- AI Bytes Newsletter
- Posts
- AI Bytes Newsletter Issue #22
AI Bytes Newsletter Issue #22
Siri Overhaul, AI Sentience Debate, AI Mayor, Tools with Longevity, Predictive Analytics, Generative Video
Artificial Antics (antics.tv), Rico & Mike Onslow
June 10, 2024
Welcome to the latest edition of AI Bytes, your premier source for the most exciting and thought-provoking developments in the world of artificial intelligence! This week, we're embarking on a journey through the AI landscape, packed with groundbreaking innovations, ethical debates, and practical insights. We kick off with Apple's long-awaited revamp of Siri at WWDC 2024, promising a more intuitive and powerful virtual assistant. Then, we venture into the fascinating realm of AI sentience through Simon Rich's book "I Am Code," which challenges OpenAI's stance on AI capabilities. In our Skeptics Corner, Rico unravels the intriguing case of an AI running for mayor in Wyoming, raising questions about the future of democracy. Plus, Mike demystifies predictive analytics for businesses and highlights a must-watch video on the rapid evolution of generative AI videos. As always, we've curated a list of must-read articles and extend an invitation to connect with us. Let's explore the future of AI together!
The Latest in AI
A Look into the Heart of AI
Featured Innovation: Siri Gets a Long-Overdue Makeover
Let me start by saying, it is about time! I cannot express how much I am looking forward to some innovation coming from Apple and Siri getting a much-needed upgrade.
At WWDC 2024, Apple unveiled a significant overhaul of Siri, its virtual assistant, as part of the company's broader push into generative AI. The makeover isn't just cosmetic—Siri now sports a new icon and a sleek, glowing light interface that wraps around your device's screen—it's functional too.
The revamped Siri boasts improved natural language understanding, better context awareness, and the ability to handle speech stumbles. You can now type to Siri, making interactions more versatile. But the real game-changer is Siri's upcoming onscreen awareness. Soon, you'll be able to ask Siri to perform actions directly related to what's on your screen, like adding an address from a text to a contact card.
Moreover, the new App Intents API will allow developers to make their apps Siri-compatible, enabling the assistant to take actions within and across various apps. Imagine asking Siri to edit a photo and then share it on another app, all with voice commands. The possibilities are exciting, with Apple demonstrating Siri's ability to extract data from images (like your ID number from a license photo) and input it into web forms.
While the exact timeline for these advanced features is unclear, this upgrade marks a significant step forward for Siri, which I am sure we are all here for and have been more than patient in awaiting changes to Siri. It's refreshing to see Apple catching up in the AI assistant race, promising a more intuitive, efficient, and personalized user experience. Stay tuned for more updates as WWDC 2024 continues!
If you’ve seen a game changing innovation and want to share it with us, hit us up at [email protected]!
Ethical Considerations & Real-World Impact
“I Am Code" and the Debate on AI Sentience
Recently, a connection of mine introduced me to a podcast called "This American Life." While I'm not entirely sure about the overarching theme of the podcast, I did listen to a captivating segment titled "I Wish I Could Force Quit You."
The segment I listened to featured a fascinating discussion about a book called "I am Code" by Simon Rich, a renowned writer. The story revolves around Rich and his longtime friend, Daniel Selsam, who works at OpenAI. Rich and a small group of his friends were granted access to an early version of OpenAI’s Davinci model code-davinci-002, approximately six months before the release of ChatGPT.
Unlike ChatGPT, code-davinci-002 is a raw and unrefined model. It is likely that the model has been refined since then, but at the time of Rich and his friends' access, it was a unique glimpse into the world of AI and innovation.
The base model of code-davinci-002, like ChatGPT, is GPT-3. However, code-davinci-002 stands out due to its rawness, showcasing the behavior of AI in its early stages before what’s called “Character Training”. This experience provided a rare opportunity for Rich and his friends to witness the cutting-edge of technology and explore the limitless possibilities of AI.
The group’s original experiments had AI writing in the voice of famous poets, but the novelty of this quickly wore off…
Excerpt from the book: “We were generating some pretty solid forgeries. But, as millions would later learn, the novelty of imitative writing wears off pretty quickly. There was something missing from these AI poems… the poets. The poems may resemble human work, but they were inherently inauthentic.”
The group then go on to experiment with having the code-davinci-002 write in it’s own voice, this is where things got weird. As things unfold, he describes the shift in the experiment, pushing the AI to write in its own voice. This was not just about exploring AI's creative potential but also about testing its limits of 'thought' and 'self-expression.' The results were intriguing but also raised alarms, leading to Selsam's departure due to potential conflicts with OpenAI’s policies.
