The Daily AI Show Newsletter Issue #12

Mentos + Diet Coke = Country Rocket Man?

In this issue, we get into:

• Are AI Robots Really The Future We Want?

• AI In Fast Food: Who Really Benefits?

• Does The 80/20 Rule Apply to AI?

Plus, we talk about AI robots and building IKEA furniture, the misleading name Agent.AI, AI songs about mentos and Diet Coke, KenTacoHut, AI mosquito zappers, and a bunch more.

It’s Sunday morning. No better time to nerd out and learn a lot about AI.

We hope you love this issue as much as we loved putting it together for you.

Your DAS Crew
-Andy, Beth, Brian, Eran, Jyunmi, Karl & Robert

Are AI Robots Really The Future We Want?

The future of AI robots in our daily lives is closer than ever, but it’s not without significant challenges. Andy introduced the concept of "embodied AI," where artificial intelligence is not just a program but something integrated with a physical body, capable of interacting with the world much like a human. "Embodied AI is the grand promise of a future where an artificial intelligence also has a body that it can use to learn about the world and act in it," he explained.

The discussion quickly moved to the practicalities of having such robots in our homes. Karl, who’s been a proponent of robots taking over mundane tasks, like taking out the recycling, questioned how soon we’ll actually see robots that can handle such complex, multi-step tasks. "It's a lot of complex tasks, which I haven't seen any robot do all at once," Karl noted, pointing out the intricacies involved in even the simplest household chores.

Beth added another layer to the conversation by asking how different members of a household might interact with these robots. She speculated on the dynamics of control and understanding, especially when younger, more tech-savvy individuals might manipulate the robot in ways that older adults wouldn’t. "Is the recycling taken out by the younger people in most households? Maybe that's coming," she pondered.

The ethical implications were not overlooked, with Jyunmi expressing concerns about robots that mimic human emotions too closely. He argued that until robots are perfected, they might cause more discomfort than convenience. "I think if you made all the robots non-humanoid or for a specific task, it might be more acceptable," he suggested.

#WHY IT MATTERS

• Embodied AI Potential: The concept of AI with physical forms opens up new possibilities for how we interact with machines, but it also presents significant challenges in terms of development and practical application.

• Complex Task Management: The ability for robots to handle complex, multi-step tasks autonomously is still a significant hurdle, raising questions about how soon we can expect functional AI robots in our homes.

• Household Dynamics: As robots become more integrated into daily life, the way different members of a household interact with and control these machines could lead to new social dynamics and challenges.

• Ethical Considerations: The closer robots come to mimicking human emotions and behaviors, the more ethical considerations come into play, particularly around how we treat these machines and what rights, if any, they should have.

• Specialized vs. Humanoid Robots: There is ongoing debate over whether robots should be humanoid or specialized for specific tasks, with implications for their acceptance and utility in everyday life.

Will AI Revolutionize the Fast Food Ordering Process?

AI is rapidly transforming the fast food industry, reshaping how customers interact with their favorite chains. From self-service kiosks to AI-driven voice ordering systems, these technologies aim to enhance efficiency, reduce labor costs, and improve the overall customer experience.

As Brian observed, "It takes the pressure off for you to necessarily move fast or maybe not take your time with the order," referring to the convenience of McDonald's self-ordering kiosks he experienced in Europe. These kiosks, which have been around for years, allow customers to order at their own pace, reducing stress and improving satisfaction.

Jyunmi discussed the ultimate goal of AI in fast food: "The ideal end result is a small kitchen with robotic workers where you order via an app or AI system and pick up your food with little to no human interaction." This vision of minimal space, minimal labor, and rapid service reflects the industry's push towards automation.

Andy provided a historical perspective, comparing modern AI innovations to the early 20th-century automats, which were essentially the first self-service restaurants. He noted, "The promise of AI in ordering is conversational AI," highlighting the potential for AI to further streamline the ordering process by replacing human interactions with sophisticated, multilingual chatbots.

