Hey everyone!

Welcome back! Guess what… it’s time! We’re dropping what we’re working on in this Monday’s edition of the newsletter! All we can say for now is that if you’re a MidJourney enthusiast, you can’t miss it!

Until then, here’s what we’ll learn about today:

• AI enables a new Beatles Song 🎺!

• Accenture Invests $3 Billion in AI 💰!

• A new AI camera that does not take photos📸!

• OpenAI keeps making ChatGPT better ℹ️

• Create Professional Product images in seconds using Flair.ai!

Reading time: 5 minutes 16 seconds

News Reel:

Highlight 1: Are you ready for the final Beatles’ Song?

Sir Paul McCartney has revealed that he employed artificial intelligence to extract John Lennon's voice 🗣️from an old demo and complete what he calls "the final Beatles record." Here are the bits:

• The specific song has not been named but is likely to be a 1978 Lennon composition called "Now And Then."

• Lennon's widow, Yoko Ono, had given McCartney the demo a year before Lennon's death in 1980🎶.

• Producer Jeff Lynne helped clean up the demo, resulting in the completion and release of two other Lennon songs, "Free As A Bird" and "Real Love," in 1995 and 1996.

• The Beatles attempted to record "Now And Then" but abandoned the session due to various issues🚫.

• Peter Jackson's ‘Get Back’ documentary played a crucial role by training computers to recognize the Beatles' voices and separate them from background noises, allowing McCartney to "duet" with Lennon on his recent tour.✌️

The announcement was made during an interview on BBC Radio 4 ahead of the launch of McCartney’s book and photography exhibition titled ‘Eyes of the Storm.’

Are you excited? We sure are!

Here’s a more detailed article from the BBC.

Highlight 2: Accenture Invests $3 Billion in AI

If you didn’t think AI is the future already, here’s another giant investing a ridiculous amount of money in AI. Accenture (the IT services and consulting giant) commits $3 billion to its Data & AI practice, aiming to support clients across 19 industries in leveraging data and AI for value generation, while doubling its AI talent pool to 80,000 professionals.

• Accenture is actively collaborating with clients on generative AI initiatives in various industries🤝.

• The "AI Navigator for Enterprise" platform will help clients define business cases and navigate AI journeys responsibly.

• Accenture will invest in new and existing relationships within its cloud, data, and AI ecosystems. 🧩

• Salesforce also announced an investment in generative AI startups 🚀and the launch of its AI Cloud service.

Here’s some more information from people matters for those interested!

Highlight 3: A revolutionary new AI camera that captures photos without a lens!

Imagine a camera that captures photos without a lens or sensor, using only AI and location data to generate unique images - introducing the Paragraphica camera by Bjørn Karmann. Here’s how it works:

• The Paragraphica camera, created by Bjørn Karmann, captures photos without a lens or sensor. Instead, it uses AI and location data from Google Maps to generate images. 🗺️

• The camera collects information about the location, weather, time of day, and nearby landmarks.

• It then uses a text-to-image AI model to transform the gathered data into a unique photograph. 💬

• The camera has three adjustable knobs that control the scanning area, image style, and image sharpness. 🎛️

• The resulting images are whimsical and nuanced representations of the location, based on how the AI perceives it.

• The camera resembles a traditional camera but features a star-shaped dish instead of a lens, inspired by the star-nosed mole. 🎭

• The Paragraphica is best suited for capturing environmental scenes and outputs single square-shaped images.

• It is currently a pet project with limitations, unable to capture faces, objects, videos, panoramas, or landscape/portrait shots.

Would you buy it if it went to market?

Here’s the detailed article from yankodesign. All images are taken from there as well!

Highlight 4: ChatGPT’s newest features!

With OpenAI’s release of ChatGPT, 7 months ago (yes, time flies… and even more when we talk about AI 🤯), and the incredible tools it brought to our lives, OpenAI continues to make it better with some important updates. And because the article is pretty technical, here are the main takeaways each with a brief explanation of what they actually mean😉.

• OpenAI introduced function calling capability in the Chat Completions API 💻:

For the coding geeks: Now developers can describe functions to GPT-4 and GPT-3.5 Turbo, and the models will intelligently output the code with the necessary arguments to call those functions. This means a more reliable way to connect GPT's capabilities with your favorite external tools and APIs. This opens up a universe of possibilities like creating chatbots that utilize external tools in your own app ♾️!

• Updated versions of GPT-4 and GPT-3.5 Turbo were released, with a 16k context version of GPT-3.5 Turbo introduced alongside the standard 4k version 🤖:

The updated gpt-4-32k-0613 offers extended context length for better comprehension of larger texts. More users will be invited to try GPT-4, with the aim to remove the waitlist entirely. The gpt-3.5-turbo-16k provides 4 times the context length compared to the standard version, supporting around 20 pages of text as a prompt in a single request at a slightly higher price 💥.

• OpenAI reduced the cost of their state-of-the-art embeddings model by 75%, and the cost of input tokens for GPT-3.5 Turbo was reduced by 25% 💰:

The embedding model, the most popular, now costs only $0.0001 per 1K tokens after a 75% cost reduction 🤑. The get-3.5-turbo chat model, which powers ChatGPT for millions, has a 25% reduction in the cost of input tokens. You can now use it for just $0.0015 per 1K input tokens and $0.002 per 1K output tokens, giving you approximately 700 pages per dollar. At the same time, gpt-3.5-turbo-16k is priced at $0.003 per 1K input tokens and $0.004 per 1K output tokens 😋.

And that is all for now, folks! here is the official announcement 📢. Great news from the OpenAI team, which always has amazing features to light those of us that use it daily. If these new updates are cool, I can’t wait to see what OpenAI is cooking for the future! 🤓

A.I. Showcase: Flair.ai

If you have a product company or want to make one, you know that taking professional product images can be a difficult and expensive task. Well… here’s flair.ai that will allow you to create professional product images in seconds! ⏳Here’s the bits of how to use it:

1. Go to https://flair.ai/

2. Click on CREATE WITH FLAIR and sign in with your email 📧.

3. Click on Create New Project

4. Click on Upload Product Photo - you can remove the background directly in flair.ai itself 😉!

5. Describe the product - I’ve typed "A non-alcoholic beer bottle" for a Coolsberg bottle:

6. Click on Generate:

7. Click on Editor:

8. Type in the information you’d like by describing your ideal product image - I’ve typed in a marble base with an ocean background. Click Generate:

9. You can use a magic eraser to remove any unwanted parts ❌. I don’t want this cloud here, for example:

10. Click Generate and that’s it! Here’s our result!

AI Art of the Week

In today’s newsletter, we want to feature some amazing work from FEEBEE300 (Discord FEEBEE300).

I recently went on a trip to London and my girlfriend gifted me a really cool jacket but FEEBEE300 designed something even cooler. Check it out!

Have a great weekend!

P.S. This was the jacket I got:

-AI Bits