You have to see this deep fake demo; sharing our indie picks for AI-powered search experiences
Welcome to another edition of the best damn newsletter in AI. Here we’ll break down AI topics that matter, open your mind to use cases, and keep you ahead of the curve.
Our #1 goal is to be useful. So please shoot us an email 📩 if you have questions or feedback, and especially if you implement something we share!
Here's what we're covering today:
You’ll be shocked by how far deep fakes are
Less-popular-but-cool AI search experiences
Let’s get to it! 👇
Would you use an Artificial… You?
What if you could clone yourself digitally, right down to the way you speak? Deep Fakes are making this a reality, and if you haven’t started taking notice, you should.
If you’re not familiar - Deep Fakes are AI-driven manipulations of video or audio content.
They create hyper-realistic representations of people saying or doing things they never actually said or did. By leveraging advanced algorithms and techniques, existing videos or audio files are altered to depict entirely different scenarios. And are often startlingly realistic.
There are many companies building version of Deep Fakes products - dubbing your lips so it looks like you’re saying something you’re not or creating entire clones of you.
Most deep fake products haven’t been very good - until now.
Yesterday, a company called HeyGen debuted their next model. And it is shockingly realistic.
The level of detail of the avatar, precision in the movements. We’ve been thinking about it all day until we wrote this. It’s crazy. (Here’s the link again… really check this out)
So you might be thinking… ok sweet, AI-me coming to all social media platforms!
The timing here is almost too good to be coincidental as TikTok has just started rolling out an “AI-generated content” flag after adding terms that ban deep fakes of different types across the platform. Instagram and others claim to be following suit.
Something interesting is that it’s not clear whether AI-generated content will actually be that bad. Hey, if videos online look like the HeyGen video, that’s higher quality than 95% of what’s out there.
What does a future of social media platforms that are a mix of AI-generated and human content look like? Will we actually like the AI-generated stuff better? Will AI-generated content lower the barrier for more stories to be told? Or steepen the learning curve?
We have our own opinions, but as quality is progressing fast - now might be the time to start forming yours.
Less-popular-but-cool AI search experiences
We covered Search as a major AI use case in our last newsletter. While powerful, ChatGPT is not actually connected to the internet (until they turn the Bing search feature back on at least).
So while you’re sitting idle, you could use Bard or Bing, But we wanted to share 3 less popular but cool search experiences we’ve enjoyed:
Perplexity.ai grabs #1 on the list. It’s fast, sleek and cites its sources - which we love. They also have a Copilot setting which turns on their generative UI mode (aka the search box can ask you follow up questions).
Metaphor grabs #2, and while it’s a bit nerdier - it’s got some serious legs. Metaphor has designed a new way of finding results, which in our experience leads to more varied results and thus a more comprehensive answer to deeper or more complex questions. The search experience is pretty basic (and it’s really built for developers to use in their tools) - but cool and worth checking out.
Andi is our #3, which we like for its super fast summarization tool. Search for something, and it’ll show you a set of results. Which with one click, you can generate a custom summary.
For your reading list 📚
Is “Big AI” still pushing ahead…?
Startups and smaller companies push to be part of the White House-level conversations around AI, saying that “Big AI” like Microsoft, Google and OpenAI shouldn’t be the only ones contributing to the conversation
Meanwhile “Big AI” is trying to be more transparent. OpenAI published their GPT Bot, a web crawler that’s specifically gathering data which could be used to train future versions of their AI models. And they’ve created a way for you to opt out.
And if you’re really nerdy…
Nvidia announced plans to release a new chip, the GH200, which will be designed to handle generative AI workflows even more efficiently
We'll see you again on Tuesday. Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected].
🪄 The AI Exchange Team