Beth raised an important point about the complexities of food preparation and the challenges AI faces in maintaining food quality. She questioned whether AI can manage the intricacies of food preparation, asking, "What is the higher order reasoning that we have achieved in the robot or in the AI automation that is doing that?"

#WHY IT MATTERS

• Efficiency and Convenience: AI-driven systems are making fast food more accessible and less stressful by allowing customers to order at their own pace, reducing wait times, and minimizing human error.

• Automation in the Kitchen: The integration of robots in food preparation, as seen with examples like Flippy the fry cook, represents a significant shift towards fully automated kitchens, potentially reducing labor costs and increasing consistency.

• Conversational AI: The future of fast food may lie in conversational AI, where orders are taken by AI systems that can understand and respond to customer preferences in real-time, enhancing the customer experience.

• Food Quality and Customization: While AI can streamline processes, the challenge remains in ensuring that food quality and customization meet customer expectations, particularly in complex or specialized orders.

• Impact on Jobs: The rise of AI in fast food could have significant implications for entry-level jobs, which are often a crucial part of the workforce, raising concerns about the socio-economic impact of widespread automation.

The Pareto Problem and AI: Does the 80/20 Rule Really Apply?

The Pareto Principle, often summarized as the 80/20 rule, suggests that 80% of outcomes come from 20% of efforts. This principle has been widely observed in various fields, but does it hold up when applied to AI? But how does this idea interact with AI efficiency and productivity? Can AI truly shift the balance of effort and results.

Brian began by referencing a LinkedIn post by Ethan Mollick, which sparked the discussion: “When an expert says an AI can do 80% of the work, that’s not usually a Pareto problem.” This comment highlights the misconception that AI merely handles the mundane tasks, leaving humans to do the difficult 20%. However, as the team explored, AI often enhances human capabilities, allowing us to focus on more creative and innovative tasks.

Andy took this a step further, suggesting that AI could actually redefine the Pareto curve. “AI might be able to do 95% of the work in just 5% of the time,” he noted, indicating that AI’s efficiency could drastically alter traditional models of effort and return. Beth added that the interaction between AI and human expertise creates a synergy where the whole is greater than the sum of its parts—a concept she humorously dubbed the “Pareto-gether” principle.

#WHY IT MATTERS

• Redefining Efficiency: AI has the potential to drastically alter the traditional 80/20 rule, possibly allowing for greater outputs with even less input, challenging our current understanding of productivity.

• Human-AI Collaboration: The combination of AI handling routine tasks and humans focusing on creative, high-level work could lead to unprecedented levels of innovation and efficiency.

• Expertise Enhancement: AI doesn’t just replace human effort; it enhances it, enabling experts to push the boundaries of what’s possible in their fields.

• Broader Implications: As AI continues to evolve, it may lead to shifts in how we define work, effort, and value across various industries, potentially disrupting established business models and societal norms.

JUST JOKES

DID YOU KNOW?

GenAI market value was $29B in 2022.

It is expected to reach $66B by the end of this year.

That’s a 128% increase in 2 years
According to Statista

Join Us For This Week’s Shows

Monday, August 26th - Is it easier to believe a real image is fake than to spot a fake one?
Discover how AI is reshaping our perception of reality.

Tuesday, August 27th - Thinking about fine-tuning an AI model?
Discover when it's truly necessary and the key steps to take before diving in.

Wednesday, August 28th - What surprising AI news stories will the DAS crew uncover this week? Join us live.

Thursday, August 29th - RAG systems can revolutionize AI, but do you know how to prompt them effectively?
Uncover the secrets to mastering prompts for optimal results.

Friday, August 30th - Are GPT Actions finally living up to their potential, or are they still falling short?

This Week’s Conundrum
A difficult problem or question that doesn't have a clear or easy solution.

Should generative AI models be rigorously filtered to remove biases, even if it means potentially altering or censoring parts of the data they are trained on?

Or should they be left unfiltered to fully represent the complexity of human expression, even if that includes harmful biases?

