The Daily AI Show Newsletter Issue #4

June 30, 2024

Welcome to Issue #4 of our newsletter.

Things have been picking up steam behind the scenes at The Daily AI Show. We have been making a bigger effort to put out more content in the form of YouTube shorts and TikTok content as well as publishing the edited YouTube version of each show on the same day as the live version. Jyunmi has been leading that charge while Beth has been running the live comments section on Youtube. The result is our comments section is now the place to be during our live shows.

We are seeing hockey stick growth across the board and that is 100% because of you and your support. Thank you!

In this edition:

  • What Happens After AGI?

  • Are KANs the Next Evolution in Neural Networks?

  • Why Model Orchestration is The Future of AI App Development

  • Will Deepfakes Ruin Elections Before We Can Catch Up?

We also talk about Toys R’ Us’ new Sora AI ad, the Pillbot, Claude’s projects and its comparison to custom GPTs, the Singularity, Descript’s shot at eliminating video redo’s, and a 1900% increase in one important area of AI.

Let’s get to it!

-Andy, Beth, Brian, Eran, Jyunmi, Karl & Robert.

THANK YOU TO OUR SPONSOR

Skaled Consulting

We help sales teams increase productivity and review with ChatGPT and Generative AI.

AI for Sales isn’t just a buzzword – it’s a phenomenon that’s reshaping how sales professionals work, enabling them to be more efficient (productivity) and effective (revenue).

Embracing AI now is how you’re going to stay competitive and future-proof your sales strategy and business.

You are not just following the latest trend or getting a new tool; you’re adopting a new way to enhance your sales approach without replacing the human touch.

We can show you how.

Why It Matters

Our Deeper Look Into This Week’s Topics

What Happens After AGI? The Future of Work

On Monday, we talked about the somewhat daunting topic of what happens after achieving Artificial General Intelligence (AGI). We explored the potential societal shifts, changes in the workforce, and broader implications of AGI becoming a reality.

We also discussed the possible scenarios, from universal basic income to changes in job roles, and how society might adapt to a world where AGI can perform tasks currently done by humans.

WHY IT MATTERS

  • Redefining Work and Purpose: As AGI becomes capable of performing most tasks currently done by humans, there will be a fundamental shift in what constitutes work. People will need to find new ways to derive meaning and purpose beyond traditional employment, which could lead to an increased focus on creative, educational, and recreational pursuits.

  • Economic Disruption: The introduction of AGI could lead to widespread job displacement, particularly in knowledge-based industries. This raises significant economic challenges, including the need for robust social safety nets and potentially implementing universal basic income to support those displaced by AGI.

  • Mental Health and Identity: For many, work is a core part of their identity. The displacement caused by AGI could lead to increased mental health issues as people struggle to find new sources of identity and purpose. Society will need to develop new frameworks for mental well-being that address these changes.

  • Increased Inequality: Without proactive measures, AGI could exacerbate existing inequalities. Those who control AGI technologies could accrue significant wealth and power, widening the gap between the haves and the have-nots. Policies and regulations will be essential to ensure equitable access and benefits from AGI advancements.

Are KANs the Next Evolution In Neural Networks

On Tuesday, Andy helped us explore Kolmogorov-Arnold Networks (KANs), a cutting-edge neural network architecture offering significant improvements over traditional models. KANs, known for their efficiency, flexibility, and interpretability, are emerging as a potential game-changer in the AI landscape.

We discussed the unique features of KANs, their advantages over multi-layer perceptrons (MLPs), and their potential applications across various industries.

WHY IT MATTERS

  • Improved Efficiency: KANs can achieve higher accuracy with significantly fewer parameters compared to MLPs. For instance, a KAN with 200 parameters can solve problems that would typically require 300,000 parameters in an MLP. This efficiency translates to lower computational and memory requirements, making AI more accessible and cost-effective.

  • Energy Savings: Due to their compact size and efficiency, KANs consume less energy, addressing one of the major concerns with current AI models. This reduced energy consumption is crucial as AI systems increasingly impact global energy resources.

  • Real-Time Decision Making: KANs can process information quickly and accurately, making them perfect for applications that require real-time decision-making. This includes high-frequency trading, autonomous driving, and real-time language translation.

  • Future-Proofing AI Development: As AI continues to evolve, the scalability and adaptability of KANs ensure they remain relevant and capable of handling increasingly complex tasks. Investing in KAN research and development today can provide long-term benefits and maintain a competitive edge in AI innovation.

Developing AI Apps Using Model Orchestration

On Thursday, we discussed the role of model orchestration in AI application development. Model orchestration involves coordinating and managing multiple AI models, evaluations, tests, and workflows to create seamless and efficient AI applications.

We explored the tools and platforms that facilitate this process, such as Respell and Vellum, and how they are revolutionizing the way we build and deploy AI systems.

WHY IT MATTERS

  • Streamlined Development Process: Model orchestration tools like Vellum and Respell simplify the development process by providing a unified interface for managing multiple models and workflows. This streamlines the creation and deployment of AI applications, reducing complexity and development time.

  • Cost Optimization: These platforms enable developers to test and compare different models side-by-side, selecting the most cost-effective and efficient options. This ensures that businesses can optimize their AI investments, balancing performance with cost savings.

  • Scalability: As AI applications grow in complexity, the ability to orchestrate multiple models and workflows becomes essential. Model orchestration platforms provide the scalability needed to manage large-scale AI systems, making it easier to integrate new models and technologies as they emerge.

