• AI KATANA
  • Posts
  • Meet GPT-4o: OpenAI's Latest AI with Real-Time Multimodal Capabilities

Meet GPT-4o: OpenAI's Latest AI with Real-Time Multimodal Capabilities

Also: NASA Names First Chief Artificial Intelligence Officer

Author

AI KATANA
May 13, 2024

Hi!

In today's newsletter, we dive into the latest from OpenAI, NASA, and other tech and AI developments making waves. OpenAI has launched GPT-4o, a groundbreaking AI model integrating text, audio, and vision capabilities, revolutionizing human-computer interactions with its fast, multilingual, and expressive responses. NASA appointed David Salvagnini as its first Chief Artificial Intelligence Officer, marking a significant step in advancing AI technologies for space exploration. Google’s new AI-driven search tool, "Search Generative Experience," is causing concern among web publishers due to potential traffic and revenue declines. The US Air Force is pioneering AI systems for aircraft navigation without GPS, preparing for electronic warfare scenarios. The race to combine quantum computing with AI is heating up, promising to transform industries with exponential computing power. OpenAI's new ChatGPT desktop app for macOS enhances user experience with a refreshed interface and seamless functionalities. Lastly, the San Francisco Bay Area leads in AI technology and funding, securing over 50% of global venture capital for AI startups. Stay tuned for more updates and insights in tomorrow's newsletter!

Sliced:

  • 🤖 Meet GPT-4o: OpenAI's Latest AI with Real-Time Multimodal Capabilities

  • 🧑🏻‍🚀 NASA Names First Chief Artificial Intelligence Officer

  • 👩🏻‍💻 Web publishers brace for carnage as Google adds AI answers

  • 🛫 The US Air Force is teaching AI to navigate aircraft in case GPS gets taken out in a future fight

  • ⏩ Gen AI Has Already Taken the World by Storm. Just Wait Until It Gets a Quantum Boost

OpenAI has unveiled GPT-4o, a groundbreaking AI model that seamlessly integrates text, audio, and vision capabilities, setting a new standard for human-computer interaction. GPT-4o responds to audio inputs with latency as low as 232 milliseconds, closely mimicking human response times, and it excels in multilingual text processing, surpassing GPT-4 Turbo's performance. Notable new features include its ability to handle and generate outputs across multiple modalities—text, audio, and images—within a single neural network. This end-to-end model allows for more natural interactions, recognizing tones, multiple speakers, and background noises, and it can output more expressive audio, such as laughter and singing.

GPT-4o also introduces significant efficiency improvements, being twice as fast and 50% cheaper in the API compared to GPT-4 Turbo, and supports five times higher message limits for Plus users. Its advanced tokenizer reduces the number of tokens required for many languages, enhancing processing speed and accuracy. Safety has been prioritized with built-in measures, rigorous evaluations, and extensive external red teaming to mitigate risks, particularly with audio modalities. Initially, GPT-4o’s text and image functionalities are available, with audio and video capabilities to follow. This release marks a major advancement in AI usability, promising broader accessibility and practical applications for developers and users alike.

NASA has appointed David Salvagnini as its first Chief Artificial Intelligence Officer, a new role expanding his responsibilities as the current Chief Data Officer. This move aligns with President Biden's Executive Order on the development and use of AI, emphasizing NASA's commitment to advancing AI technology responsibly. Salvagnini, with over two decades of experience in technology leadership and a background in the intelligence community, will spearhead AI strategy and planning across the agency. He will also enhance collaboration with other government entities, academia, and industry. NASA leverages AI to support various missions, from analyzing Earth science imagery to managing communications for the Mars rover. Salvagnini’s appointment marks a significant step in integrating AI to accelerate space exploration and research efforts, ensuring that NASA remains at the forefront of technological innovation.

As Google rolls out its new AI-driven search tool, the "Search Generative Experience" (SGE), web publishers are bracing for a significant impact on their traffic and revenue. The AI-generated answers provided by SGE often displace traditional links to websites, which threatens the visibility and viability of millions of online creators who rely on search engine traffic. The shift is expected to reduce web traffic from search engines by up to 25% by 2026, potentially resulting in billions of dollars in losses for digital content creators. This move has raised concerns about the centralization of information and the future of an open internet, as it may force websites to purchase ads to maintain visibility. Critics argue that this change could undermine the diverse ecosystem of web content that has flourished over the past two decades.

🛫 The US Air Force is teaching AI to navigate aircraft in case GPS gets taken out in a future fight

The US Air Force is developing AI systems to navigate aircraft in environments where GPS is unavailable, addressing concerns that future wars could see electronic warfare and anti-satellite weapons disrupt GPS-dependent operations. This initiative, part of broader AI projects within the military, includes using AI to navigate C-17 cargo planes via Earth's magnetic fields, a method complicated by electromagnetic noise. Successful tests indicate AI can effectively navigate without GPS, which is critical given the lessons from the ongoing conflict in Ukraine where GPS jamming is common. This AI-driven approach aims to maintain operational capabilities in GPS-denied scenarios, potentially reshaping military navigation and strategy.

⏩ Gen AI Has Already Taken the World by Storm. Just Wait Until It Gets a Quantum Boost

China, the U.S., and major tech firms are in a heated race to integrate quantum computing with AI, potentially revolutionizing the technology landscape. Quantum computing, which leverages qubits capable of existing in multiple states simultaneously, offers exponential boosts in computing power and the ability to mimic natural processes more intuitively than traditional binary computing. This synergy is poised to enhance generative AI applications, which are already transforming various industries by enabling rapid creation of content. Companies like IBM and startups like Quantinuum are making significant strides, each employing different quantum technologies such as superconducting qubits and trapped ions. China's aggressive investment in quantum research, exemplified by the Jiuzhang quantum computer's impressive performance, underscores the geopolitical stakes. As the U.S. also ramps up funding and imposes restrictions on Chinese firms, the global competition is set to accelerate advancements in both AI and quantum computing, potentially leading to breakthroughs in solving complex problems like climate change.

🛠️ AI tools updates

OpenAI has launched a new desktop app for ChatGPT, currently available only for macOS. Announced by CTO Mira Murati, the app features a refreshed user interface and allows users to interact with ChatGPT by typing or speaking. It includes functionalities such as using a keyboard shortcut (Option + Space) to ask questions and taking and discussing screenshots within the app. While both free and paid users will have access, initially, it will be exclusive to ChatGPT Plus subscribers before a wider release. A Windows version is expected later this year. Additionally, OpenAI has introduced a new model, GPT-4o, which is faster and available to all users for free. This update follows ChatGPT's availability on iOS and Android, aiming to make AI interactions more natural and integrated across platforms.

💵 Venture Capital updates

The San Francisco Bay Area has emerged as the global leader in AI technology and funding, securing over 50% of worldwide venture capital for AI startups in 2023. This surge, highlighted by OpenAI's $10 billion investment from Microsoft, underscores the region's dominance. Bay Area AI companies raised over $27 billion last year, a significant increase from $14 billion in 2022. Major players like OpenAI, Anthropic, and Inflection AI have established substantial real estate footprints in the area. The region's concentration of talent, big tech companies, and top-tier universities like UC Berkeley and Stanford drives its leadership. Despite global competition from AI hubs in China, the U.K., and Canada, the Bay Area's unique ecosystem continues to attract startups and investors, solidifying its status as the epicenter of AI innovation.

🫡 Meme of the day

⭐️ Generative AI image of the day