In 1956, computer scientist John McCarthy coined the term “artificial intelligence” (AI). In the years since, major technological milestones in the evolution of AI include the supercomputer Deep Blue defeating a world champion chess player in 1997 and the commercial availability of technologies like speech recognition, robotic process automation, and deep learning.
Moreover, In 2022, generative AI exploded into mainstream popularity and usage with the release of the chatbot ChatGPT from OpenAI. Led by CEO Sam Altman, the company was founded in 2015. While generative AI is far from a new phenomenon (it dates back to the 1960s), OpenAI’s ChatGPT and applications from other developers have made the technology more accessible with simple, conversational user interfaces.
“The mainstream availability of generative AI is a phase-shift in human development,” says Patrick Chung, General Partner of venture capital firm Xfund. “A synthetic dry-intelligence mind is starting to interact with our wet-intelligence brains, shaping them both.” Chung was Sam Altman’s very first investor in Altman’s first company out of Stanford, Loopt. Chung served on his board for almost eight years.
What Is Generative AI?
Generative AI is a powerful form of AI that is primarily used for content creation. It represents an advancement in machine learning techniques. Earlier machine learning approaches used labeled datasets to train algorithms to classify data and predict outcomes. In contrast, generative AI models rely on unsupervised learning—they can determine the structure of unlabeled datasets, recognize patterns on their own, then generate new data based on this. The upshot is that generative AI can produce content without a defined or binary “correct” output based on the user’s instructions. The technology can create text, audio, images, or synthetic data within seconds.
OpenAI’s ChatGPT, the most widely known generative AI chatbot, is based on the company’s Generative Pre-trained Transformer-4 (GPT-4) model. Its conversational interface makes it easy for basically anyone to use, and it is able to generate complex responses to detailed prompts written in plain language.
ChatGPT was used by more than 1 million people within five days of its release to the general public in November 2022. Since then, users have demonstrated its impressive content-generating proficiency by sharing examples of the chatbot’s poems, computer code, and college-level essays. Amherst College professor Corry Wang gave ChatGPT a “solid A-” for an essay it produced in 10 seconds on Ernest Gellner’s and Benedict Anderson’s theories of nationalism. Someone else programmed ChatGPT to run a 24/7 stream of AI-generated Seinfeld episodes on Twitch. In March 2023, OpenAI released GPT-4 and has integrated this more advanced model into its ChatGPT Plus product.
Generative AI Applications
While generative AI can be employed for fun and creative parlor tricks, it has the potential to transform and optimize business processes. Specifically, it can quickly create effective targeted marketing and advertising campaigns, assist with product development, and write code for software companies. Chatbots and virtual assistants can also accomplish many customer service and support tasks. In manufacturing and supply chain management, generative AI can enable predictive asset management.
Generative AI also has the potential to improve medical research, drug development, healthcare, and creative industries. In healthcare, it can create personalized patient care plans, advance medical imaging and diagnosis, and provide administrative support for tedious, time-consuming tasks. For media companies, generative AI can assist with translation, content creation, and social media monitoring, among other use cases.
AirOps
Xfund supports startups that are creating generative AI tools that demonstrate the technology’s power and wide spectrum of applications. Two examples from the Xfund portfolio include AirOps and Delphi.
AirOps, which recently announced a $7 million seed funding round with support from Xfund and other investors, is an AI-based app-building platform that allows users to integrate their business data and logic, along with ChatGPT, to optimize various processes and activities in a structured and predictable manner. The platform is data-driven, flexible to changes in the world of AI, and easy to deploy without expensive engineering cycles.
While users will soon be able to create their own apps on the platform, several are already available for use. These include Text Generate, Text Summarize, Text to Takeaways, Video Summarize, and Magic Transform, the latter of which will instantly format or transform any type of data. There are also coding apps, such as SQL Explain, which immediately translates SQL queries into easy-to-understand terms. Other coding apps include SQL Insight Generate, SQL Write with Pasted Schema, and SQL Write with Database.
AirOps’ code-free apps support media (including MP3, MP4, and .doc), run at scale, and can be accessed via web browser, Google Chrome Extension, Google Sheets, API, and Snowflake.
“AirOps makes it possible for anyone to employ generative AI in a turnkey fashion,” says Patrick Chung of Xfund.
Delphi
Delphi, another of Xfund’s AI portfolio companies, is deploying generative AI in a different, more personal direction by allowing users to create and interact with AI versions of their heroes, whether living or dead. Marketing itself as the “Waze for Life,” Delphi imagines a world where its customers use its innovative tech to create their own board of advisors from which they can glean tips and insight into any problems or life obstacles they may have. If you’ve ever wondered how Steve Jobs would manage a specific business decision or how various Greek philosophers would react to modern problems, Delphi has answers.
Delphi maintains a collection of AI clones that includes Margaret Thatcher, Warren Buffett, and Jeff Bezos. Delphi offers plus, pro, and premium subscription packages, the last of which includes access to unlimited clones in addition to unlimited messages and documents and priority support. Soon, users will also have the ability to monetize their clones.
“Although our users know they’re not actually chatting with Warren Buffett, they are still drawn to Delphi’s AI clones because the insight is real,” added Patrick Chung of Xfund.
