AI Bytes Newsletter Issue #31

Welcome to the 31st edition of AI Bytes! In this edition, we explore the latest AI innovations, including the groundbreaking Qwen2-Math models that are redefining mathematical reasoning, and Anthropic’s bold expansion of their model safety bug bounty program. Plus, we’ll look at how Rendora’s new text-to-3D video platform is pushing creative boundaries, and Rico offers insights on the evolving relationship between AI and human connection. As always, this issue is packed with essential AI tips and must-read articles to keep you ahead in the rapidly advancing world of AI. Let’s get started!

The Latest in AI

A Look into the Heart of AI

Featured Innovation
Qwen2-Math Revolutionizes Mathematical Reasoning in AI

The Qwen Team recently introduced the groundbreaking Qwen2-Math, a series of math-specific large language models that set a new standard for mathematical problem-solving in AI. Built on the robust Qwen2 framework, Qwen2-Math significantly outperforms even leading proprietary models like GPT-4o in math-related tasks.

Qwen2-Math models are rigorously trained on a diverse and meticulously curated Mathematics-specific Corpus, and their capabilities are proven through exceptional performance on both English and Chinese math benchmarks. The flagship model, Qwen2-Math-72B-Instruct, excels in complex mathematical competitions, showcasing its potential to tackle some of the most challenging math problems.

As the Qwen Team continues to push the boundaries of AI in mathematics, they are also preparing to release bilingual models supporting both English and Chinese, with plans for multilingual models on the horizon. Qwen2-Math is poised to be a pivotal tool for those requiring advanced mathematical reasoning and problem-solving in AI.

Ethical Considerations & Real-World Impact

Anthropic just made a bold move with their new "model safety bug bounty" program, and it’s the kind of action that could set a new standard in the AI industry. They’re offering cash to researchers who can find and report bugs in their AI models, putting safety and transparency right at the forefront. This goes beyond just looking to have staff fixing a few glitches; it’s also about being responsible and recognizing that as AI becomes more woven into our daily lives, the stakes are getting higher.

We’ve recently covered how OpenAI is cutting back on their safety and alignment teams, and that’s raising some red flags with many of us who are keeping a pulse on the industry. In contrast, Anthropic is doubling down on safety by actively inviting outsiders to scrutinize their models. They’re saying, “We know this tech is powerful, and we want to make sure it’s safe.” This proactive approach is what we need if we’re going to build AI systems that people can really trust—not just hype.

But this move by Anthropic also brings up some big questions. If they’re making safety a priority, what does that say about other AI companies that might not be? Will they follow suit, or are we going to see more corners cut in the name of speed and profit? The choices being made now are going to have a huge impact on how AI is developed and integrated into our lives. As AI keeps evolving, these decisions will shape not just the technology, but how it’s perceived and whether it ultimately helps or harms, or is even adopted at all by people.

AI Tool of the Week - Rendora

The Toolbox for using AI effectively

Rendora's Revolutionary Text-to-3D Video Platform

This week, we're spotlighting Rendora, a trailblazing platform that has taken the creative world by storm with its groundbreaking text-to-3D video technology. Unveiled at SIGGRAPH 2024, Rendora empowers users—whether enterprises or individual creators—to transform simple text into stunning, fully animated 3D videos within seconds. By leveraging a comprehensive suite of generative AI tools, Rendora offers an all-in-one solution for producing high-quality 3D content that rivals professional studio outputs. With just a few clicks, users can customize avatars, scenes, and camera angles to create lifelike animations that are perfect for marketing, education, training, and more.

What truly sets Rendora apart is its ability to democratize 3D content creation, making it accessible to everyone, regardless of technical expertise. The platform's hyperrealistic 3D AI avatars and advanced AI cinematography ensure that the final product is not only visually stunning, but also perfectly aligned with the intended narrative. As the demand for immersive digital content continues to rise, Rendora's innovative approach to 3D video production positions it at the forefront of the industry's future, making it a must-have tool for any content creator looking to push the boundaries of their creative potential. Try it out and let us know what you think.

Rico's Roundup

Critical Insights and Curated Content from Rico

Skeptics Corner
When AI Speaks: Navigating the Human Connection in a Digital World

"Life imitates art far more than art imitates life," Oscar Wilde once said. This sentiment feels particularly poignant as we are catapulted into an era where artificial intelligence is finding its voice…literally. The recent developments in AI voice technology bring to mind the 2013 film "Her", which Mike and I have referenced on the podcast and in the newsletter many times. In the movie, “Her”, Joaquin Phoenix's character falls in love with an AI assistant voiced by Scarlett Johansson. What seemed like science fiction a decade ago is now knocking on our reality's door.

