AI systems are prone to "hallucinations," where they provide incorrect or inappropriate responses, particularly when dealing with recent events: Meta's AI initially asserted that the attempted assassination of former President Donald Trump didn't happen, despite the incident being widely reported. Joel Kaplan, Meta's global head of policy, acknowledges that this is an "industry-wide issue" affecting all generative AI systems, presenting an ongoing challenge in handling real-time...
read Jul 31, 2024Generative AI is Straining Aging US Power Grid
The rapid rise of generative AI is straining the aging US power grid, raising concerns about its ability to meet the technology's growing energy demands. Power grid challenges: The United States' electrical infrastructure is struggling to keep up with the power-hungry nature of generative AI models: Transformers, which convert raw electricity into usable power, are on average 38 years old and have become a major cause of power outages. Building new transmission lines has faced resistance due to the increased costs that are often passed on to local residents in the form of higher electric bills. Water scarcity and data...
read Jul 31, 2024Meta Ditches Celebrity Chatbots, Shifts Focus to AI Studio
Key takeaways: Meta has discontinued its celebrity-based AI chatbots and is now prioritizing user-generated chatbots through its new AI Studio tool: The celebrity chatbots, which were a significant part of Meta's announcements at its Connect event last September, allowed users to interact with alter-ego versions of celebrities like Charli D'Amelio, Dwyane Wade, and Paris Hilton. Meta spokesperson Liz Sweeney stated that the company learned from building the celebrity chatbots and is now focusing on enabling users, creators, and celebrities to create their own AI chatbots through AI Studio. Pivot to AI Studio: The launch of AI Studio marks a shift...
read Jul 31, 2024Data Center and AI Chip Demand Fuel AMD Q2 Earnings Boost
A resurgent AMD proves its AI mettle in strong Q2, boosting hopes for continued growth in a challenging environment for chip stocks. Key financial highlights: AMD's second-quarter results exceeded expectations, driven by strong performance in its data center and artificial intelligence businesses: Revenue grew 9% year-over-year to $5.84 billion, surpassing estimates of $5.72 billion. Adjusted earnings per share (EPS) increased 19% year-over-year to $0.69, beating estimates by a penny. AI chip success: AMD's MI300X AI chip is gaining traction in the rapidly growing AI market, contributing significantly to the company's data center revenue: Sales of MI300 chips exceeded $1 billion...
read Jul 31, 2024Samsung Q2 Profit Soars on Rising AI Chip Demand
Samsung's Q2 earnings surpass expectations, driven by strong AI memory chip demand and smartphone sales, signaling the company's resilience amidst industry challenges. Robust financial performance: Samsung Electronics reported impressive second-quarter results, exceeding analysts' estimates for both revenue and operating profit: Revenue reached 74.07 trillion Korean won (approximately $53.45 billion), surpassing the average estimate of 73.74 trillion Korean won. Operating profit soared to 10.44 trillion Korean won, significantly higher than the average estimate of 9.53 trillion Korean won, representing a staggering 1,458.2% increase from the previous year. Strong demand for AI memory chips: Samsung attributed its strong performance to robust demand...
read Jul 31, 2024OpenAI’s “GPT-4o Long Output” Produces 16x Longer Responses
OpenAI announces GPT-4o Long Output, a variation of its GPT-4o model that significantly extends output capabilities, allowing for responses up to 64,000 tokens in length. Key features and improvements: The new GPT-4o Long Output model offers a substantial upgrade in output length while maintaining the same overall context window as the original GPT-4o: GPT-4o Long Output can generate outputs up to 64,000 tokens, a 16-fold increase from the original GPT-4o's 4,000-token limit, enabling the model to produce responses as long as a 200-page novel. The maximum context window, which includes both input and output tokens, remains at 128,000 tokens for...
read Jul 31, 2024MIT’s Thermometer Method Efficiently Calibrates Large Language Models
Researchers from MIT and the MIT-IBM Watson AI Lab have developed a new method called Thermometer to efficiently calibrate large language models (LLMs) and improve their reliability across diverse tasks. Key Takeaways: Thermometer addresses the challenges of calibrating LLMs, which can generate inaccurate responses and exhibit misaligned confidence levels: The method involves building a smaller, auxiliary model that runs on top of an LLM to calibrate it, enabling better-calibrated responses on unseen tasks. Thermometer is more efficient than other approaches, requiring less power-hungry computation while preserving the model's accuracy. Understanding the Problem: LLMs can be applied to a wide range...
