- Bot Eat Brain
- Posts
- Microsoft partners with Suno to bring song creation to Copilot
Microsoft partners with Suno to bring song creation to Copilot
PLUS: Lawyer with AI > lawyer with brain
TOGETHER WITH
Good morning, human brains. Welcome back to your daily munch of AI news.
Here’s what’s on the menu today:
Burn your million dollar recording studio to the ground 🎸 🔥
Microsoft and Suno bring text-to-complete song creation to Copilot.
Does VideoPoet create poems? Nope. Not at all. 📽️ ✨
Google unveiled its new video generation model, VideoPoet.
Normal lawyers suck. Get one that uses AI 🤖 👩⚖️
LexisNexis is giving US law students access to AI tools.
MAIN COURSE
Create a banger in five seconds 🎸 🔥
A couple of weeks ago, we reported on Google’s Instrument Playground. It’s an AI that generates 20-second music tracks using text prompts.
That was… eh. 🤷♀️
This is… better. 🤘😤
I don’t like new things. What is Suno?
It’s a startup based in Cambridge, MA. It’s a team of musicians and AI experts with experience in major tech companies. Their goal is to make music creation widely accessible.
I have the music skills of an octopus.
No matter, friend. You can create complete songs from a single sentence.
Is it any good?
It’s seriously impressive and very easy to use.
No cap. 🧢
It allows you to create full audio compositions, lyrics, instrumentals, voices, and more using text prompts.
The best part? It’s believable, and the songs are really catchy.
Really?
Bruh.
We typed in: “A neo soul song about a robot that eats brains.”
Here are some of the lyrics:
“An appetite beyond the realm of reason
A robot with a taste for human cognition (ooh-yeah)”
So the lyrics are a little cheesy, but the beat seriously slaps.
It be spittin’. 🤖 🎵
Yes, it created the album art too.
How do I use it?
Prior to this partnership, creating songs with Suno’s tech required using Discord.
Now, you can enable the Suno extension in Microsoft Copilot. After that, just put in a text prompt and watch it do its thing.
Isn’t Microsoft Copilot expensive?
Yes. It’s cheaper just to use it through Discord.
Plus, the first 50 credits are free.
From our partners
Learn No-Code & AI skills in 100 days
Learn No-Code and AI skills by committing to the 100DaysOfNoCode or 100DaysofAI challenges.
Receive a 30-minute bite-sized lesson straight to your inbox every day.
It’s free, fun, and effective!
Both challenges start Jan 1st, 2024 ⚡️
SIDE SALAD
Google’s short, sweet video model 📽️ ✨
On Tuesday, we reported on Google DeepMind’s FunSearch. It’s an LLM that solved a notorious scientific problem, a first for LLMs.
Later that day, Google unveiled VideoPoet. It’s a powerful, context-aware video generation model.
After the Gemini let down, I’ll pass.
Don’t sleep on this. 😴
VideoPoet is more versatile than a lot of video generators out there.
It animates static images, edits videos for inpainting and outpainting, generates audio from video, uses different inputs like text, images, videos, and more.
I’m almost impressed.
It offers interactive video editing, which allows you to guide the model in modifying video motions and actions.
You can even tell VideoPoet to respond to camera motion commands.
Like this:
That’s pretty cool.
Right? It can also generate relevant audio for videos without any instruction from you.
This is where the whole “context-aware” thing comes into play.
There’s always a “but…”
By default, VideoPoet only outputs 2-second videos.
I knew it.
If you really wanted to make a longer video, you could give VideoPoet a 1-second clip and have it output videos indefinitely.
Since it understands what it’s looking at, the clips should cut together seamlessly.
We haven’t tested it out.
Alright, I’ll give it a chance.
You can’t yet.
A LITTLE SOMETHING EXTRA
Lawyers + AI > lawyers + books 🤖 👩⚖️
Huh?
LexisNexis is an analytics company that specializes in legal/business data. It developed a generative AI tool called Lexis+ AI.
Lexis+ AI offers conversational searching, document drafting, summarization, analysis, and more.
Why are they doing this?
According to a LexisNexis survey, 78% of law professors plan to teach generative AI to students.
Have any law schools agreed to this?
Oh, yeah.
Students from Berkeley School of Law, Wake Forest School of Law, and many more praise Lexis+ AI.
What’s so great about it?
Students claim it interface is easy to use, it saves time, and it retains context better than other AI tools.
RECOMMENDED READING
MEMES FOR DESSERT
YOUR DAILY MUNCH
Tools
Pawtrait Studio — generate over-the-top images of your beloved pets.
Creatify — enter a URL and it automatically generates product video ads.
Dewstack — Create and host AI docs for FAQs, user manuals, and more.
DeepMake — AI video editing software with upscaling, face swapping, and more.
Think Pieces
Nature Journal picks ChatGPT as one of its 2023 Scientists of the Year. Ilya Sutskever also made the list.
Is this better than Stable Diffusion XL? This new text-to-image model, Playground v2, outperforms it 2.5 times in user satisfaction.
Here’s why Anthropic changed its legal protection terms. It will protect customers from copyright claims started on the 1st day of the new year.
Startup News
Rite Aid has been banned from using AI facial recognition tech. They agreed to halt for five years after falsely identifying innocent customers.
Midjourney is pre-releasing Version 6. They invited paying members to give feedback and rate images created by its newest version.
Research
A Paper on Gemini — Gemini Pro falls behind the original GPT 3.5 Turbo in most areas. It performs well in translation tasks in some languages.
gAvatar — NVIDIA’s paper on how to leverage Gaussian splatting for text-to-3D Avatar generations.
Amphion — Audio, music, and speech generation toolkit that’s open-source.
TWEET OF THE DAY
Ethan Mollick, a renowned Wharton University Professor, uses Suno to instantly create a song about cheese.
Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.
AI ART-SHOW
Until next time 🤖😋🧠
What'd you think of today's newsletter? |