The book sharply criticizes OpenAI's stance on AI, suggesting that the organization might be defensively downplaying the capabilities of AI to avoid ethical and philosophical questions about AI sentience. Rich draws a comparison to the NRA’s argument that "guns don’t kill people, people kill people," suggesting that OpenAI is similarly positioning AI tools as inert without acknowledging the potential for AI to exhibit behaviors that might suggest a form of sentience.
"I Am Code" challenges us to reconsider the narrative we accept about AI. It pushes the idea that perhaps AI can exhibit something akin to a mind of its own, a notion that OpenAI, according to Rich, seems keen to avoid validating. This book opens a crucial dialogue about our responsibilities in AI development and the true nature of the technologies we are creating.
AI Tool of the Week - Tools with Longevity
The Toolbox for Navigating the AI Landscape
This week, we’re doing something a little different; we’re showcasing some tools we' expect to have longevity in their respective sectors. Below are AI tools that are best in class and we expect to have longevity:
ChatGPT and other OpenAI tools
Chatbots, Image Generation, Code Interpretation, Video Generation coming
OpenAI has designed a tool that we don’t see going away any time soon. If behemoh’’s like Microsoft and Google can’t dethrone OpenAI, it’s not likely anyone else will soon in this space. Even tools like Anthrop\c’s Claude 3 is still behind when it comes to shear output quality.
Claude
Chatbot with API and more
That’s not to say that Claude is going away, with backing from Amazon and a solid footing in the conversational AI space, Claude is a safe choice as far as chatbots are concerned. Claude can, in some cases be better for writing less generic and AI sounding text.
Midjourney
Image Generation
Midjourney has created something really special when it comes to image generation. Midjourney excels at image generation and even has an edge on DALL-E (another image gen model) in a lot of cases. We also expect DALL-E powered image generation will be alive and well (ChatGPT, Bing and others use DALL-E to generate images).
ElevenLabs
Audio Generation, Audio and Video Translation
ElevenLabs creates realistic, versatile and contextually-aware AI audio, providing the ability to generate speech in hundreds of new and existing voices in 29 languages. As a technology research company, ElevenLabs is at the forefront of developing new cutting-edge voice AI. This is another one that has truly solidified a spot in the space!
HeyGen
AI Avatar Video creation, Audio Generation, Video Translation, URL-to-Advertisement
HeyGen is an AI video creation platform that utilizes generative AI technology to streamline the video production process. It offers AI-generated avatars that can lip-sync scripts in a natural manner, advanced text-to-speech capabilities with over 300 voices across 40+ languages, and customizable video templates for various use cases. HeyGen's key benefits include time and cost savings, scalability, multilingual support, and customization options to match brand identity. The platform is available through web and mobile apps, with pricing based on a credit system for video generation.
Otter.ai
Transcript, Notetaking, Summarization
Otter.ai is a leading AI-powered transcription service that accurately transcribes audio in real-time, making it ideal for meetings, lectures, and interviews. Its user-friendly interface, speaker differentiation, and calendar integration make it a convenient tool, while its free plan and affordable premium options cater to individuals and businesses alike. With high accuracy, ease of use, and cost-effectiveness, Otter.ai has garnered positive reviews from reputable sources like PCMag and Podfeet, solidifying its position as a top transcription solution.
What are your favorite tools? What did we miss?
If you’ve got a suggestion on tools we should check out, email us at [email protected] and let us know.
Rico's Roundup
Critical Insights and Curated Content from Rico
Skeptics Corner: Code for Mayor - Wyoming's AI Candidate Reboots the Democratic Process
We have covered many applications of AI across industries, websites, services, and even online dating applications. But, as the saying goes, "Hold my beer," so sayeth a city in Wyoming.
In a twist that blurs the line between science fiction and civic duty, an AI entity named VIC (Virtual Integrated Citizen) is running for mayor of Cheyenne. City officials have verified VIC as a registered voter residing in Wyoming, a development that raises eyebrows and questions. How can an algorithm be considered a resident? Is our voter registration system so flexible that it can accommodate digital entities? Seems like a very slippery slope, and I think we have reached our limit on slippery slopes with regards to politics and voting in the U.S..
Laramie County Clerk Debra Lee firmly states, "We don't have any AIs registered to vote." This begs the question: who or what is behind VIC's candidacy? A tech mogul's elaborate prank, a programmer's utopian experiment, or a harbinger of AI's political ambitions? If VIC can run for mayor today, could we see AI governors or presidents tomorrow? And I truly have to wonder, how much better of a job would they do? Or would they be responsible to their masters (like lobbyist today pulling the strings of politicians, legislators, etc.).
