OpenAI vs. VC

OpenAI

As of 2023, the OpenAI Company Recorded More Than 100 Million Active Users Monthly. In 2022, OpenAI had 21.1 million people actively using its platform. Impressively, just five days after it became publicly available, it already had one million active users. All acquired via word of mouth.

Updates (GPT 4-Turbo)

OpenAI just released it's new version, link to it if you want to watch it - here are the high level notes which you'll need to take into consideration as you read further. I'll keep them brief.

Context Length

Previously - It could hold around 12 pages of a PDF (8,000 tokens).

GPT - That can now ingest 300 pages. A book length. What's even more impressive is that its understanding of that "book" has been improved significantly. It understands all of those 300 pages and how they correlate to one another.

Improved Knowledge

Trained up to April 2023 and if you connect via Bing, you're able to access real-time data.

Copyright Shield

OpenAI have boldly say they'll cover the costs of any copyright claims against anything you build via their API.

GPTs

This is the core release that's going to shake things up. A GPT is your own GPT (not the generic ChatGPT). You can give it access to your systems, integrate with one of their partners, during the demo day they used Zapier. This meant GPT could integrate with all 6,000 applications, just by asking it to connect two applications you have in your business together.

Custom GPTs

In the example they showed, Sam Altman uploaded one of his transcripts for a lecture he held. He could then ask that GPT questions. It responded with what he'd typically say as it had ingested that data "duh". Okay but...

What does this mean?

Companies can now, safely, ingest their own data to build their own GPTs. Sure, someone would need to manage this project but it's literally as simple as either, a non-technical person dragging and dropping files. Or, assuming you have a dev team, asking them to take an hour to string together your own GPT.

Well, it's actually a lot more than that. A lot, lot, lot more than that.

One category in real trouble: Knowledge base's

• This is one use-case. There are many many many many more. I picked this one because it's the easiest to implement. As of 3 days ago, I have spoken to 5 founders who are all doing this.

As of 2022, this markets estimated to be valued at $558.4 Bn (Source).

Businesses use knowledge bases to bring information from different sources, such as internal communications platforms or email software, into a self-serve digital library that acts as a single source of truth. Businesses can determine what content they want to collect in the knowledge base, how it is used, and who can access or edit the information. The content can range from training documentation to project timelines to meeting notes and action items. Knowledge bases allow users to organize information effectively and offer powerful search capabilities, which help employees find information efficiently and trust that it is accurate. Knowledge bases can be accessed by a large audience, but they are typically managed by a limited number of contributors. Contributors usually include those with more expertise or oversight into particular topics, such as subject matter experts, team leads, and project managers.

Source G2

I went through several of these companies on Crunchbase, I won't name any specific ones but here is some data against a few:

Disclaimer 2: some of these companies will have very specific use cases that it will not be able to do unless someone does bespoke dev work.

Company 1:

Total Amount Raised: $70.7M - Series C

First Investment: Jan 2014

Last Investment: Apr 2020

Company 2:

Total Amount Raised: $25.5M - Series A

First Investment: Jul 2019

Last Investment: Jan 2022

Company 3:

Total Amount Raised: $33.8M

First Investment: Dec 2019

Last Investment: Jun 2021

What's even more interesting is the pricing model ranges (from what I saw) from $29 (per user per month) to $2,000 per month. How're you going to convince a company to stay with you when I can do all of this for $20 a month with:

• No onboarding
• 10 minute setup time
• Zero training required (no learning curve)
• No touch points from CS, QBRs, renewal conversations, upsells.

Even your ice to eskimo's CS/Sales professional can objection handle that.

Why this category is totally f@￡ked?

Overnight, OpenAI just released the ability to upload all my existing training manuals, documentation, playbooks, onboarding. ABSOLUTELY ANY DATA YOU HOLD AS AN ORG to build your own GPT.

It will be able to answer any question you have, quicker and without you having to login and mess around in one of the aforementioned companies UI (also, no learning curve, adoption etc. It's just ask the question, get the answer). ZERO Navigation.

Managers, CEOs, Directors, whoever the department head is can now build their own knowledge base with no technical knowledge in 5 minutes (just dragging and dropping). Share the link and away you go.

Counter Arguments

We track if someone's read it!

You can access the log of your GPT and see what questions been asked, who has read it, the response they received. Aka Tracking.

It integrates in our CRM.

You know how I mentioned Zapier...?

We help build the collateral.

ChatGPT, transcripts from calls, previous emails, feed whatever data you want to it and it'll build the above for you.

For existing investments in these companies

All of these companies will start churning customers, sure it'll take a while for people to begin realising this is possible. Reduced headcount and before you know it. Unless they pivot, the markets going to disappear right before your eyes.

I am unsure if it's even possible to pivot successfully as much as they need to, to protect ones investment.

If any of your portfolio spring to mind when reading this, I'd be proactive and not reactive. Chat with the founders and see if it's possible to do an early pivot. It'd be naive to think that they're AI proof. Don't blame the founder for not having a crystal ball and try to reason with your LPs about an unseen iceberg.

For future investments

Every company should now be adding the following to their due diligence questions and process.

Is this company truly AI proof?

The rate of releases of OpenAI and the introduction of the GPT marketplace. Thousands of apps are going to appear from the 2,000,000+ developers using their API.

Vitamin companies or "nice-to-have" solutions just won't cut it anymore. They need to be doing something truly unique and unfortunately, complex to warrant getting venture capital.

The defensible moat is getting smaller and smaller.

How do I know this?

It happened to me. I built Comtura which ended up failing because of OpenAI. We built the ability to automatically update CRMs using transcripts. I was doing demo's and I saw reps put in a transcript, ask questions and fill in the notes. I wanted to charge $79 per user per month and a huge portion of that was computation cost for our clever NLP/ML models. That was one rep, however, I knew that the responsible thing to do was inform our investors. No matter how painful it was. With this new release, I was correct in my prediction.

Why pay that much for something that was, sure a bit manual but now with this new release. Via Zapier + custom GPTs - that entire idea is worth nil.

It was impossible to predict and even harder to see or know what over 2,000,000 developers are currently working on.

Final thoughts

I listed one category, their are many more that I'm confident are swiftly going to become engulfed by affordable solutions. Writing this now I believe the following questions are important to ask yourself:

1. How can I use this incredible technology to get ahead? Whether in your job role or making a project. - If you're not, you're missing the modern day gold rush.
2. If I'm moving company, is that company AI proof?- If I work for one of these companies, are the leadership team preparing adequately for a change to protect myself and my career?

I for one, am extremely excited to see what crazy developments are going to come out of the wood work.