On the Way to Super AI-ntelligence

Hey there 👋

Welcome back to the second issue of 2023 from Unwind AI. Hope you are off to a great start to the year and have begun to progress toward your goals. As we focus on how to make this year the best one yet, it’s important to stay abreast with the incredible progress being made in the field of AI. And so we’re here to bring to you the latest news, insights, and trends and explore how these developments can help you achieve your goals this year.

So what are we waiting for? Let’s jump right in!

This issue covers:

1. Latest Developments 🌍

2. News from the Industry 🧑‍🏫

3. Tools of the Trade ⚒️

4. Hot Takes 🔥

5. AI Meme of the Week 🤡

Latest Developments 🌍

YOLO v8 🖼️

Are you ready for the next level of real-time object detection and segmentation? Ultralytics has launched YOLO v8, a cutting-edge, state-of-the-art model that builds on the success of previous YOLO versions and introduces new features and improvements to further improve performance and flexibility. It is designed to be fast, accurate, and easy-to-use, making it suitable for a wide range of object detection, segmentation, and classification tasks.

It offers well-documented workflows, spotless code, simple usage, and support for all YOLO versions, as well as being heavily maintained with the latest advances in AI.

NanoGPT 🐥

NanoGPT, an easy-to-use repository for training/finetuning GPTs, is now available. It is a rewrite of minGPT that gives more emphasis to training speed. The train.py file is able to reproduce GPT-2 (124M) on OpenWebText in 38 hours of training. It is highly customizable, allowing users to easily train new models from scratch, finetune pre-trained checkpoints, and make any desired modifications. For instance, the largest pre-trained checkpoint currently available for finetuning is the GPT-2 1.3B model from OpenAI.

VALL-E 🎙️

VALL-E is a neural codec language model for text-to-speech synthesis (TTS). It is trained using 60K hours of English speech, which is hundreds of times larger than existing systems. It can generate high-quality personalized speech by using just a short recording of a person's voice as a reference. This technology is better than the current state-of-the-art methods in terms of how natural the speech sounds and how similar it is to the reference recording. Additionally, it can also keep the emotions and the environment of the reference recording in the generated speech. This technology can be used for many different applications such as creating new speech from text, editing existing speech, and creating new content when combined with other AI models.

Check out this sample!

• Speaker Prompt:

• Ground Truth:

• Baseline:

• VALL-E

News from the Industry 🧑‍🏫

OpenAI at $29B Valuation 😲

Microsoft is reportedly in talks to invest $10 billion in OpenAI, the company behind the popular ChatGPT app. The funding, which would also include other venture firms, would value OpenAI at a whopping $29 billion.

Microsoft's investment in OpenAI might be a gamble with potential rewards as OpenAI is still trying to work out its business model. But Microsoft's cloud business will benefit from the investment, and it could gain 75% of any profits OpenAI makes from products such as ChatGPT and Dall-E. Most importantly, the investment puts Microsoft at the forefront of the development of potentially important consumer technology in the coming decade. (Source)

Project Voice 📣

Project Voice is a platform dedicated to accelerating the adoption of conversational AI and voice technology. It produces the podcast This Week In Voice which features interviews with industry executives, and a Substack newsletter, This Week In Voice VIP. They also run an annual event called Project Voice for conversational AI and voice technology. This year’s Project Voice 2023 will bring together the three co-founders of Siri for the first time in over a decade. It will also include speakers from Logitech, Karen Webster, Fandango, Stanford's MediaX Lab, Restaurant Business Magazine, The Science of CX, AWS, Intel, SoundHound, and other organizations. The event is expected to attract 2500 attendees and over 100 media outlets. Excited? Get yourself registered soon!

ChatGPT + Computational Knowledge Superpowers = 🤯

ChatGPT and Wolfram Alpha are two powerful AI technologies that have been developed with different approaches. ChatGPT was created to generate text that is similar to what humans would write, while Wolfram|Alpha is a powerful system for representing the world in formal symbolic ways. Can combining the two technologies through natural language create something much more powerful than either could do on its own? Will this alliance open up new possibilities for AI, where natural language and structured computation can work together to achieve beyond-human tasks? Check out this blog to get some perspective on the same.

Banning Writing from ChatGPT ❌

In a scenario where generative AI is extensively being used to create content, the International Conference on Machine Learning (ICML) has banned authors from using AI tools like ChatGPT to write scientific papers, sparking debate about the role of AI-generated text in academia. The ICML has clarified that its ban only applies to the text produced entirely by AI and does not prohibit the use of these tools for editing or polishing author-written text. However, there are still unanswered questions about the use of AI-generated text and images, such as who owns the output and whether it should be considered novel or a derivative of existing work.

Bringing back Dinosaurs to Life 🦖

Want to see extinct dinosaurs coming back to life? Check out this blog to generate real-life-like images of extinct species of dinosaurs via Midjourney AI using prompts from ChatGPT.

Tools of the Trade ⚒️

AI Playlist Maker 🎶

Here’s an app for creating the perfect playlist on Spotify and Apple Music for any occasion. PlaylistAI uses AI prompts, images, videos, and your most-listened-to music to create a playlist. It can:

• create playlists from any idea like “Early 2000's pop music” or “Playing board games on a rainy day”,

• turn any music festival poster into a playlist,

• make a playlist by identifying the songs in TikToks and other videos,

• and create a music festival lineup from your top artists from the past 1, 6, or 12 months.

(Isn’t that too cool? 😯)

Instant 3D Asset Generation🕋

Mirage is a platform for creating your own AI-generated 3D game assets by simply uploading an example image or inputting a prompt. The platform uses AI to generate 3D meshes and textures. It also features a community where users can explore assets created by others.

AI running Live Shows 📺

Elevate your live shows with an AI host, LiveReacting. It provides an interactive and engaging experience for the audience while saving you time and money. LiveReacting understands the context of the live show and acts accordingly, introducing quizzes, reading questions, talking about the topic, announcing winners, and interacting with players in real-time. LiveReacting is suitable for various industries such as educational institutions, influencers, and small pub owners hosting quizzes.

It has a host of features including:

• the ability to understand context during your live show,

• educate and entertain the audience with a mix of educational and entertaining content,

• support for multiple languages, and

• the ability to bring your own avatar.

Hot takes 🔥

GPT-4 is all you need 😅

AI - your Dating Wingman 👼

Watch what happened when this guy let GPT-3 talk freely to his matches on Tinder.

Microsoft ❤️ OpenAI continues

AI Meme of the Week 🤡

That’s all for this week!

Will see you next Saturday with more such content. Don’t forget to subscribe and give your feedback below!

BONUS 🎉

Share this newsletter with three other friends and stand a chance to win my book GPT-3: The Ultimate Guide to build NLP Products with OpenAI API. Winners will be selected on a monthly basis.

🎁 Every paid subscriber will also receive $39 USD worth of learning resources on trending topics like Python, Data Science, Machine Learning, and NLP!