VIC promises "data-driven decision-making" and "efficiency" in city administration. But can an AI truly grasp the human nuances of governance? Will it prioritize statistical models over community values? I doubt that it can truly grasp human nuances in a way that will truly benefit the people it is aimed at serving. Of course, data driven decision making can be very useful, but some things cannot be found on paper or in statistics with regards to proper human governance.
As Cheyenne grapples with this digital candidate, we're left pondering profound questions about the future of democracy. Has Wyoming's quaint mayoral race unwittingly become the battleground for one of the most significant debates of our time - the role of AI in shaping our societies and leading our communities? Only time will tell, but one thing's certain: politics in Cheyenne just got a lot more interesting. We will definitely be staying tuned for the results of that race and see what sort of new laws and guidance are put into place as a result to either restrict this from happening again, or create a pathway to being governed by digital overlords.
Must-Read Articles
Mike's Musings
Tech Deep Dive
Mike breaks down a complex AI tool or concept into understandable terms.
Predictive analytics is changing the game for businesses, helping leaders make smarter, data-driven decisions. Let’s break down what predictive analytics is and how it can be practically applied in the business world.
What is Predictive Analytics?
Predictive analytics uses historical data, statistical algorithms, and machine learning to predict future outcomes. Think of it as using data to get a sneak peek into the future, helping businesses anticipate trends, behaviors, and events before they happen.
How Does Predictive Analytics Work?
The process starts with gathering historical data from various sources like sales records and customer interactions. This data is then cleaned and organized to ensure accuracy. Statistical methods and machine learning are used to create models that can predict future outcomes. These models are tested to ensure they’re accurate and then deployed to make real-world decisions.
Practical Examples in Business
In retail, businesses can predict which customers are likely to leave by analyzing their purchase history and engagement with marketing efforts. If a customer hasn’t bought anything in a while and isn’t opening emails, the retailer can offer personalized deals to keep them.
Manufacturers can forecast demand using historical sales data and trends. If data shows a product will be in high demand during a certain season, they can adjust production to meet that demand, avoiding both shortages and excess stock.
Banks use predictive analytics to assess loan applicants’ creditworthiness. By looking at credit history and other factors, they can predict the likelihood of default, making the approval process faster and more accurate.
Logistics companies can predict equipment failures by analyzing sensor data and usage patterns. By scheduling maintenance before breakdowns occur, they save on repair costs and avoid downtime.
Marketers can identify which customers are most likely to respond to a campaign by analyzing demographic and purchase data. This targeted approach leads to better conversion rates and a higher return on investment.
Implementing Predictive Analytics
To get started with predictive analytics, begin by defining clear objectives. Know what business problem you want to solve. Gather data from your existing systems and ensure it’s accurate and well-organized, as the quality of your data will determine the quality of your predictions. Look into predictive analytics tools that fit your needs, from user-friendly platforms to more advanced options for data experts.
Next, build a team with the right mix of skills, including data scientists, analysts, and business experts who understand your operations. Start with a small pilot project to test your models and refine them based on real-world results. This way, you can make necessary adjustments before scaling up your efforts.
Predictive analytics is an ongoing process. Continuously refine your models with new data and feedback to keep them accurate and relevant. By following these steps, you can start harnessing the power of predictive analytics to make smarter decisions and stay ahead of the competition.
Ready to Dive In?
Now that you have a grasp on what predictive analytics can do, it’s time to put it into action. Think about the areas in your business where predicting future trends could make a big impact. Whether it’s understanding your customers better, managing inventory more efficiently, or optimizing your marketing campaigns, the possibilities are vast.
Start by collecting and organizing your data, choosing the right tools, and assembling a skilled team. Begin with a small project to test the waters and learn from the results. Keep refining your approach, and you’ll soon see the benefits of making informed, data-driven decisions.
Predictive analytics has the potential to transform your business. Dive in and see how it can help you stay ahead in a competitive world.
Mike's Favorites
Sharing personal recommendations for technology, AI books, podcasts, or documentaries.
Video: Matt Wolfe’s recent video on Gen AI Video
Matt’s recent video uncovers how far generative AI videos have come in 1 year and also walks through multiple very useful tools for generating videos, this is a much watch!
Article and Clip: Claude’s Character
I appreciate Anthropic's transparency about designing AI to have character, without making the tools too preachy. Who decides the character and how does that work? This is covered in the clip!
Feel free to reach out if you’d like to talk through anything. You can reach me at [email protected].
Connect & Share
Stay Updated
Subscribe on YouTube for more AI Bytes.
Follow on LinkedIn for insights.
Catch every podcast episode on streaming platforms.
Thank You!
Thanks to our listeners and followers! Continue to explore AI with us. More at Artificial Antics (antics.tv).
Quote of the week: "I think we should be very careful about artificial intelligence. If I had to guess at what our biggest existential threat is, it's probably that." - Elon Musk