  • Future-Proofing AI Development: Investing in model orchestration tools prepares businesses for the future of AI development. These platforms are designed to evolve with technological advancements, ensuring that AI applications remain cutting-edge and effective.

AI Deepfakes and Elections

On Friday, we used the recent US debate to talk about AI deepfakes and their potential impact on the upcoming 2024 US election.

WHY IT MATTERS

  • Accessibility and Spread: The term "cheap fakes" refers to the ease with which sophisticated deepfakes can be created using readily available tools on smartphones. This democratization of deepfake technology means that even individuals with minimal resources can produce convincing fake content, making it harder to control the spread of misinformation.

  •  Trust Erosion: The widespread use of deepfakes erodes public trust in media and official communications. When people can no longer trust what they see and hear, it undermines the foundations of democracy and informed decision-making.

  • Technological Arms Race: As detection technologies improve, so do the methods for creating undetectable deepfakes. This ongoing arms race between creators and detectors of deepfakes necessitates continuous investment in research and development of more sophisticated detection tools.

  • Personalized Misinformation: AI’s ability to micro-target individuals with personalized deepfakes poses a significant threat. These tailored messages can exploit personal biases and manipulate individuals on a highly granular level, making it crucial to develop countermeasures that address both the creation and dissemination of such content.

HEARD AROUND THE AI COOLER
What We Are Chatting About This Week In Slack

Descript For The Win

Eran shared news that Descript will soon offer AI video regenerate which aims to eliminate re-recording. Our team uses Descript as part of our post-show processing, so we will be watching this closely to see if it helps with creating video shorts or doing minor editing. If it works, we will definitely share on a future show.

Insights from Iain Banks’ Culture Series 

Eran posted an after-show video talking about author Iain Banks’ book series that touches on a future society that doesn’t work or need money. Eran talked about how we change our relationship with work and focus on paying it forward.

@justin..1262 commented on the video and asked about the likelihood of a technological singularity and whether historical references and future projections will become less useful.

Beth responded by talking about Climate Futurist, Alex Steffen, and his reference to the Age of Discontinuity where enough factors are changing that the past data has less predictive power for current and future experience.

Go check out Eran’s video and see the entire conversation in the comments section.

Can Claude Projects Take A Bite Out of Custom GPTs?

Beth shared a tweet from Matt Wolfe saying he goes to Claude almost 100% of the time. And while we are looking forward to seeing what can be done with projects, Brian was saying a key missing element is real time internet searches. Until Claude can pull UTD info into a project, there will always be a place for the simplicity and ease of setting up a custom GPT. The low usage limits on Clause certainly do not help.

Keeping Us Up at Night:

@justin..1262 had some great comments on Monday’s show in the YouTube comments.

Justin said:

“It's fascinating to see the different perspectives on how AGI might impact jobs. While Andy thinks it will likely hit knowledge workers first and Brian feels it may also hit blue-collar workers, I absolutely agree with both positions but honestly could see it surprisingly affecting high-level white-collar jobs too.

What really keeps me up at night is how unprepared we seem. This could hit us like dominoes falling, forcing us into reactive mode. Now feels like the inflection point where we need to be proactive. We need to set ourselves up to make informed decisions when AGI arrives, not just scramble to respond.

Thanks for breaking this down in a way that makes sense. You've got me pondering big questions, but also thinking practically about what's next. Keep these thought-provoking episodes coming!”

Thanks Justin!
It keeps us up at night too.
The best thing we can do is keep talking about it.

Did You Know?

The global market for AI in enterprise applications is on the brink of a transformative explosion. Revenues in this sector are expected to skyrocket from $1.62 billion in 2018 to an astonishing $31.2 billion by 2025 (source: Statista). This represents an almost 1,900% increase, underscoring the profound impact AI is set to have on the business landscape.

Fact Details:

As AI technology matures and becomes more accessible, businesses across all industries are seeking innovative AI-driven solutions to gain a competitive edge.

Here are some key areas where AI is expected to revolutionize enterprise operations:

  • Enhanced Customer Experience:

    • AI-Powered Chatbots: Intelligent virtual assistants that can handle customer inquiries 24/7, providing personalized support and freeing up human agents for more complex tasks.

    • Sentiment Analysis: AI algorithms that analyze customer feedback in real-time, helping businesses tailor their products and services to meet customer needs more effectively.

  • Operational Efficiency:

    • Predictive Maintenance: AI systems that predict equipment failures before they happen, reducing downtime and maintenance costs.

    • Supply Chain Optimization: AI tools that forecast demand, optimize inventory levels, and improve logistics to ensure smooth operations.

  • Data-Driven Decision Making:

    • Business Intelligence: AI-powered analytics platforms that provide deep insights into business performance, enabling data-driven strategies and decisions.

    • Market Analysis: AI models that analyze market trends, competitor strategies, and customer behavior to identify new opportunities.

  • Automation of Routine Tasks:

    • Robotic Process Automation (RPA): AI-driven automation of repetitive tasks such as data entry, invoice processing, and compliance reporting, improving accuracy and efficiency.

    • Intelligent Document Processing: AI systems that can read, interpret, and process documents, reducing manual effort and speeding up workflows.

  • Innovative Product Development:

    • AI-Driven Design: Using AI to design new products, simulate performance, and optimize materials, leading to faster and more cost-effective development cycles.

    • Personalized Products: Leveraging AI to create customized products and services tailored to individual customer preferences.

Did You Miss A Show?
Catch Up Now On YouTube

How'd We Do?

Let us know what you think of this newsletter so we can continue to make it even better.

Login or Subscribe to participate in polls.