• NATURAL 20
  • Posts
  • OpenAI's Foundry Leaked Pricing Shows AI Potential

OpenAI's Foundry Leaked Pricing Shows AI Potential

The pricing for OpenAI's "Foundry" project gives us great insights into where they think things are heading.

Author

Wes Roth
March 01, 2023

The Norse God, Odin, had two ravens who brought him information from all over the world.

*YOU* however, have something even better... this newsletter!

We will bring you news of AI, tech and business from all over the world.

So forget your plan of trapping/training birds to do your bidding.

It doesn't end well:

Rick and Morty Two Crows

Rick and Morty Two Crows

OpenAI Leak!

The pricing for OpenAI's "Foundry" project gives us great insights into where they think things are heading.

Foundry is the "platform for serving OpenAI's latest models".

Basically, the enterprise level of AI that OpenAI will make available for purchase.

Basically it's ChatGPT on steroids and also improving really, really fast.

foundry

Foundry

What is Foundry?

Foundry will come with "robust fine-tuning". Meaning that OpenAI will use their impressive knowledge base to help companies mold the AI platform to do their bidding.

An industrial foundry is where metal tools are produced.

So an AI Foundry would be where businesses produce their "cognitive" tools.

For example HR Block could automate 90% of what their tax professionals do by hooking up with Foundry and letting OpenAI fine-tune the outputs to create AI that helps people complete their tax returns.

Businesses with large call centers can create an AI that answers phones, emails and texts for a specific customer base (think something like tech support or banking).

Already Google/Deepmind have plans for AI models that approaching the performance of human clinicians called Med-PaLM.

75% of the US economy is services.

AI tools like this will allow most businesses to attempt to replace their service workers, cognitive workload and knowledge work, away from humans and unto AI.

The other impact can be on optimizing things like sales calls, for example.

Let's say some corporation kept a massive amount of call recordings spanning decades.

Feeding this data into AI to try to learn from to optimize the sales pitch and then creating the perfect "AI salesperson" capable of handling an unlimited number of calls for a fraction of the cost of it's flesh and blood counterparts.

The Cost

Foundry pricing is as high as $1.5 million per year.

This is for the large 32k "context window" (see explanation below)

Eagle-eyed Twitter and Reddit users spotted that one of the text-generating models listed in the instance pricing chart has a 32k max context window. 

(The context window refers to the text that the model considers before generating additional text; longer context windows allow the model to “remember” more text essentially.) 

GPT-3.5, OpenAI’s latest text-generating model, has a 4k max context window, suggesting that this mysterious new model could be the long-awaited GPT-4 — or a stepping stone toward it.

TechCrunch (article links below)

Read More About It:

Cognitive Revolution on Substack

TechCrunch:

Poll on Yesterday's "Elon Musk and Woke AI" Article

So 62% of you feel that 

"ELON IS RIGHT - AI should not be censored and it should not reflect the biases of it's developers."

Not too many people voted, but thank you for all who did!

What'd you think of today's edition?

Login or Subscribe to participate in polls.

What are MOST interested in learning about AI?

What stories or resources will be most interesting for you to hear about?

Login or Subscribe to participate in polls.

Join the conversation

or to participate.