How OpenAI Works with Startups
Pointers directly from OpenAI's Head of Strategic Startups, Marc Manara
Last week, we had the chance to interview Marc Manara, Head of Strategic Startups at OpenAI, about how they work with startups. He shared some great insights including:
The best resources to use when you’re building with OpenAI
How startups can incorporate AI into their plans in the most high-impact way
Info on the Converge 2 program (applications due Jan. 26)
And and other juicy nuggets. Whether generative AI is core or tangential to your business, there is helpful advice for you in the summarized highlights below.
OpenAI's Startup Team Works Most Closely with Startups with AI at the Core
Marc and team aim to support startups across all stages, particularly in building with AI to showcase its impact. They provide resources like quick starts, documentation, a playground, and Cookbook (an open-source collection of examples and guides for building with the OpenAI API).
Startups are essential to how OpenAI and show the impact of AI in our lives, since they’re the ones doing some of the most innovative stuff with the tech - so the team wants to make it as easy as possible for startups to get started, and to support scaling of the technologies so startups can build faster and more effectively. They have a small team, so they try to find the biggest impact things they can do to support startups across the board.
That means that the team tends to work most closely with larger VC-backed startups that are building AI into the foundation of their products. These are the startups pushing the limits of OpenAI’s technology and testing them in the wild with their customers. They try to be helpful and supportive to enable startups to get the most out of the tech as they scale.
They also want to be the listening device for the org on what kind of feedback they’re receiving from startups across all stages, and how this feedback can be incorporated and prioritized on the OpenAI product roadmap, so they can be aligned with their customers’ needs and help them move fast.
Advice for Non-AI Startups
There are plenty of startups and businesses using AI in ways that are important to them, though it might not be core to their product offering.
Marc’s advice for how these companies should best utilize OpenAI is to identify high-value AI use cases to improve the customer experience. This tends have the most impact and value for them.
Often, Marc hears about companies whose board, investors or management team are asking about a company’s AI strategy, and the company ends up shoehorning AI into their products without clear business objectives. This can just end up being a waste of time.
Instead, he recommended checking out the customer stories on the OpenAI website if you are in need of some inspiration of where might be the highest value place inside your product to use AI. Then when you have some ideas, there are pretty good quickstart guides on the OpenAI API documentation, and you can use the playground to test these things out with some of your real data. Check out the OpenAI Cookbook which has a lot of examples with actual code and explanations for solving different types of problems with sample code you can lift and apply.
Trends in AI Startups and what Startups Do Well
There are tons of people creating AI applications in every single possible niche vertical, and yet every day, they come across people with new, exciting applications. Investors seem to be most interested in the ones where the founders have extensive domain expertise and some kind of differentiation in their execution and traction, not the ideas alone. For example, access to a unique data set is a big differentiator, whereas incorporating generic chatbots into an existing feature set is not as likely to create a differentiated experience or catch investor attention.
When there is a lot of competition converging in the same niche area, the way for a startup to stand out tends to be the same as for any startup success: fantastic founders, strong execution, product-market fit, growth, etc.
Practical Advice
Use the right tool for the right job. You can be most time and cost efficient by using GPT-3.5 Turbo for certain tasks like summarization, and GPT-4 when you need more reasoning capabilities.
Focus on the most high-impact use cases. For startups, this typically means customer-facing applications. Larger enterprises tend to enable a variety of internal use cases (e.g., chatbots for internal data, automating administrative tasks).
You can use your proprietary data. There are misconceptions around this, but rest assured, OpenAI does *not* train its models on any data passed through the API. Privacy is paramount and your data is yours only.
Apply to Converge 2, OpenAI’s Startup Fund program, by Jan 26! It’s a 6-week program for early-stage founders aiming to push the boundaries of AI. It provides support from OpenAI's teams, $1M investment and access.
Thanks again to Marc, and for the many founders who joined us for the informative talk!
Upcoming Lynx Collective events
Bookmark or subscribe to the full list of upcoming Lynx Collective events via Luma.
NYC Startup Founders and Operators Lunch (NYC, Jan 31 - 12pm ET)
Join us for lunch with other NYC startup founders and operators. This is an opportunity to get to meet about a dozen other startup people based in NYC in a casual setting. (Open to the public.)
Startup Bootcamp: Using Goals and Structure to Drive Outcomes
Startup Bootcamp is a monthly series for startup founders & operators covering all the essential themes relating to managing a fast-growing startup from 0 to 200 employees. Each session includes lecture, workshop & (optional) discussion covering a specific theme.
This Session: Using goals and structure to drive outcomes
In this session we'll cover:
The role of a leader within a startup org
Adapting organizational structure for different stages
Using structure to drive more autonomy and speed
Driving outcomes through OKRs
Who is this event for?: Anyone who wants to learn how to grow a startup organization - founders, operators, execs & employees. The material we cover is generally designed for startups that are in growth mode, however, it may be valuable for anyone interested in scaling organizations.