Published Oct 13, 2023

A week of horror, an AI conference of contrasts

Host Alex Volkov delves into the dual narratives of innovation and fear at an AI conference, exploring its diverse insights on multimodal technologies, evolving business models, and the transformative potential of AI amidst the emotional and spiritual turmoil of modern progress.
Episode Highlights
ThursdAI logo

Popular Clips

Episode Highlights

  • Multimodal AI

    The conference showcased significant advancements in multimodal AI technologies, highlighting their potential applications and implications. emphasized the importance of integrating AI beyond text, featuring a remarkable demo involving DALL-E and GPT-4 Vision, which demonstrated AI's ability to understand and generate art through image comparison and critique 1. This approach mirrors the process of code interpretation, where AI generates a draft, critiques it, and refines the output 1. and Swyx discussed the potential of GPT-4 Vision in real-world applications, such as navigating environments using visual cues, which could revolutionize how AI interacts with the physical world 2.

    Multimodality is a thing and I think vision is first, is what I'm saying. After text, what comes after text? And there's two candidates for what comes after text. One is voice and one is vision.

    ---

    The exploration of multimodal AI also included innovative demos by Sam and Jason, who are working on creating a new kind of operating system that seamlessly integrates AI into various modalities, potentially redefining human-computer interactions 3.

       

    AI Business Models

    The evolving business models in AI were a focal point, with discussions on revenue generation and sustainability. highlighted the financial challenges faced by AI companies, such as GitHub Copilot's losses, and emphasized the need for a strategic approach to business operations 4. He criticized the sensationalism in media narratives that often misrepresent the financial realities of AI ventures 4. shared his perspective on maintaining positivity and truth in the community, advocating for a focus on constructive discourse rather than sensational headlines 4.

    Nobody ever makes money on AI because look at Copilot is losing $20 per user. If that's your narrative, which a lot of people are on my timeline today doing that, then I don't think you're going to do very productive things of your life.

    ---

    The conference also featured a standout demo by Dedy from Codium AI, showcasing innovative approaches that differentiate them from competitors like Copilot, underscoring the diverse strategies AI companies are employing to achieve sustainability 5.

       

    AI Engineering

    AI engineering innovations were prominently featured, with new tools and techniques unveiled at the conference. categorized AI engineers into three types: AI-enhanced engineers, AI product engineers, and AI engineer agents, each playing a crucial role in advancing AI technologies 6. The event included workshops on AI engineering fundamentals, providing attendees with hands-on experience and foundational knowledge 6.

    AI enhanced, which is engineer using AI tools but working on non AI products. And then there's AI product engineer, which is aiengineer using AI APIs to build AI enabled products for the end customer.

    ---

    Open source collaboration was another key theme, with leading efforts to establish standards that facilitate interoperability among AI tools, aiming to accelerate innovation and reduce redundancy in development efforts 7.

Related Episodes