read Jul 30, 2024Bipartisan Bill Aims to Carve Out AI Deepfakes from Section 230
A bipartisan bill aims to combat AI deepfakes by amending Section 230 protections for tech platforms that fail to address the issue, potentially signaling a new approach to regulating online harms. Key provisions of the Intimate Privacy Protection Act: The proposed legislation, introduced by Reps. Jake Auchincloss (D-MA) and Ashley Hinson (R-IA), targets cyberstalking, intimate privacy violations, and digital forgeries: The bill would amend Section 230 of the Communications Act of 1934, removing legal immunity for platforms that fail to combat these harms. It establishes a "duty of care" for platforms, requiring them to have a reasonable process for addressing...
read Jul 30, 2024AI Wearable ‘Friend’ Offers Constant Companionship
The AI-powered wearable 'Friend' aims to provide constant companionship through an always-listening pendant, without focusing on productivity enhancement. Key Details of the Friend Wearable: The Friend is an AI-powered pendant that listens to the wearer's surroundings and provides commentary and conversation through text messages and push notifications: Developed by Avi Schiffmann, the Friend uses Anthropic AI's Claude 3.5 Sonnet large language model to engage in conversation, offer encouragement, and even tease the wearer. The device has a microphone that listens to everything around the wearer by default, and can be tapped to ask questions or receive unprompted messages. The Friend...
read Jul 30, 2024Data Strategies for Marketers in a Privacy-Focused World
Marketers face significant challenges due to signal loss caused by privacy regulations and technological changes, but innovative strategies and tools can help them adapt and thrive in this new landscape. Understanding signal loss and its implications for advertisers: Signal loss refers to the reduction or elimination of data signals that advertisers rely on to track, measure, and optimize their campaigns, which can happen due to privacy regulations, browser updates, mobile operating system changes, and increased use of ad blockers: Signal loss reduces targeting precision, complicates attribution, and leads to inefficient budget allocation. Advertisers must increasingly rely on first-party data, invest...
read Jul 30, 2024How AI Publishing Enabled Thad McIlroy to Launch His New Book
A game-changing book release showcases the transformative potential of artificial intelligence in the literary world, despite initial distribution challenges. Overcoming AI-driven publishing hurdles: Thad McIlroy's new book, "The AI Revolution in Book Publishing: A Concise Guide to Navigating Artificial Intelligence for Writers and Publishers," faced a brief setback when Ingram flagged and removed it from distribution due to concerns over AI-generated content: Despite McIlroy using AI for only about 1% of the book, such as demonstrating ChatGPT for image alt-text generation, Ingram's filters mistakenly identified the book as AI-generated. The issue was quickly resolved after McIlroy raised concerns on social...
read Jul 30, 2024Meta’s SAM 2 Model A Big Step for Object Segmentation in Videos
The Meta Segment Anything Model 2 (SAM 2) represents a significant advancement in object segmentation for both images and videos, with the potential to revolutionize video segmentation and enable seamless application across various image and video use cases. Key features and capabilities: SAM 2 is the first unified model for real-time, promptable object segmentation in images and videos, offering improved accuracy and performance compared to existing solutions: SAM 2 achieves better video segmentation performance than current methods while requiring three times less interaction time. The model can segment any object in any video or image without the need for custom...
read Jul 30, 2024Runway’s Gen-3 Alpha Transforms Still Images into Stunning Videos
Runway's Gen-3 Alpha AI video model now allows users to generate highly realistic videos from still images, opening up new possibilities for filmmaking and video creation. Key features and capabilities: The latest update to Runway's Gen-3 Alpha model introduces the ability to use still images as prompts for generating AI videos: Users can upload a still image and optionally add text prompts to guide the video generation process, with the option to create 5 or 10-second videos. The image-to-video feature is surprisingly fast, taking less than a minute to generate a video, while maintaining high quality and consistency with the...