NVIDIA's Stormcast for Weather Forecasting

NVIDIA has introduced Stormcast, an AI model capable of forecasting thunderstorms with high precision over a few kilometers. Jyunmi explained, "AI models traditionally struggle with weather variables, but Stormcast nearly matches the current NOAA HRRR system in accuracy." This development could revolutionize how we predict and respond to severe weather events.

NVIDIA's Nemotron 4B Instruct for Gaming

NVIDIA also unveiled Nemotron 4B Instruct, designed to enhance conversational abilities in game characters. Jyunmi highlighted, "This on-device model improves role play and function calling, potentially extending its use beyond gaming into mobile apps and desktop interactions." The model could pave the way for more interactive and responsive AI in gaming and other applications.

Data Center Power Demand Surge

The increasing demand for AI-driven data centers has led to a 515-megawatt increase in power supply in North America in the first half of 2024. Andy noted, "The demand for transformer-based inference is driving massive growth in data centers, yet it still might not be enough to meet the AI industry's needs."

Final Spark's Brain Cell-Based Biocomputers

Final Spark, a Swiss startup, launched a service allowing scientists to rent cloud access to brain cell-based biocomputers for $500 a month. Beth emphasized the significance, stating, "These biocomputers are 100,000 times more efficient than current data centers, which could drastically reduce energy consumption."

AI for Government in Wyoming

Victor Miller, a mayoral candidate in Wyoming, has promised to run the government with the help of an AI bot named Vic. Karl found the concept intriguing, noting, "While it's unlikely he’ll win, using AI to assist in decision-making could lead to more efficient and transparent governance."

Eric Schmidt on AI Agents

Former Google CEO Eric Schmidt discussed the future impact of AI agents, predicting that by next year, they will be capable of independently achieving complex objectives. Andy highlighted, "Schmidt believes we're not prepared for the explosive transformation AI agents will bring to business operations."

Meta's Self-Taught Evaluator for LLMs

Meta has released research on a new method called the self-taught evaluator, which allows LLMs to improve themselves without human feedback. Andy explained, "This could eliminate the need for human reinforcement in training, making AI models more autonomous and efficient."

Unitree Robotics' Affordable Humanoid Robot

Unitree Robotics has optimized its $16,000 G1 humanoid robot for mass production. Andy pointed out, "While the AI component might not be as advanced, Unitree is making significant strides in making humanoid robots more accessible."

Washington State's Safe Robotics Algorithm

Washington State University developed an algorithm that increases robot safety by 80% by making them more aware of human inattentiveness. Jyunmi explained, "This innovation could greatly reduce accidents in environments where humans and robots work closely together."

Renovate Robotics' Rooftop Robot

Renovate Robotics introduced Rufus, a rooftop robot that can install shingles and, in the future, solar panels. Beth commented, "This robot not only reduces the risk of human injury but also makes solar panel installation more accessible."

MIT's Tiny Batteries for Nanobots

MIT has designed tiny zinc-air batteries to power cell-sized robots, which could be used for medical applications like drug delivery. Jyunmi remarked, "These batteries could revolutionize how we approach nanotechnology and medical treatments."

Trunk Tools' Construction AI Platform

Trunk Tools secured a $20 million Series A for their AI platform designed for the construction industry. Andy explained, "Their platform organizes vast amounts of documentation and links it with project schedules, making construction management more efficient."

Eight-Year-Old Builds Harry Potter Chatbot

An eight-year-old used Cursor and AI to create a Harry Potter chatbot without any coding knowledge. Karl was impressed, noting, "This example shows how accessible AI tools are becoming, even for young children."

Waymo's Self-Driving Success

Waymo announced that it is now completing 100,000 autonomous rides per month, primarily in San Francisco. Andy shared, "Their Gen 5 vehicles are larger and more advanced, signaling rapid growth in the self-driving car industry."

BZIGO's AI Mosquito Detector

BZIGO launched IRIS, an AI-powered mosquito detector that tracks mosquitoes with a laser pointer. Andy humorously added, "This might just be the gadget I need, especially with its ability to work in the dark using infrared."

Did You Miss A Show Last Week?

Enjoy the replays here on YouTube or take us with you in podcast form on Apple Podcasts or Spotify.