🎵 AI Tunes are legit

PLUS: TikTok lets creators label AI-generated content more easily

Author

Haroon Choudery
August 11, 2023

New here? Welcome to NAB, and make sure to subscribe for more!

Happy Friday, fellow humans. 👋

Mood: Up

Stress: Down

Outlook: Bright

Horizons: Expanded by AI insights

Let’s dive into it…

Have anything you’d like to share with over 40k subscribers? 

🧵 In today's edition:

  • 🧠 Claude’s Instant IQ Boost

  • 🎵 AI Tunes are legit

  • ✍️ Hands off the Media

  • 🤑 AI Fundraising News

🤖 Top AI News

🧠 Claude’s Instant IQ Boost

Anthropic has supercharged its text-generating AI, Claude Instant, with the release of Claude Instant 1.2 (CI 1.2).

CI 1.2 incorporates the power of Claude 2, showing significant improvements in math, coding, reasoning, and safety compared to previous versions.

  • In tests, CI 1.2 scored nearly 60% on coding challenges, up from 53% for the prior release.

  • It also aced 86% of math questions versus 81% previously.

More updates:

  • The bot is less likely to hallucinate, make mistakes, or be manipulated into generating harmful content.

  • CI 1.2 features a 100k token window, the same as Claude 2.

  • The bot also generates longer, better-structured responses and follows formatting instructions more closely.

  • It's better at extracting quotes and handling multiple languages.

Overall: Anthropic ultimately aims to create an advanced "self-teaching" algorithm that could power virtual assistants for email, research, art, and more.

Claude Instant has started to provide glimpses of this, competing with basic AI services from OpenAI, Cohere, and Bard.

Excited to see where they go next.

Read more: TechCrunch

🎵 AI Tunes are legit

I know April feels like it was three years ago, but who remembers that Drake x The Weeknd AI song? Or the time when Grimes was willing to split royalties with anyone who created an AI song with her voice?

Those were pivotal moments in helping us determine the future of the music industry in this AI era.

As it turns out, the future is now here. Google is cooking up an AI music collab with Universal Music Group (UMG) to let fans legitimately create tracks using artists’ voices.

According to the Financial Times:

  • UMG artists could opt-in to work with the AI tool

  • A licensing deal would pay artists for using their vocal styles and melodies in AI-generated songs

Why are they doing this? Broadly speaking, it conveniently resolves AI copyright issues that created unease in the industry earlier this year. (read: UMG calls AI music ‘fraudulent’)

But there are bigger questions to answer:

  1. Will musicians embrace AI mimicking their craft?

  2. Does licensing artists’ voices to create music cheapen their art?

  3. Should tech and labels decide how creators’ styles are used?

Overall: A deal like this would solve the legal troubles. But ethical quandaries persist around reputational risk and AI creativity versus human expression.

Similar issues underlie current disputes, like the SAG-AFTRA strike against using AI actors and novelist Jane Friedman finding AI books illegally under her name.

What are your thoughts on this AI music situation?

Login or Subscribe to participate in polls.

Read more: Mashable

✍️ Hands off the Media

Source: The Atlantic

Major media groups like AP, AFP, Getty Images, Gannett, and authors’ associations have co-signed an open letter calling for new rules to protect their data as AI models feast on their content.

Their argument is simple: Models like ChatGPT and Google’s Genesis absorb media material without payment or consent from creators. This “undermines core business models” that fund quality journalism through ads, licensing, and subscriptions.

To help resolve this, publishers are requesting

  • Transparency into AI training sets

  • Approval from copyright holders

  • Negotiations with AI firms

  • Labels on AI content

  • Elimination of model bias/misinformation.

Don't get them wrong - they believe AI can offer great benefits, but they want to understand what the AI is being trained to do.

Overall: This issue isn’t just with news media. The Senate is probing AI copyright issues with art-generating AI. Even comedians and authors are suing OpenAI.

As these conversations progress, one thing is for certain: Big AI Tech has to answer tough questions about AI, copyright, and consent.

And for the media’s survival, the stakes couldn’t be higher.

Read more: The Verge

🤑 AI Fundraising News

Virtualitics Closes $37M Series C Funding Round to provide AI capabilities that enable advanced analytics for organizations, automatically discovering hidden patterns in complex data and delivering immersive experiences that guide informed decisions.

Middleware Raises $6.5M in Seed Funding to build an advanced AI advisor based on generative AI to improve the cloud observability stack further.

🗞️ AI Quick-Bytes

What else is going on?

  1. CHART: AI stocks enter correction territory

  2. ChatGPT expands its ‘custom instructions' feature to free users

  3. Jeli is bringing generative AI to incident report analysis

  4. The AI boom might be leading to a spike in GPU prices

  5. TikTok lets creators label AI-generated content more easily

  6. Report slams generative AI tools for helping users create harmful eating disorder content

  7. Supermarket AI meal planner app suggests recipes that would make chlorine gas

🐦️ Tweet of the day

🥳 Unreal. Congrats to the whole PlayHT team!

What did you think of today's newsletter?

Login or Subscribe to participate in polls.

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 :)

Reply

or to participate.