Will AI be too expensive?

Why the cost of AI won't be an issue; Create full videos entirely with AI

Author

The AI Exchange
March 09, 2023

Welcome to another edition of what we’re determined to make 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: 

  • An argument for why the cost of AI won’t be an issue

  • Create visually stunning videos, completely with AI

  • Salesforce joins the AI arms race after some exciting announcements

... and if someone forwarded this email to you, thank them 😉, and subscribe here!

Let’s get to it! 👇

TODAY'S PERSPECTIVE

The argument for why the cost of AI won’t be an issue

For most of us.

Last week, Open AI dropped the cost of their APIs - which give any business or product the ability to start using their models - to 1/10th of the previous price.

What most people don’t know is that this isn’t the first major price drop we’ve seen since Open AI started commercializing its AI models in 2020.

The initial models were priced around $0.04 to $0.08 per 1000 tokens. Then the price dropped to $0.02. Now, it’s $0.002. That’s a 97% price drop in a little over 2 years.

Ok woah.

But how is that even possible?

The “cost” of creating AI models comes from 3 main things:

  1. The initial training cost of the model, which is mostly an upfront cost (with some exceptions)

  2. Then, the cost of using the model (also known as inference and happens each time you use ChatGPT)

  3. Then, the cost of running the company that provides it (aka all the smart machine learning engineers, servers, office space, etc)

#1 and #3 are relatively fixed costs. Meaning that as usage of an AI model grows, these costs do not. And they happen to be the most expensive parts of running this type of company.

So as more people use Open AI’s models, they are able to use that revenue to recoup the initial training cost, cover operating expenses for the business, and ultimately have a higher profit margin over time.

Will the price go up later?

This is informed speculation, but the main argument for why it might is Capitalism. If Open AI wins the market, they could certainly hike the prices.

However, that feels unlikely.

Instead, we are likely to have multiple major players in the foundational AI model space. We’ve talked about this in the newsletter before. And more competition will mean more options and cheaper prices.

And ongoing research is pushing both the hardware and model training costs for #1 above to be more efficient.

USE CASE DEEP DIVE

Create customized videos entirely with AI

Are you a creative looking to stay on top of the latest advancements in generative AI? If so, you’ll be blown away by this next use case deep dive!

Community contributor Claire Xue, Founder of Moodelier, shared her incredible workflow for creating entire videos, from start to finish, with AI.

Claire explains that by combining the power of various AI tools, she is able to create visually stunning and compelling content that resonates with her audience.

From ChatGPT for the voice narration script to Midjourney for still images and Kaiber AI for video clips, this impressive workflow is a great example of the potential of generative AI in the creative industry.

While Claire notes the videos created with AI tools are far from perfect, they represent an exciting step forward in video creation. By using this workflow, Claire is able to create impressive AI-generated content that attracts more people to her free resource page, "The AI Stuff Creators and Artists Should Know".

If you’re interested in learning more about the tools and workflow Claire used, head over to the full Stack to find out more!

LINKS

For your reading list 📚

Following Hubspot's lead, Salesforce is joining the AI wars by...

And if you're really nerdy...

Member only links (more info on joining here!)

That's all!

We'll see you again on Tuesday. Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected]. 

... and if someone forwarded this email to you, thank them 😉, and subscribe here!

Cheers,

🪄 The AI Exchange Team