read Jul 30, 2024Google to Bundle Gemini Advanced with Pixel 9 Pro Smartphones
According to a new report, Google will bundle a valuable AI subscription with its upcoming Pixel 9 Pro smartphones, significantly increasing the appeal of these devices for consumers interested in advanced AI capabilities. Key details of the bundled offer: The report reveals that Google will include a one-year subscription to Google One AI Premium, worth $239.88, with each purchase of a Pixel 9 Pro, Pixel 9 Pro XL, or Pixel 9 Pro Fold smartphone: Google One AI Premium provides 2TB of cloud storage and access to Google's premium Gemini Advanced AI features, enabling the use of the powerful Gemini 1.5...
read Jul 30, 2024The Cautious Apple Intelligence Rollout Risks Frustrating iPhone Upgraders
The iPhone 16 is set to launch without some of the most anticipated Apple Intelligence features, potentially disappointing customers who are looking to upgrade for the AI capabilities. Apple's cautious approach to AI rollout: Apple is taking a slow and measured approach to introducing its new AI features, even for developers: The first iOS 18.1 developer beta provides access to a limited set of Apple Intelligence features, such as writing tools, Siri improvements, and natural language searches in Photos. Some expected features, like the AI image generation tool Image Playground, were not included in the 18.1 beta. Apple is labeling...
read Jul 30, 2024The Top Business Use Cases for Generative AI
The generative AI revolution is transforming business processes and unlocking new opportunities for innovation across industries. Key use cases for generative AI in business: From chatbots to coding assistants, generative AI is being applied in a wide range of business contexts to streamline operations and enhance productivity: Advanced chatbots powered by generative AI can engage in more natural conversations with customers and handle a greater variety of requests, improving customer service efficiency. Digital assistants or copilots enhanced with generative AI capabilities can perform tasks such as searching for information, creating documents and presentations, summarizing emails and videoconferences, and generating supply-chain...
read Jul 30, 2024Ferrari Exec Targeted in Deepfake Scam
Deepfake technology used in attempt to scam Ferrari executive, highlighting growing trend of AI-powered impersonation attacks targeting businesses. While the attempt was ultimately unsuccessful, the incident underscores the increasing sophistication of these scams and the need for heightened vigilance. The Scam Attempt: A Ferrari executive received seemingly legitimate WhatsApp messages from someone impersonating CEO Benedetto Vigna, discussing a confidential acquisition and requesting the signing of an NDA: The impersonator used a convincing imitation of Vigna's southern Italian accent and had a profile picture of the CEO, adding to the scam's credibility. The fake Vigna claimed to be calling from a...
read Jul 30, 2024Apple Paper Suggests Apple Intelligence Was Trained with Google
Key revelations from Apple's research paper: A quote buried in Apple's "Apple Intelligence Foundation Language Models" paper suggests that the company initially relied on Google's hardware for training its AI models: The paper states that Apple's Foundation Model (AFM) and its underlying server technology were initially built on "v4 and v5P Cloud TPU clusters" using Apple software. While a CNBC report suggested that Apple rented time on Google-hosted clusters, the research paper does not explicitly mention Google or Nvidia, implying that Apple likely purchased the hardware directly and used it within its own data centers. Apple's evolving AI infrastructure: As...
read Jul 30, 2024Deepfakes and Algorithms: How Bad Actors Weaponize AI to Manipulate Minds
The rise of artificial intelligence (AI) has brought about unprecedented opportunities, but also significant dangers as bad actors exploit the technology to manipulate people and undermine trust in the digital ecosystem. The dark side of AI: Bad actors, from cybercriminals to unethical corporations and rogue states, are weaponizing AI to craft sophisticated strategies that influence individuals and groups, often without their knowledge: Deepfakes, hyper-realistic video or audio recordings that make it appear as if someone is saying or doing something they never did, pose a significant threat to personal reputations and the integrity of information. AI-powered social media bots and...
read Jul 30, 20245 Popular Myths About Generative AI
The rapid rise of generative AI has led to widespread hype, confusion, and fear, with many misunderstandings about its true capabilities and implications. Myth 1: Generative AI is intelligent; While generative AI can mimic certain aspects of natural intelligence, such as language processing and creation, it lacks the full range of human cognitive abilities: Generative AI's "thought processes" are based on probability rules learned during training, making them far more limited and less sophisticated than human intelligence. It is called "artificial" intelligence for a reason, as it can perform impressive tasks but is still fundamentally an algorithm. Myth 2: Generative...
read