😺 Mona Lisa rapping

PLUS: Introducing The Neuron PODCAST!

Welcome, humans.

BIG NEWS!

After delivering 300+ editions of The Neuron newsletter, we’re launching The Neuron podcast!

TLDR: It’s still about AI, but we’ll dive deeper and cover more than we normally do in the newsletter: extra thoughts, info, tips, and juicy insider secrets (ooh!).

And while we dig Lex Fridman and Call Her Daddy, we don’t think you need another marathon 2-3-hour podcast. We’re gonna keep it tight, The Neuron style.

Today’s the first episode. Come listen on Apple Podcasts, Spotify, YouTube, or wherever you get your pods.

To celebrate, we’re giving away a $500 Visa gift card to one supporter of our launch!

Enter here by subscribing to our podcast or by sharing it on social media!

Here’s what you need to know about AI today:

Microsoft demoed an AI that animates images of people to make them speak.
Nearly 50% of AI legislation in state congresses concerns deepfakes.
Groq is a way to use Llama 3 that responds insanely fast.
61% of Neuron readers think the media overhypes AI.

Microsoft developed an AI that animates images into talking figures.

Just when we thought AI couldn’t get any stranger, Microsoft demoed a new AI model dubbed "VASA-1” that can make any static image “talk” and “sing”.

The videos sound creepily real:

Two questions:

Why did this clip sound better than 95% of AI music we’ve heard?
Why is a $3T company like Microsoft spending its time and money making Mona Lisa rap?

At its best, this type of AI video generation tooling could help businesses engage customers in new ways.

Instant videos of anyone saying anything that look/sound real = massive time savings. Imagine using a single image of Apple CEO Tim Cook to create endless marketing videos of him promoting the Apple Vision Pro

This is essentially what companies like HeyGen and Synthesia (specifically with AI avatars) are working on:

At its very worst, however, this type of tech could exacerbate the very valid concern that AI might be used to produce hyper-realistic deepfake content that distorts reality.

It’s a legit concern cause it’s already happening at scale:

AI deepfake videos imitating celebrities like Tom Hanks and Mr. Beast endorsing bogus products have gone viral.
Two A-list Indian actors, Ranveer Singh and Aamir Khan, were victims of AI cloning videos that misrepresented them as endorsing certain political parties.

Coming up: Microsoft acknowledged potential misuse in its announcement, and has “no plans to release an online demo, API, product, additional implementation details, or any related offerings”.

That’s likely a wise decision, particularly since deepfakes are currently the #1 AI-related issue being addressed by regulators.

Nearly half of the 407 AI-related bills this session in state legislatures focus on deepfakes, with 11 states having recently enacted laws prohibiting the use of deepfakes in political advertising.

FROM OUR PARTNERS

Largest price correction in 30 years in this overlooked sector

64,387 investors have started buying into a market that used to be restricted to the billionaire class: postwar and contemporary art. What do they know that you don't?

Until now, prices haven’t declined for 2 years in a row in nearly 30 years
Trends can change, but when prices last “dipped” in 2009 they shot up 132% over the next 5 years
Postwar and Contemporary art prices have outpaced the S&P by 64% (1995-2023)

Now, anyone can easily invest in blue-chip art with Masterworks. Masterworks’ team scours the art market for the best deals, buys them at a discount, and offers shares to members.

Want to get in before the art market comes roaring back? As a trusted partner, The Neuron readers can use this exclusive link to skip the waitlist.

Around the Horn.

Apple acquired a French AI and computer vision startup.
Bank of America noted that there’s pressure on Big Tech to deliver on their AI promises during this week's earnings calls.
AI for Data Journalism: demonstrating what we can do with this stuff right now.
The three major job hotspots for AI are SF, Seattle, and San Jose.

Treats To Try.

*Sana AI is your ultimate AI assistant for work—it helps analyze documents, summarize meetings, automate repetitive manual tasks, and more. Work faster AND smarter—try Sana AI for free today!
Groq makes using Llama 3 insanely fast. See how it measures up to GPT-4, how to instantly refactor and document code, & explore a ‘llama3-researcher’ that digs into any topic.
Anthorpic’s Prompt Library offers dozens of optimized prompts you can use with Claude for business and personal tasks.
Stepify.tech transforms YouTube videos into step-by-step tutorials.
VideoGigaGAN is a new Adobe tool that can upscale a video up to 8x with rich details.

*This is sponsored content. Advertise in The Neuron here.

Tuesday Ticker.

Here are the results from last week’s survey:

H. chose No: “I think they underhype to real potential and awesome use cases & overhype some features which aren't as useful or play of folks' fears, both leading to slower adoption.”
D. chose Yes: “The media is all doom-and-gloom or utopian unicorns. All clickbait. That's what makes the neuron so helpful. Y'all show reality.”
N.L. chose It’s Just Right: “It’s the world we're living in. Cover it all and educate people on the pitfalls.”