Caution: Gen AI rabbit holes ahead

Author

Renga P
December 08, 2022

Hello there (cue Obi-wan intro)

I am Calin and this is the first edition of ‘AI in the middle’ newsletter. Wait, what? Why is AI in the middle? Now that I have your undivided attention, let’s unpack what we have here: Generative AI.

Before you groan ‘Not another Generative AI news snippet’, I am here to calm you down. This is not just any AI-centric newsletter, this is the ONLY newsletter you may need to become a Jed-I (see-what-I-did-there) and take you to the promised generative AI land. Are we seated and comfortable? Let’s begin.

What does ‘AI in the middle’ mean?

I know you are pressed for time, so let me make this quick. I made this newsletter, so you get the best news and insights across the Generative AI landscape in one place, so you don’t have to go scour the cesspool we know as the Internet.

Until now, AI has been used as an innocuous sidecar companion for the work we do, helping us get better at grammar, or write new headlines. But now AI gets a seat at the thought process table, right in the middle, between monotonous computer-driven automation, and human creativity. I have seen this first-hand with generative AI helping us get creatively unstuck with MidJourney, Stable Diffusion, and the latest kid on the block from the OG who started it all, Open AI and its offspring ChatGPT.

Knowing this space is only going to go deeper and wider, I decided to take notes, introspect, and share some of these with you, with you willingly obliging.

And what makes you interesting enough, Calin?

Because I speak nicely. I am kidding, but hear me out (I do speak nicely fr). I am building in the Gen AI space myself. Rather than guarding my turf, I want to be a messenger and put all such Gen AI projects in the spotlight, so that you can have a look at all the wonderful builders, making progress in this new wave.

Hmm, you may be onto something here…

If you are thinking that way, thank you so far. My team and I start this week with putting the spotlight on some of the best written thought pieces on endless AI, from folks I hold high praise for - Mario and Ben Thompson.

Along the way, we also talk about the possibilities of AI in education and community-building, and then wind up new deals and startup announcements that are cooking up a storm. Enough of my millennial babble, let's get to it.

Mario on the rise of ‘endless media’

endless media

We have seen recent developments in gen AI show promise on how media is consumed, with "endless media" allowing consumers to generate their own TV shows, albums, games and other forms of media on demand. While AI can't yet create the next Game of Thrones, it has already demonstrated the ability to write blog posts and conjure images faster than humans.

Mario from the Generalist speculates on how AI could be used to tailor media to consumers' specific preferences, while copyright enforcement could become a major issue. He believes that for the world of endless media to emerge, generative artificial intelligence must succeed in three dimensions: quality, scalability and control.

We’re keeping an eye to see if it can do so without becoming sentient and wiping us all out in a Matrix-style uprising.

Ben Thompson talking about my homework (for the future)

homework meme

Ben from Stratechery is making us think about the future of homework. What happens if everyone starts using the new technology to “generate” their homework?

AI-generated text, such as that from the ChatGPT app, is a probabilistic device that does not have any internal record of right or wrong, but rather a statistical model of what bits of language go together in different contexts. This means that, unlike calculators which have one definitive answer, AI-generated text can offer multiple answers to a question. And often, it gives out the wrong answer.

The potential for AI-generated text to create essays and other homework assignments raises the question of how schools and teachers will adapt to this new technology. Will the “knowledge skill” of the future be the ability to assess AI responses and “edit” them?

Meanwhile, whilst we figure that out, the new wave of excuses are starting to flow: It’s not my fault teacher, I instructed it to do the homework right, but it’s not listening!

Ethan Mollick talks about my Professor (why is everyone in Gen AI obsessed with school!)

The ChatGPT AI-powered chatbot could have a more significant impact on analytical and creative industries than it will on repetitive and dangerous ones, according to Ethan Mollick of the Wharton School of the University of Pennsylvania.

Mollick demonstrated the capabilities of ChatGPT by asking it to create a syllabus for an MBA course, create an assignment and grade a student's work. He also asked them to create a lecture, explain a concept and write in a warm tone. Mollick said the AI's performance was impressive and suggested that it could be used as an intern-like resource to help extend human capabilities.

Mollick added that, while the AI is not yet as good as a human professor at any task, it can do a ‘shocking amount right now’. He urged readers not to be scared of AI, but to consider the opportunities it presents. (This is the cop version of ‘everybody stay calm’...I guess)

So now that we’re automating both homework and teacher’s jobs, everyone can go on Spring Break indefinitely, right? RIGHT?

David Spinks on Gen AI being the neighborhood watch (where do I sign up?)

bad actor

The one thing that surely is left for us humans is building communities. Or is it? David Spinks believes that communities will be built by robots.

In his experiment, he goes through a scenario where he uses ChatGPT to invite new members to a community, drive engagement, manage bad actors and deal with moderation.

David goes on to say we’ll all be “cyborg community builders” in 3-5 years and the AI will help us drive data and insights from a community. Finding the right expert in a community, leveraging and sharing knowledge and matching community members can all be powered by AI in the future. Scaling communities is hard, but AI can make all of this easier.

Gen AI Deals that make your eyes (and mouth) water

What caught our eye?Cristóbal Valenzuela’s (Runway’s CEO) vision: “Picture you’re watching a movie and you are the lead actor in the movie: your voice, your body, your face.” Imagine the best you (no, not the I-just-had-a-face-massage-you, the Lensa-AI-you) in a movie of your choice?

What caught our eye?Founder Ishan Sharma, says many clients ‘have seen up to 70% more qualified conversions after they started using SellScale.’

Is this the end of pesky, unpersonalized sales emails? If you’re not the best person to speak to about this, please let us know.

New developments to sound suave at the coffee cooler

Things to learn when you need a raise…

The best resources we came across that will help you become a better founder, builder, or investor in Gen AI.

And things to know when you are feeling low

  • AI prompts don’t always work as intended, here’s a list of failur..err..experiments

That's a wrap! See you next time!