How can YOU start a Generative AI company?

Three insights to help you start a new GenAI company and more.

Barun Pandey
March 10, 2023

GM! Welcome to the Status Code.

We're like skilled pilots, navigating you through the vast expanse of AI.

Here’s what we have for today:

🐣 Companies are flocking to build their own GenAI
🤑 VCs have plenty $$ for GenAI companies
⏲ Three insights to help you start a new GenAI company

(Estimated reading time: 4 minutes 30 seconds)

Not subscribed yet? Stay in the loop with AI weekly by reading the Status Code for five minutes.

Subscribe To The Status Code

Two Headlines

Two main stories of last week. If you have only ~2 minutes to spare

1/ 🐣 Companies are flocking to build their own GenAI

It seems companies are doing two things in 2023. Laying off employees and devising a Generative AI roadmap.

There were five major announcements last week:

Salesforce and Hubspot’s ChatGPT bots

Both top players in CRM announced they are integrating ChatGPT API in their platforms.

The value to customers? Writing content faster.

Salesforce is going to offer different models for their customers to pick from. But, they will start with OpenAI's GPT model for launch.

Meanwhile, Hubspot is going one step further. They also announced ChatSpot.

ChatSpot will allow you to do everything in Hubspot but with chat-based commands. One step towards natural language becoming the way we interface with software.

Slack’s new assistant bot

Slack, owned by Salesforce, is bringing a ChatGPT-enabled bot too.

The bot will be able to summarize conversations and draft new messages. Users can also use it as a research tool.

DuckAssist’s Wikibot

Step aside, Microsoft and Google. There’s Privacy AI.

DuckDuckGo announced its AI-assisted search. The privacy-focused search engine trained its model on Wikipedia.

But it’s completely anonymous and private. I know my go-to place to ask questions like: “Who holds the record for drinking most beer in a single day” now.

CFOs don’t need interns now

Brex, a financial platform for startups, is also joining the party. They announced a new suite of AI-powered tools for CFOs.

The tools will allow startups to get real-time financial insights through natural language. What kinda insights? Like saving opportunities and spending trends.

So, the next time a CFO wonders, “What’s our subscriptions spend compared to other companies like us?” an intern doesn't need to be all sweaty.

Instead, CFOs can put their query in Brex and voila, an answer will pop up, with graphs and all that.

2/ 🤑 Have a GenAI startup? Here’s 1B$ for you

Global VC funding fell to $18 billion in February 2023, down 63% from last year.

Yet, Generative AI startups don't seem to be on bended knees for the VC war chest.

There were many funding announcements last week.

The biggest was Anthropic AI (yeah, Google’s $400 million investment wasn’t enough). They raised $300 million at a $4.1 billion valuation.

But there was more movement.

Reid Hoffman, a top investor in Silicon Valley, is announcing he’s leaving OpenAI’s nonprofit board to avoid a conflict of interest.

Reason? He co-founded another AI startup, Inflection, and it’s making moves. Reports suggest Reid is in discussions to raise $675mn. Inflection is working on an AI personal assistant.

Besides, Reid wants to up his GenAI portfolio. He wrote on LinkedIn: “There are future trillion-dollar companies being built and invested in right now, which will change markets and launch new ones”.

Wait, it doesn’t stop.

Cohere, another Google alumni startup, is also raising at a valuation of $6bn.

Stability AI, creators of Stable Diffusion, is raising at a valuation of $4bn.

No wonder, Salesforce is announcing a new investment arm. It’s a $250 million generative AI fund, which they say will help bolster the startup ecosystem.

Those are fancy words for saying: “I want in on the hype train”.

One Trend

1 trend you can pounce on. Reading time: ~2 minutes 30 seconds

⏲Three insights you can use to start a new startup

Is GenAI a once-every-15-years opportunity?

Well, that’s why VC spending has gone up. But VCs said the same for Crypto. So who knows?

But Generative AI can be different. Many studies suggest it can be as effective in productivity as Steam Engines.

And unlike Crypto, the use cases are simple.

It’s not like VR; you don’t need fancy hardware.

And compared to self-driving cars, it doesn't have to be perfect to avoid risking someone’s life.

Generative AI may be the next wave of the internet.

For founders, this is a golden opportunity. Any new technology platform leaves plenty of low-hanging fruits to pounce upon.

I researched a few frameworks/models that you can use to create new businesses:

Shift in how tech teams work.

(Insights from Tomasz Tunguz)

Traditionally, we had a core engineering team and a data science/machine learning team working independently.

Core Engineering Team focused on building products and maintaining them.

Data Science/Machine Learning Team analyzed data and built ML models to support business functions.

But, ChatGPT is bringing new expectations to consumers. They want products to be intelligent.

So, Tomas predicts companies will merge their data science team into core engineering. Core Engineering teams need data scientists’ expertise to optimize the end-user experience.

This will open a new opportunity for startups: to build platforms that unify both and allow for that collaboration.

Generative AI as a ‘co-worker.’

(Insights from Ethan Mollick)

A recent study lists the top 10 occupations AI will likely affect the most.

Most of them include teaching. It is opposite to what people believed - that AI will first replace manual labour.

So, there will be no teachers? No. It's not replacing the entire job. It's automating job functions where automation was rare.

Why was it rare? Well, because those job functions included many tasks.

Teaching requires more autonomy. And it involves many tasks like prepping lessons, giving grades and writing recommendations.

But, Generative AI has opened a way for teachers to drop repetitive tasks. So, they have incentives to adopt AI.

And it will be like a co-worker relationship.

People will work with technology instead of having it work for them.

And for startups, it's an opportunity. Find job functions where automation was rare before. And discover if Generative AI can contribute to the co-worker relationship.

Start new businesses that redo old businesses.

(Insights from NFX)

There are a few core characteristics of Generative AI that you can use to redo old businesses:

Zero To One -> Zero To Ten

Most Generative Tech companies start with solving a zero-to-one problem. Take Jasper, for example. Copywriters can use it to write the first drafts of their copy.

But, they will evolve into providing zero to ten. They will meet the complete needs of the user.

Jasper will write a copy that’s going to be publish-ready.

Animation GenAI companies will create movies that are Disney-quality.
Replace Curation with Creation

Curation was the name of the game for the past decade. Spotify won the Music game with their A-star recommendation. Netflix knew your choices for movies better than you.

But, with GenAI, a new paradigm for personalization will appear.

Instead of suggesting movies based on preferences, apps that replace Netflix will create new shows for you.
Low-friction interfaces

One defining attribute of Generative AI is how it allows for low-friction interfaces.

Traditional UX experience was limited.

Take Hubspot, for example. Even with massive investments in UX, it took a new user a great amount of time to get used to its interface.

But, if they get ChatSpot to work, users can use simple text prompts to get the desired result. AI will do most of the heavy lifting.

New startups can use this 10x change in interfaces to redo old platforms.