🗞️ AI News Recap

Hey TEDchats Fam!

Happy Monday! Ever tried catching up on AI news and felt like you were juggling a dozen spinning plates? It's a bit like being a circus performer in the tech arena – exhilarating, but with the constant challenge of keeping up with every twist and turn. This week, we're your skilled ringmasters, orchestrating the highlights of the AI circus that unfolded in the past week. So, step right up, and let's marvel at the feats of innovation – it's a spectacle you won't want to miss!

🐇 QUICK BITES — THE PAST WEEK

Major Tech Companies Commit to Combat Election-Related Deepfakes

In 2024, a pivotal year unfolds for democracy and technology, marked by a record number of global elections. Simultaneously, the rapid advancement of AI introduces both extraordinary benefits and risks, notably the creation of deceptive "deepfakes." Recognizing this challenge, 20 tech companies, including industry giants such as Microsoft, Google, Amazon, Meta, IBM, and Adobe, united at the Munich Security Conference to announce a Tech Accord addressing deceptive AI use in the 2024 Elections. This initiative aims to counter manipulated videos, audio, and images of political figures while prioritizing the protection of free expression and voter choice (read what Microsoft's President has to say about it here).

ChatGPT's Brain Upgrade: Introducing the 'Memory' Feature

OpenAI enhances ChatGPT with ‘memory’, allowing it to remember user preferences, creating a more personalized and efficient experience. While promising a revolution, this innovation triggers privacy concerns. OpenAI addresses this by giving users control over what the AI retains, offering a 'Manage Memory' section and 'Temporary Chat' for incognito interactions. This user-centric approach empowers individuals, emphasizing a balance between personalization and privacy. The introduction of memory in ChatGPT sets the stage for a future with context-aware AI, but ongoing efforts are crucial to navigate the delicate balance between user needs and privacy concerns (check out OpenAI’s full announcement here).

OpenAI Introduces Groundbreaking Text-To-Video Tech

OpenAI introduces Sora, a groundbreaking text-to-video AI model, a new competitor in the field set to rival Runway’s Gen-2 and Kaiber's AI. Sora's innovation lies in its capacity to generate intricate, photorealistic videos from text prompts, featuring multiple characters, dynamic motions, and detailed backgrounds. While facing challenges in simulating complex physics and cause-and-effect scenarios, OpenAI takes a responsible approach by limiting initial access to Sora, seeking feedback from diverse experts and red teamers to enhance accuracy and reliability. As Sora undergoes refinement, it signifies a broader advancement in AI capabilities, envisioning a future where text can effortlessly translate into compelling videos, potentially transforming industries and personal creative expressions (see OpenAI’s announcement of Sora).

Nvidia May Be Getting a $100 Billion Competitor!

SoftBank's founder, Masayoshi Son, is reportedly seeking up to $100 billion for a new chip venture, codenamed Izanagi. The project aims to rival Nvidia and supply essential semiconductors for AI. This marks a significant move for Son as SoftBank scales back startup investments. According to a Bloomberg report, the envisioned company would complement Arm Holdings' chip design unit, positioning Son to establish an AI chip powerhouse. Funding may include $30 billion from SoftBank and a potential $70 billion from Middle Eastern institutions, as per Bloomberg sources (read more).

Trademark Blow: OpenAI's “GPT” Denied by Patent Office

The U.S. Patent and Trademark Office (USPTO) has rejected OpenAI's attempt to trademark "GPT", deeming it "merely descriptive" and ineligible for registration. The patent office argues that "GPT" describes a feature, function, or characteristic of the product, and OpenAI's claim to popularize the term doesn't meet trademark standards. Although this absence of trademark may dilute OpenAI's control over GPT-related terms, their widespread recognition gives them an edge in the market, even without legal protection (read USPTO’s full statement).

Google's Gemini 1.5, Featuring a 1 Million-Token Context Window

Google has recently launched Gemini 1.5, a significant upgrade to its AI model, featuring an unprecedented context window of 1 million tokens (learn more about tokens here). This advancement far exceeds the 128,000-token capacity of OpenAI's GPT-4 Turbo and the 32,000 of the Gemini Pro 1.0. The larger context window allows Gemini 1.5 to process and interpret extensive data sets, from analyzing the entirety of your favorite movie to parsing complex business data, all in a single query (see Google’s announcement + Youtube demo).

