MARCH 22ND, 2023

TALES OF TOMORROW

ISSUE #4

Hey there!

Issue #4 comes a little late, sorry for that.

Nevertheless, we're packed with news and excited to share some really amazing advances in the field of AI-powered creative storytelling. 🚀

Let's dive in!

🗞️ Latest Updates

Google, Microsoft, Adobe, LinkedIn, Stripe join the AI bandwagon

After Google announced to rollout generative AI tools within Google Workspace (starting with Google Docs and Mail) over the coming months, Microsoft added 365 Copilot, a generative AI extension for Word, Excel and PowerPoint, to the fast-paced AI tech race.

On the visual side of things, Adobe introduced Firefly, a full-fledged AI-powered illustration suite bringing AI image, 3D & font generation to the center stage.​

LinkedIn and Stripe will also integrate AI text generation tools into their platforms: the former with GPT-4 supporting profile creation and GPT-3.5 enhancing job descriptions, the latter with an AI-powered developer tool.

​

Giddyup! ALPACA is on the rise!

While Big Tech is trying to get as much of the pie as possible, the AI open-source community is also having its moment of glory: ALPACA, a 100% open-source large language model, is said to deliver results almost as good as GPT-3. And it's completely free, transparent, and open to all. Here's an article that shows you how to use it on your devices: Meet ALPACA!​

​

Will AI take your job?

It has been said over and over again: AI will impact every aspect of our lives. A newly published paper now looks at the impact of GPT-like AI tools on the job market. It is still an early assessment, but creative professionals should really take a close look: GPT-4 and the future of work.

🔦 Project Spotlight

Open-source text-2-video on Modelscope

While this model is still in its early stages and far from the realistic AI imagery we’re used to seeing these days, it’s a first step toward AI text-to-video creation. And it’s open-source, which will likely catalyze the development of more powerful models in the near future.Want to try it yourself? We got you covered: "How to setup & run Modeslcope's text-2-video in under 10 minutes" (old version)

EDIT: There's already a new version with better quality! 🤯 (the tutorial above will automatically use the latest one)

AK @_akhaliq New modelscope text to video model is out, better quality, trained for a month longer (old model on left, new model on right) @Gradio demo: huggingface.co/spaces/damo-vi… model: huggingface.co/damo-vilab/mod… reddit thread: reddit.com/r/StableDiffus… "An astronaut riding a horse." pic.twitter.com/uEUAPJWuWS March 21st 2023 63 Retweets 316 Likes

​

Text-to-Image-to-Music to Video

A neat combo of Stable Diffusion, Riffusion, Mubert AI and CLIP! Enter a text prompt and create music, album art, and a "music video" all at once. Check it our here.
​

🔬 Paper corner

New Framework Challenging Diffusion Models

It happened almost overnight: diffusion models became the gold standard of AI image generation with the advent of Midjourney and DALL-E. Until then, so-called GAN models (short for Generative Adversarial Network) were the way to go for AI image generation, but diffusion models were too powerful and took over.

Now GANs are back. Read more about GigaGAN outperforming diffusion models here.

What else?

​Runway's Gen2 has been announced and Directed Diffusion lets you place AI-generated images inside other images. Meanwhile, the government of South Korea is about to create a course to train prompt engineers. 🎉

GPT-4 poetry

We hope you enjoyed reading Tales Of Tomorrow and found some inspiration! 😊 If so, please share "Tales Of Tomorrow" on Twitter.

Also, if you want your project featured or if there is a topic you would like us to cover, please don't hesitate to reply to this email - don't be shy, we would love to hear your feedback and do our best to include your wishes in issue #5!

Until then, keep creating! ❤️

Peace.

Tristan

113 Cherry St #92768, Seattle, WA 98104-2205
​Unsubscribe · Preferences​