My Money Where My Mouth Is

Future Telescope 7

This Lord Echo track is nice to play while reading today’s post:

You may remember from the post last week that I advocated saving $20 every month on GPT-4 and just use Bing AI instead.

Well, turns out, I didn’t follow my own advice!

In fact, I went ahead and paid for another AI service - Midjourney v5. Because check out these images I can make with Midjourney now:

Impressed?

It gets better still:

Yes, even these photorealistic examples were all generated with the help of an AI.

Welcome to the future.

1. Why Now?

I could have paid for API access to GPT 3. I could have bought ChatGPT Plus when it was first unveiled.

I could have paid for Midjourney the day it was unveiled.

But I was able to get by with the basic features without feeling compelled to shell out money. If it wasn’t possible to use a feature on these services, there were competing services like Bluewillow for image generation that always scratched my itch.

But now, it’s different.

Now I have you!

49 subscribers have signed up to read Future Telescope every week. In just about 45 days of writing this newsletter, it feels extremely gratifying to have this substantial audience and a chance to talk to you inside your inbox.

Thus, your presence encouraged me to dive deeper into this world and put my money where my mouth is.

2. Early Impressions

I got my hands on GPT-4 just this week and am still experimenting. First cut impressions are:

1. It is very specific about answering exactly the stuff you’ve asked it about. It doesn’t meander or talk beside the point.

2. It can code. It can also give you step by step explanations of code.

3. It is a good editor.

   LLM AIs as Editors

There are many more things I want to try with GPT-4, but I want your help here. If you are curious about what GPT-4 can or cannot do, please share your suggested GPT-4 prompt in the comments below, and I will share with you the answers I get.

This could be a lot of fun. I’ll look forward to your comments!

Now let’s talk about Midjourney.

Midjourney v5 has legitimately blown my mind.

Just look at these outputs!

The photorealism here is just mind-blowing. Here are some new things that my use of v5 have taught me:

1. The Midjourney team specifically prioritized hand and face generation in its training data for v5. Thus, it is now a goldmine for photoreal images.

2. The prompts for this version are now more natural. In the earlier models, you would write “Boy, holding bucket, photoreal, lush green background”
   Now, you write “A photograph of a boy in a red tshirt holding a yellow plastic bucket, standing in front of lush trees.”
   By the way, here’s the output you get for this prompt in v5:

3. It’s faster when you pay for it.

4. You also get an access to the Midjourney bot, which allows you to send prompts for image generation outside of the rookie channels that Midjourney assigns to free users.

3. Catalogue

The past 10 days have been insane in the world of AI. We have the following new developments to consider now:

1. Google Bard - the GPT competitor that claims to do a lot but apparently doesn’t get up to the mark

   Marques Brownlee @MKBHD

   I’ve been playing with Google Bard for a while today and I never thought I’d say this, but… Bing is way ahead of Google right now (at this specific chat feature)

2. Bing image generation - here we go, Bing gets yet more capable. Also, check out the domain name - bing.com/create - I wonder what other creation tools is Microsoft developing.

3. Adobe Firefly - let’s welcome Adobe to the generative AI world, and boy, what an amazing feature set they are bringing to the table. I have been using Photoshop now for 13+ years, and I can honestly say this is legit the most profound set of upgrades to the Adobe Creative Suite. Just look at those AI brushes! And AI vectors!

   Bilawal Sidhu @bilawalsidhu

   Been hands-on with the beta of Adobe's cutting-edge Generative AI tool, and I'm impressed! 🤯 Here's a taste of the power of #AdobeFirefly 🎇 and what sets it apart in the increasingly crowded world of #AI art. Thread 🧵🎨

   

   

4. Nvidia brought a lot of enterprise capabilities to the table. But one thing that stood out to me was the text-to-3d service they announced with Shutterstock. Impressive.

5. Microsoft brought AI to office. It’s wild.

There’s a lot more that happened in our AI world, and here are the places you can find out more:

GPT-4's killer use-case, Microsoft and Google's AI battle comes to your workspace, plus LLaMAs for everyone

This article mentions the story about the guy who gave ChatGPT $100 and made money using it. Isn’t that crazy?

And that’s the power of this technology to me. Seek ideas from it. Talk about it with the world. Become better at working with it, and see your impact on the world grow.

It’s a beautiful new world, and it’s time to start adopting it. If that means buying a monthly subscription, I’m in!

A beautiful world, imagined by Midjourney v5

That’s it for the seventh draft, see you next week!