Good Monday Morning!
Today, we’re learning about Japan’s upcoming AI model, and crafting original audio from text.
Grab your coffee and let’s get started…
Today’s Menu 🥡
- Japan Is Building Its Own ‘ChatGPT’: But Why?
- Text-To-Music Models Are Here: Stable Audio Launch
- Generate Prompts To Make Music: Audio Revolution
Why Japan Is Building Its Own Version Of ChatGPT 🇯🇵
Fugaku, the Japanese supercomputer in charge of digesting Japanese texts.
According to Keisuke Sakaguchi, an expert from Tohoku University, while ChatGPT excels in English, its Japanese performance is hindered by complex alphabets, limited data, and other nuances.
The underlying challenge?
English revolves around 26 letters. In contrast, Japanese boasts two sets of 48 characters, coupled with over 2,000 regularly used kanji characters. And ChatGPT sometimes… fumbles.
Big players like NEC, Fujitsu, and SoftBank are pouring massive resources, amounting to hundreds of millions, into developing AI systems tailored for the Japanese language, not just mere translations of the English counterpart.
Because it’s not just about language. It’s about culture.
When asked to draft a job application in Japanese, ChatGPT might miss out on those intricate expressions of politeness so inherent to the Japanese ethos. Such oversights highlight the gaps in ChatGPT’s understanding of cultural subtleties.
Using Fugaku, one of the world’s fastest supercomputers, a significant project is underway. Aimed at launching next year, this LLM, backed by luminaries like the Tokyo Institute of Technology, promises to be a game-changer, sharing its open-source code with all.
Japan’s journey in the AI realm is just getting started, and the world watches with bated breath.
A Banksy Got Everyday Investors 32% Returns? 👀
Mm-hmm, sure. So, what’s the catch?
We know it may sound too good to be true. But thousands of investors are already smiling all the way to the bank, thanks to the fine-art investing platform: Masterworks.
These results aren’t cherry-picking. This is the whole bushel. Masterworks has built a track record of 13 exits, including net returns of +10.4%, +27.3%, and +35.0%, even while financial markets plummeted.
But art? Really? Okay, skeptics, here are the numbers. Contemporary art prices:
- Outpaced the S&P 500 by 131% over the last 26 years
- Have the lowest correlation to equities of any asset class
Got your attention yet? The Daily Bite readers can skip the waitlist👇
See important disclosures at masterworks.com/cd
Turn Your Text Prompts Into The Audio You Want 🎧
Screenshot of Stable Audio Dashboard
Stability AI, the company behind Stable Diffusion, introduced new audio generation tech last week converts text prompts to original audio.
Yup, text-to-music is here.
Unlike models that mimic specific artist styles, Stable Audio doesn’t replicate iconic bands – it gives you creative freedom rather than boxing you in.
How does the technology work?
Traditional audio tracks generation used ‘symbolic generation’ methods, primarily involving MIDI files.
By interacting directly with raw audio samples, Stable Audio allows for the creation of entirely new music, moving away from the repetitive notes of MIDI.
Its model is trained on over 800,000 licensed music pieces from the AudioSparks library, ensuring a comprehensive dataset and high performance.
How much does it cost? .
- Free version / 20 generations per month of up to 45 second tracks.
- Pro plan / $12/mo, 500 generations per month up to 90 second tracks.
- Enterprise plan / Unspecific amount & generations.
You can hear a sample track in this tweet:
Stable Audio: Turn Your Text Prompts Into Music 🎶
Stability AI, the company behind Stable Diffusion, introduced new audio generation tech last week converts text prompts to original audio.
Yup, text-to-music is here.
Unlike models that mimic specific artist styles, Stable… pic.twitter.com/FCZyCISaL7
— Snack Prompt 🍭 (@SnackPrompt) September 18, 2023
Make Life Simple With AI
This prompt will generate quality text-to-music prompts from your suggested inputs.
You’ve probably used ChatGPT and Midjourney more than you’d like to admit, and the introduction of text-to-audio likely won’t be any different.
Turning every imaginative thought into tangible sound.
From cars honking during traffic to a Halo-themed guitar solo, you’ll be able to generate it all.
However, learning the proper audio terminology won’t come overnight, so we’ve made a text-to-music prompt generator just for you:
Click the prompt above, input the game, movie, or band you’d like to mimic, and ChatGPT will provide you with a prompt to get the song done.
Try it out: Text-To-Music Prompt Generator
Today’s Fast Snacks 🥡
The World’s First AI CEO Humanoid
Mika, Dictador’s AI CEO, made a powerful appearance at the SALZ-21-Home of Innovation conference…
Defense AI Startup Helsing Raises $223M
Swedish heavy industry and defense group Saab is joining as a strategic investor, deepening what was previously an existing partnership…
Advertise With Snack Prompt
Our newsletter is read by thousands of tech professionals, investors, engineers, and business owners around the world. Learn more or book your spot today…
Share on Pinterest