Training AI to Sound Like You: The Future of Text Messages Has Arrived

Clubhouse, a formerly popular live audio app, has introduced a feature enabling users to text friends using a personalized voice. With the latest update, users can train AI to mimic their voice for the texts they send. Friends receive these messages in the recreated voice, maintaining the sense of a real-time chat with a friend. In case this sounds too intimidating, rest assured that the text messages are accompanied by an indicator when the AI voice is reading the message (see Clubhouse’s announcement here).

Nvidia Launches Chat with RTX

Nvidia's latest release, Chat with RTX, is an AI chatbot application designed to run locally on PCs equipped with RTX 30- or 40-series GPUs. This tool enables users to feed YouTube videos and documents for summarization and detailed analysis, functioning entirely on the user's computer. Notably, it assists in rapid document search and video content analysis, offering near-instant responses and reducing reliance on cloud-based AI systems. The application, however, is still in its early stages, with some bugs and limitations, reflecting its status as a developmental demo rather than a final product (see Nvidia’s Youtube preview).

Airbnb Acquires AI Firm GamePlanner to Enhance User Experience

The recent acquisition by Airbnb of GamePlanner.AI, a discreet AI company co-founded by Siri's creator with an estimated value of around $200 million, marks a significant shift in the company's technological strategy. CEO Brian Chesky emphasizes the goal of transforming Airbnb's app into a more intuitive, AI-driven interface, enhancing personalization for users. This initiative marks a significant step in leveraging generative AI for a more dynamic and responsive user experience (read more).

Cambio's AI Bots Can Negotiate

Cambio, backed by Y Combinator, introduces AI bots that negotiate debt for consumers and handle sales calls for banks. Originally designed to assist consumers with debt collections, these AI bots have helped approximately 70% of users improve their credit scores by resolving outstanding debts. Expanding its scope, Cambio now offers this technology to banks and credit unions through an API, enabling these institutions to enhance customer service and sales outreach. This advancement represents a notable application of AI in the financial sector (read more).

Slack Adds AI-Driven Search and Summarization Tools

Slack has introduced new features with the aim of improving information accessibility to address repetitive administrative tasks. These include an AI-powered search tool and a function to summarize channel content, making it easier for users to access and understand the vast amounts of informal, unstructured data stored in corporate Slack channels. The update, which leverages generative AI, provides detailed summaries and context-rich search results, facilitating more efficient information retrieval and comprehension in workplace communications (read Slack’s announcement).

Sierra Emerges as a New Player in Business AI

Sierra, co-founded by former Salesforce co-CEO Bret Taylor and ex-Google employee Clay Bavor, represents a significant advancement in conversational AI. Unlike typical customer service bots, Sierra is designed to interactively solve customer problems and execute tasks directly within enterprise systems. This includes complex actions like managing subscriptions or coordinating deliveries. Additionally, Sierra addresses critical issues in AI, such as reducing inaccuracies and hallucinations, which are vital for maintaining brand reputation in large companies (read the latest company announcement here).

Next Big Tech Event Starts in One Week!

In a week (Feb 26-29), around 85,000 attendees will gather at Fira de Barcelona for Mobile World Congress (MWS) 2024, a pivotal event in the mobile industry. Despite recent COVID-related disruptions, MWC remains influential for phone makers, telecoms, and tech vendors. After hitting a peak of 109,000 visitors in 2019, the event faced setbacks, but attendance is gradually recovering, with 88,500 last year and a similar figure expected this year. MWC, though below its peak, continues to be a significant event in the tech calendar (find out more details here).

🧰 AI TOOLBOX — RECAP

👋 UNTIL NEXT TIME

As we lower the virtual curtain on this week's AI circus, it's been a spectacular show filled with technological wonders and jaw-dropping feats. The thrill of the AI arena never ceases, and as we bid adieu to the spotlight, we're left with the anticipation of what the next act will bring. What astonishing innovations and mind-bending breakthroughs await us in the future performances of the AI circus? Join us for our next newsletter as we continue to unveil the next set of breathtaking acts in the ever-evolving world of artificial intelligence.