🛫 Top Gun: AI
PLUS: Samsung is building generative AI, but not for you
New here? Welcome to NAB, and make sure to subscribe for more!
Happy Tuesday, fellow humans. 👋
Congrats to the Denver Nuggets for winning their first championship in franchise history! Feel bad for Jimmy and the Heat, but they will be back 💪
Let’s dive into some AI news…
Have anything you’d like to share with over 40k subscribers?
🧵 In today's edition:
🎼 Meta’s AI-powered music generator
🛫 Top Gun: AI
🤝 OpenAI will share AI models with the UK government
🤑 AI Fundraising News
🤖 Top AI News
🎼 Meta’s AI-powered music generator
AI-generated songs have flooded the internet recently, and Meta is joining in on the creative fun.
Meta released its own music generator, MusicGen, using AI to turn text descriptions into music.
The generator is trained on a whopping 20k hours of music, including licensed and instrument tracks from Shutterstock and Pond5.
The best part? Meta is sticking to their promise and has made the project open source - you can find it on HuggingFace.
It’s also important to note that MusicGen only generates about 12 seconds of audio - clearly not enough to replace human musicians. Labels - breathe easy.
Despite the increase in AI music generation tools, streaming platforms have already started to ban AI music. Companies like Deezer are taking it further, building "cutting-edge tools" to detect AI-generated content on their streaming platform.
Why is this important: It’s clear that generative music is improving with each iteration, but labels and streaming services are facing new challenges every day.
While some musicians are embracing the technology, others are calling for regulation to protect their livelihoods and prevent fraudulent activity.
With companies using licensed tracks to train their models, will the industry start accepting AI-generated music?
Read more: TechCrunch
🛫 Top Gun: AI
I didn’t believe that anyone other than Tom Cruise could complete the mission in TopGun: Maverick.
I was wrong.
{Spoiler Alert}
The mission in Top Gun involved flying close to a rocky canyon bed and maneuvering through the rock formations as quickly as possible to avoid being spotted by the radar as they tried to reach their destination.
Until recently, autonomous aircrafts had issues doing this since their main objective was to reach their target in the quickest way they knew how - a straight line with maybe a few turns here and there.
But now, MIT researchers have developed a machine-learning technique that uses deep reinforcement learning and mathematical optimization to address the 'stabilize-avoid problem' in autonomous robots, resulting in a tenfold increase in stability and improvements in safety and stability.
The new technique has been effectively tested on a simulated jet aircraft. In the future, it could be a starting point for designing controllers for highly dynamic robots like autonomous delivery drones.
Read more: SciTech Daily
🤝 OpenAI will share AI models with the UK government
Yes, yes, more regulation talk - but this seems important.
UK has seemingly been at the forefront of AI regulation.
UK leaders have proactively called to conduct an initial assessment of AI model accountability, safety, transparency, and other ethical concerns.
They committed £100 million (about $125.5 million) to create a Foundation Model Taskforce to develop "sovereign" AI.
Now: At the launch of London Tech Week, PM Rishi Sunak stated that Google DeepMind, OpenAI, and Anthropic have promised to give the UK government "early or priority access" to their AI models for research and safety purposes.
The reason? The move aims to improve inspections of AI models and help the government identify potential risks and opportunities before the models are open to the public.
That’s big time.
Read more: Engadget
🤑 AI Fundraising News
Gensyn AI Secures $43M in Series A Funding for Decentralized Machine
Percepto Raises $67M in Series C Funding to provide autonomous inspection and monitoring solutions using autonomous drones, robots, and AI.
🗞️ Byte size: AI article summaries
Disclaimer: AI is (partially) used to summarize these articles.
Samsung is building a generative AI, and you won’t be able to use it [Sam Mobile] - Samsung is reportedly developing a generative AI service similar to ChatGPT and other widely available tools, but it will only be for the company's internal use. Employees will be able to leverage the AI's power for projects and input sensitive company data into its database without concern for leaks. The tool will run on closed circuits and be available only to Samsung semiconductor and software developers. (Read more)
OpenAI’s CEO Calls on China to Help Shape AI Safety Guidelines [Yahoo Finance] - OpenAI CEO Sam Altman believes that China's expertise in AI talent should be used to develop global standard regulations that ensure AI safety. OpenAI brought the AI initiative ChatGPT to China in 2020, but currently, it is not available due to China's rules around data and censorship. However, China is highly invested in AI, which will play a strategic role in defining the deepening tech rivalry between China and the US. (Read more)
A special mention: If you have some spare time, I highly recommend you check out AI Whiteboard by Emad Hasan - a great read for business leaders in AI.
🐦️ Tweet of the day
🤖 Applicable AI!
Artificial Intelligence.
A technology everybody talks about.
But what are the 7 real world applications and what can AI do today?
Let's find out:
— Tim O'Shea (@timosheahq)
4:57 PM • Jun 12, 2023
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 Robot and @haroonchoudery on Twitter
P.S.: Check us out on our socials - snacks on us if you also hit follow :)