⚔️ The Fierce Face-off in Generative AI

Hello AI Enthusiast,

Last week's OpenAI turbulence left many ChatGPT users pondering the reliability of their favorite AI tool. It's a big question now: should AI product developers start considering a backup plan? Meanwhile, this week has been a display of new releases, with tech giants and startups unveiling updates and new features, intensifying the competition in the generative AI business.

As the AI landscape shifts and evolves, one thing remains constant - the ability of creative individuals to take these advancements and build something unique. Gone are the days when developing sophisticated AI-driven products required extensive technical knowledge and months of development. We’re living in an era where anyone with curiosity and a problem-solving spirit can leap into creating AI products without coding.

The projects of our students from the current third edition of the Master in Prompt Engineering are shining examples of this new wave of creators:

• 🍏 Diet Assistant: Guiding healthier food choices through tailored recommendations.

• 📧 Customized Cold-Calling Email Generator: Elevating business outreach with personalized emails.

• 🐾 A Dog Rescuer Platform: Connecting rescue dogs with their ideal new families.

• 🤝 Candidate Summaries for HR: Streamlining the hiring process with automated candidate insights.

• 🎯 Goal Assistant: Aiding focus and perseverance in achieving personal and professional goals.

This is the reality of AI today – a field not just for the tech elite but for anyone with a dream to innovate and create. If you're itching to make your mark, our Master in Prompt Engineering is your gateway. But don't wait too long; the 300€ early bird discount is almost up, ending tomorrow!

And now, let's get you updated with this week's News Bytes. There's plenty to catch up on in the ever-exciting world of AI!

News Bytes 🗞️

• ChatGPT's voice chat feature, previously exclusive to premium users, is now accessible to everyone, offering a choice of five unique voices. This expansion comes despite the recent tumult at OpenAI, showcasing the team's resilience in enhancing ChatGPT's capabilities.

• In this context, Inflection AI steps up with Inflection-2, a user-friendly AI that excels in factual knowledge and reasoning aiming to power their personal AI product, Pi. This model is trying to offer a viable option for those needing dependable and sophisticated AI capabilities.

• Anthropic releases Claude 2.1, an AI model with its 200K token capacity (roughly 150k words), 2x reduction in false statements, and enhanced accuracy and summarization, emerges as a strong contender in the rapidly evolving generative AI field. Features like enhanced comprehension, API tool use, and improved honesty in Claude 2.1 could attract users seeking reliable and sophisticated AI platforms.

• Anthropic didn’t stop here. It has reduced its per-token pricing. This move, aimed at retaining clients and competing with both closed-source firms and open-source models, intensifies the debate around the potential for free AI models. Gianluca Mauro, our founder, dove deeper into this topic in an Instagram video.

• Google's Bard AI has also got itself an upgrade. It now offers a notable YouTube integration, analyzing videos to extract specific details like key points or main topics. This feature, while convenient for users, raises concerns about its impact on content creators, especially when it bypasses paywalls and traditional content consumption methods, potentially affecting creators' revenue.

• Elon Musk couldn't resist jumping into the AI fray, and now his chatbot Grok is set to debut this week for X's Premium+ subscribers. Last week, teasers of Grok's integration in X's web app hinted at its imminent release, marking Musk's entry against rivals like OpenAI's ChatGPT and Google's Bard. Grok could potentially liven up X's Premium subscription amidst a tough phase for the company.

• In a "surprising" turn of events, Amazon throws its hat into the crowded AI chatbot ring with 'Q', unveiled at the AWS Reinvent conference. Designed to ease AWS navigation and integration with numerous enterprise systems, Q is competitively priced for business users. This move marks Amazon's latest stride in challenging tech giants like Microsoft and Google in the productivity software domain.

• Google's playing it cool in the AI race, pushing back the launch of its GPT-4 rival, Gemini, to early 2024. Amidst the rush from OpenAI and Elon Musk's ventures, Google is taking extra time to polish Gemini, aiming to ensure it lives up to its high expectations and stands out in the competitive AI landscape.

• Stability AI enters the growing AI video-generating arena with its new model, Stable Video Diffusion. This open-source offering aims to transform industries like advertising and entertainment, but it's not without concerns, such as unclear training data origins. Amidst these challenges, Stability AI is keen to carve out its niche in this burgeoning field.

• In the unfolding narrative of AI video generation, Pika Labs has just secured a significant $55 million in funding. Pika 1.0, aims to revolutionize video creation with AI-powered tools that can convert captions and images into various video styles. This surge of innovation in AI video generation is part of a broader trend where startups are rapidly advancing, challenging tech giants.

• ElevenLabs has pushed the boundaries of speech technology by adding Speech to Speech (STS)¹ to their Speech Synthesis. STS can mimic the emotional content and tone of a voice, making it sound like another person entirely. As Gianluca mentioned in his recent workshop, AI is more than capable of faking emotions. Does it make it a psychopath?

• To keep making these AI models more powerful, Scale AI is collaborating with NVIDIA to develop next-generation language models (LLMs). By combining Scale's customized training datasets with NVIDIA's SteerLM technique, developers can create applications for industry-specific purposes.

This week’s glossary 📖

1. Speech to Speech (STS): A technology that transforms spoken words from one voice into another, preserving the original message while altering the voice's identity and characteristics.

Educational Pill 💊

Exploring Speech to Speech (STS) Technology

Unlike Text to Speech, which turns text into voice, STS changes one spoken voice into another. It's like having the ability to take what someone says and repeat it in a different voice, complete with its own style and emotion.

But with great technology comes great responsibility. Imagine if someone could copy your voice without you knowing – it could lead to confusion or even misuse of your identity. Also, STS could make fake voices sound very real, which might make it hard to tell if what we're hearing is true or made up. And then there's the question of keeping our voices private. In a world where any voice can be imitated, we need to be careful about protecting our own.

As we learn more about STS, it's important to think about how we use it responsibly. It's not just a cool tech trick; it's a tool that needs to be used wisely.

From our community 🤝

With all the news about changes at OpenAI and new AI tools coming out, our community's been talking about the need for trust and practicality in AI companies, highlighting how these values endure beyond industry turbulence.

That is the end of our newsletter.

Remember, if your company is looking to implement AI technologies, we also offer customized corporate training.