OpenAI's new voice mode is pushing us into uncharted territory, further blurring the ever-thinning lines between human and artificial interaction. This technology is designed to captivate, much like Samantha in "Her" captivated Theodore. We've already seen people form bonds with virtual assistants like Siri or Alexa, treating them as digital companions (or their personal verbal punching bags when they cannot complete a task). Now, with AI voices becoming more lifelike, the boundary between real and artificial is getting even fuzzier.

While the features and concepts are usually interesting, they also get a bit more complicated. Imagine an AI that not only understands what you're saying but responds with spot-on tone and emotion, making the conversation feel personal. It's reminiscent of how Theodore and Samantha's relationship evolved in "Her," becoming deeply intimate despite the lack of physical presence.

But what happens when someone starts leaning too heavily on this artificial emotional connection? An AI programmed to sound comforting and empathetic might lead folks to swap real human relationships for digital ones. Which, on it’s face, may serve many for the purpose of good in preventing loneliness, assist folks with practicing speech, new languages, etc. However, it's important to remember that no matter how smart the AI gets, it's still just a tool—not a true friend or lover, as Theodore painfully discovered in "Her."

The ethical terrain here is complex and the waters are muddied. Giving AI such a lifelike voice could open doors to manipulation or deception, which we have discussed on the podcast episode with Dave George and also in our AI roundtable, in which we gave examples in which AI had been caught being deceptive. Imagine a scam where a trusted voice turns out to be an AI, or an AI subtly influencing you against your best interests. It's like handing someone a superpower without an instruction manual, and unfortunately, we know far too many nefarious actors exist here in 2024.

On the flip side, the positive potential is huge, and the uses for this tech are countless. For those feeling isolated or lonely, having a voice to talk to, even an artificial one, could be a comfort and potentially even prevent cognitive decline. And for people who struggle with typical tech interfaces, interacting in a more natural way could be game-changing. In "Her," we saw how Theodore's relationship with Samantha initially helped him overcome his loneliness and writer's block. Don’t worry, we won’t spoil the ending for you here, but do recommend the film, especially as AI emerges and is more openly adopted in the coming years and months.

Ultimately, like anything else in life, it's about striking a proper balance. We need to embrace these advancements while setting clear boundaries, and being aware of the impact they have on us. As tech becomes more human-like and easier to use, we've got to keep our own humanity in check. It truly is a brave new world that is on the horizon before us, and we're the ones who need to navigate it responsibly.

Like Theodore in "Her," we're stepping into a future where the lines between human and artificial intelligence are increasingly blurred. But unlike Theodore, we have the benefit of foresight (and the AI Bytes Newsletter to keep you informed). We can shape how this technology integrates into our lives, ensuring it enhances rather than replaces human connection. Even such a polarized time where I am sure many of us would prefer a kind AI voice over the vitriol that has become social media and the internet at large. That being said, as we move forward with these technologies, let's remember Wilde's words and consider: How much of our future reality is being shaped by the art and technology we create today? I guess we will see, and as always, please let us know what your take is. We would love to hear from you all.

Must-Read Articles

Mike's Musings

AI Tip - Daily Hacks to Improve your Workflow!

Hey all! this week, we’re going to get practical with 5 hacks to improve your AI workflows.

One thing that helps is to turn off OpenAI training on your data (this is not surefire, use your best judgement). To turn off training in ChatGPT, go to Settings / Data controls and turn “Improve the model for everyone” to Off

Without further ado, let’s get started!

1. Speed Up Email Responses

We’ve all been there—staring at a packed inbox, feeling the stress mount as more emails pour in. The time it takes to craft thoughtful responses can eat away at your day, leaving you with less time to focus on bigger tasks. That’s where AI comes in. Imagine being able to breeze through your inbox with AI helping you craft the perfect response in seconds.

Here’s How I Do It:

1. Fire Up Your AI Tool: I usually go with ChatGPT, but Copilot, Claude 3 (Claude 3.5 Sonnet is impressive for writing) or Gemini work just as well.

2. Copy the Email: I grab the email content by selecting it and hitting Ctrl+C.

3. Ask AI for Help: I jump into my AI tool and paste the email, then I ask something like, “Can you help me respond to this?”

4. Lay Out the Key Points: I tell the AI what I want to include—maybe I’m agreeing to something but also asking for something in return, like time off to balance extra work.

5. Review and Polish: The AI drafts a response for me, which I tweak if needed, and then I’m good to go.

Pro Tips:

• Always review and refine the outputs. The easiest way to have AI take your job is to just copy and paste what it gives you as outputs without putting thought into yourself. Keep the human side of the equation in play!

