Is a new investment era here? OpenAI and Fifty Years discuss AI in Venture Capital

table of contents

As with so many other industries right now, innovations in artificial intelligence are transforming Venture Capital. VCs are using AI to automate tasks, improve decision-making, and identify new investment opportunities. 

And while AI oracles don’t exist, those at the forefront of this shift have valuable insights to share. We recently spoke to Peregrine Badger, Instigator at Fifty Years and Adam Perelman, Engineering Manager of ChatGPT at OpenAI to unpack how VCs can start folding AI into their processes. 

Watch the full conversation or keep reading for the highlights, with time-stamps to help you dive deeper at key moments.

Task automation: Screening potential investments and performing due diligence

Starts at 4:31

VCs can benefit from AI by eliminating repetitive, manual tasks so they can focus more on building relationships with founders and other valuable connections in their network. Adam Perelman said the key to finding great automation use cases is to identify areas of your workflow that require lots of information and use AI to synthesize that data for you. 

“What’s consistent across a lot of use cases,” Perelman said, “is that one of GPT’s superpowers is to give structure to a high volume of unstructured data. Whether your research process involves a first pass at scientific papers or reading a large amount of user feedback, AI is a great filtering system that can speed up the research process.”

Peregrine Badger explained how he’s doing that to source opportunities at Fifty Years. The firm’s mandate is to help PhD founders translate basic science into solutions for meaningful world problems, and they source most of their opportunities from peer-reviewed scientific papers. 

“A human could read papers all day, every day, and crack only a tiny fraction of what’s out there,” Badger said. “We were already scanning Twitter ourselves for papers, so I built a tool called Paper Scraper that sources and ranks a large volume of papers posted on Twitter. Then we pass the highest ranked papers through a ChatGPT API, and we can use the tool to narrow our scope based on some fundamental questions.”

Badger uses ChatGPT to distill papers down to insights like:

  • What did the study achieve? 
  • What’s novel about the study?

After those questions are answered, a human scientist on Badger’s team evaluates the papers for real-world translation. “If we identify a piece of work that might have potential for real-world commercial application, we reach out to the author and ask if they’ve ever thought about starting a company.”

Better and faster decision making: Identifying investment opportunities and assessing risk

Starts at 15:00

AI can act as a gut check against an investor’s own instincts regarding risks and potential. Badger said AI is particularly useful for early stages when there’s little data, and when you’re having first conversations with founders to answer some fundamental questions.  

“AI can act as a second voice when we’re able to record conversations and pass them through an LLM [large language model]” Badger said. “Then we can ask the AI, which questions have we not asked yet?”

Badger says most early conversations share one significant characteristic: You’re looking for a single big risk you’re not comfortable with. “AI can help you weigh those risks against the amazing attributes of these founders and help determine whether that one risk is something that's going to be hard to overcome,” he said. 

Perelman described how the team at OpenAI uses ChatGPT to assess risk, too. “We're not going to let ChatGPT run our product strategy,” he said, “but we do ask it to critique our strategy and help us find gaps. We also use it to synthesize interviews with important customers to help us make sure we didn't miss any of the key themes.”

Engaging with LPs and entrepreneurs: More personalized or less with AI?

Starts at 21:20

It’s an existential question: Can AI personalize communication and help us form meaningful relationships?

The consensus on the panel was that while AI can perform the task of surface-level personalization, beyond first contact it’s a human’s job to form a bond. 

Badger added to this assessment, saying, “When I really think about it, I don’t think we spend that much time trying to personalize outreach. But I do think that as AI tools become more common, people are going to receive more outreach messages from automated systems. But then these kinds of messages will be everywhere, and personalized automation will start to become meaningless.”

Perelman agreed and said the role of AI is more to “take care of everything else” so people can focus on relationships. “Before you go into that first meeting with a founder, you can use AI as your personal teacher, to help you ramp up on an area of expertise. But it’s still up to us humans to build human relationships.”

To find out how you can start using AI to improve your workflows, watch the whole webinar here. You’ll also leave with an understanding of how to:

  • Identify and researching potential investment opportunities
  • Assess the financial and operational health of potential investments
  • Track the performance of investments and making exit decision

author
posted in
share this

Interested in learning more?

Reach out to us and get a personalized demo

Talk to Sales