🎶 AI lacks soul
PLUS: Google helps non-profits adopt AI
Sponsored by
New here? Welcome to NAB, and make sure to subscribe to stay informed on the latest developments in AI!
Happy Thursday, fellow humans 👋
AI is creating music that isn’t necessarily all that great, Apple’s Siri could be in for a serious upgrade, and Google plans to invest over $20M into non-profits to help them adopt GenAI…
Let’s dive into it!
Have anything you’d like to share with over 40k AI enthusiasts?
🧵 In today's edition:
🎶 AI lacks soul
🍎 Apple enters the AI ReALM
🔨 Google helps non-profits adopt AI
🤑 AI Fundraising News
🗞️ AI Quick-Bytes
🐦 Tweet Post of the Day
🎶 AI lacks soul
Stability AI has launched Stable Audio 2.0, a feature that allows users to generate AI-composed "songs" up to three minutes long, a significant increase from the previous 90-second limit.
Here are the details:
The updated model allows you to import your own audio samples to influence the generated output.
You can also adjust parameters like how closely the AI follows text prompts and the extent of transformation for uploaded audio.
There's even the option to overlay jazzy sound effects.
However, there is one problem: the music still noticeably lacks the "soul" factor.
Other companies like Google and Meta are also wrestling with this "listener repulsion" problem as they develop their own audio AI models.
The root of the issue? AI models are trained on databases of pre-existing audio, not genuine musical creativity.
So, while AI can compose longer "songs," infusing them with authentic artistry and emotion remains something only a human can do.
For now, the beeps and glitchy sounds resembling whale songs will remain firmly placed in the "fascinating experiment" category.
Read more: The Verge
🍎 Apple enters the AI ReALM
Source: Try AI Pen
Apple researchers say they've developed a powerful on-device AI model that can comprehensively understand and reference items on your screen, in conversations, and even from background processes.
Plus, Apple claims it "substantially outperforms" OpenAI's flagship GPT-4.
Called ReALM (Reference Resolution As Language Modeling - yep, that’s a mouthful), the system can contextualize on-screen apps, ongoing conversations, and ambient processes like alarms or music playing, allowing it to be a smarter, more useful assistant.
To test its abilities, Apple pitted ReALM against GPT-3.5 and GPT-4 in benchmarks, challenging the AI models to identify relevant entities from provided data.
The result?
While GPT-4 got a performance bump from seeing screenshots, Apple's smallest ReALM model matched its scores for on-screen tasks. The larger ReALM versions blew past GPT-4 entirely.
Apple also claims ReALM tops GPT-4 on "domain-specific user utterances," making it an "ideal choice" as a practical, high-performance assistant that runs locally on-device.
Read more: Toms Guide
🔨 Google helps non-profits adopt AI
Google wants to ensure that non-profits are not left behind in this rapidly growing AI world.
To help this cause, Google announced a $20M+ investment to accelerate GenAI adoption across the social impact sector.
Here’s the TL;DR:
Google’s new "Accelerator: Generative AI" program will provide funding, training, and hands-on mentorship to 21 non-profits over six months.
The chosen organizations span various industries like healthcare, climate, education, and more - all exploring cutting-edge AI use cases.
One example is the World Bank’s development of an AI assistant to help policymakers quickly extract insights from dense data.
For context, the program builds on Google's recent AI Opportunity Agenda to upskill a global "AI Corps" workforce through training pathways.
For the non-profits, Google's accelerator could prove transformative:
They'll receive AI implementation guidance, technical mentorship from Google experts, and vital resources to scale their AI tools across communities in need worldwide.
The program's unveiling signals Google's determination to be a vanguard for "AI for Good" as GenAI explodes.
By empowering mission-driven organizations, the accelerator could be responsible for a tidal wave of positive, equitable AI disruption across crucial sectors.
Read more: Mashable
🤑 AI Fundraising News
NobleAI raises $10M in Series A extension funding to help accelerate product development and reduce costs through science-based AI models.
Brandtech Group raises $115 million in funding to continue providing its brands, Oliver, Jellyfish, and Gravity Road, with AI capabilities.
Need to schedule a meeting with 7+ people?
Need to take meeting notes while staying engaged?
Need to manage your deliverables and deadlines?
Meet Xena by Xembly, your company’s new AI executive assistant that does the work for you — giving you 400 hours extra a year.
Boost your team’s productivity with Xembly today (free for a limited time only).
🗞️ AI Quick-Bytes
What else is going on?
Women in AI: Kathi Vidal at the USPTO has been working on AI since the early 1990s.
Nvidia might be clouding the funding climate for AI chip startups, but Hailo is still fighting.
Opera allows users to download and use LLMs locally.
Microsoft 365’s Copilot gets a GPT-4 Turbo upgrade and improved image generation.
AI George Carlin case settled as performers demand better protection.
TSMC facilities will resume production overnight after the quake.
Now there’s an AI gas station with robot fry cooks.
Replit launches a new product in the race for AI coding assistants.
🐦 Tweet Post of the Day
A cool tutorial on how to edit DALL-E images directly in ChatGPT 🧑🎨
OpenAI has just added image editing to Dall-E 3
This means you can add or modify anything on the image you generate.
I'll show you where this feature is and how to use it directly in ChatGPT:
— Paul Couvert (@itsPaulAi)
2:34 PM • Apr 3, 2024
What did you think of today's newsletter? |
And that does it for today's issue.
As always, thanks for reading. Have a great day, and see you next time! ✌️
— Haroon: (definitely) Not A Bot and @haroonchoudery on Twitter
P.S.: Check us out on our socials - snacks on us if you also hit follow. :)
Autoblocks | Twitter | LinkedIn | YouTube