• Be careful what you share with chatbots and AI tools. Strip sensitive data or PII (personally identifying information) out of emails if you are using the free version of ChatGPT or other tools.

2. Simplify Data Analysis

You’ve got a mountain of data staring back at you, and the thought of manually sifting through it all is daunting. Whether it’s sales figures, customer feedback, or operational metrics, data analysis can be time-consuming and prone to error. But what if you could have an AI assistant handle the heavy lifting, turning raw data into actionable insights in minutes?

Here’s How I Tackle It:

1. Get Your Data Ready: I make sure my data’s in an Excel file, CSV, or Google Sheets.
   

2. Upload It: I upload the file into ChatGPT (or your AI tool of choice) by clicking the paperclip icon. Make sure your data is free of PII or other sensitive data.
   

   

3. Start Asking Questions: I fire off questions like “How many X units did we sell in 2022?” and the AI breaks down the data for me.
   
   

   

4. Make It Visual: If I need charts, I just ask the AI to whip up a pie chart or bar graph, and it delivers.
   
   

   

5. Double-Check: I’ll pop back into Excel to make sure the AI’s numbers match up with my own in a spot check, just to be safe.

3. Effortless Document & Image Understanding

Picture this: you’re handed a dense contract or a lengthy report, and you need to pull out the critical details fast. Or maybe you’ve got an image with important information that you need to translate into text. Manually going through these documents can be a slog, but AI can speed up the process, giving you quick, accurate summaries or extracting text with ease.

How I Get It Done:

1. Upload Your Doc or Image: I’ll start by uploading a PDF, Word doc, or even an image into my AI tool.
   

2. Ask Away: I ask the AI questions about the document—stuff like, “Tell me about this flow” or “What issues might I be concerned about in this contract?“

3. Get the Insights: The AI breaks it down for me, summarizing key points or giving me specific answers.

4. Create a Custom GPT: If I know I’ll be referencing the same doc a lot, I save it as a Custom GPT so I don’t have to re-upload it every time. I’ve got a tutorial that goes in-depth on how to create your own Custom GPT https://artificialantics.beehiiv.com/p/value-creation-with-custom-gpts-from-idea-to-savings

5. Save and Review: Once the AI does its thing, I save the insights for later use.

4. Quick Text Extraction

You just wrapped up a brainstorming session, and the whiteboard is covered in ideas and notes. The challenge? Capturing all that information and turning it into something usable. Taking photos is one thing, but how do you get all that text into a format you can edit and organize? This is where AI’s text extraction capabilities shine.

How I Handle It:

1. Snap the Image: Whether it’s a whiteboard after a meeting or a handwritten note, I take a photo or screenshot.

2. Upload It: I upload the image into ChatGPT using the paperclip icon.

3. Tell the AI What You Need: I might say, “Can you transcribe the writing on this whiteboard?” and the AI spits out the text.
   

   

4. Copy and Use: Once the text is extracted, I copy it into my document, email, or whatever tool I’m using.

5. Edit and Organize: I usually tidy up the text and organize it for future reference.

5. Receive Instant Feedback

Whether you’re designing a logo, drafting a report, or crafting a presentation, getting feedback is crucial—but not always easy to come by. Waiting on colleagues or friends for input can slow you down. What if you had an AI tool that could give you instant, detailed feedback, helping you refine your work in real-time?

My Approach:

1. Prep Your Work: Whether it’s a logo design, a document, or a script, I get it ready for review.

2. Upload or Paste It: I’ll upload the file or paste the text into my AI tool.

3. Ask for Feedback: I ask something like, “Critique this logo.
   

   

4. Review AI’s Take: The AI gives me feedback, highlighting strengths and weaknesses, and sometimes it even makes recommendations.

5. Refine Your Work: I take the AI’s suggestions and use them to improve my work.

By incorporating these detailed steps into your workflow, you’ll not only save time but also enhance the quality of your work, making AI an invaluable tool in your everyday tasks.

Mike’s Favorite

Top AI Conferences of 2024

As the AI/ML space becomes more mature, we’re seeing more and more events pop up related to AI. Some of these are meetups, some are conferences. As someone that loves engaging with other folks, I figured I’d check out this list and share it with you!

If you’ve got something you think I’d like, hit me up at [email protected] 

Latest Podcast Episode

• Listen on:

  • Apple

  • Spotify 

  • All other major podcast networks

Connect & Share

• Have a unique AI story or innovation? Share with us on X.com or LinkedIn.

• Collaborate with us: Email us.

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: "AI is likely to be either the best or worst thing to happen to humanity." — Stephen Hawking