Technology that replicates the complexity and "human-ness" of people. Part of our Request for Startups series.
By Ryan Hoover and Vedika Jain on August 22, 2023
The world is moving from atoms to bits, decreasing costs and increasing accessibility while delivering a better solution.
We don’t write postcards, we text. We rarely pickup paper magazines, we get email newsletters delivered to our inbox. We don’t visit Blockbuster Video (RIP), we press the Netflix button on our remote. We no longer blow dust from video game cartridges, we tap an icon on our phone. We never fast forward a cassette tape, we hit “next” on Spotify.
The next big shift is in synthetic humans, software that simulates the non-deterministic behavior and “human-ness” of people with the programmability, scalability, and delivery of bits.
Of course LLMs are core to the “why now” but they’re just a part of the driver. The combination of advancements in LLMs, computer vision, generative AI, and a more connected API ecosystem have unlocked a long tail of new use cases once only achieved through real humans.
The rise of synthetic human technology is best described in early examples:
Synthetic Users and Roundtable offers user research without the people. It’s used to instantly gather intelligence or feedback on an idea from AI participants, eliminating the need to recruit, schedule, and interview humans.
Synthetic Traffic creates simulated human behavior to help designers test their product usability and experience before pushing to production.
Fictive and Outset take the reverse approach to Synthetic Users and Roundtable. They simulate the UX researcher, asking participants questions through a human-like conversational UI.
Deeptune and Unilingo translate and dub videos for creators to extend their revenues and reach to new audiences without the costly and time-intensive process of human-powered dubbing.
Synthesia, HeyGen, and InVideo create virtual people for sales pitches, marketing collateral, training videos, and more. The traditional approach to filming people is 100x slower and more costly. See an example at the bottom of this essay. ;)
Roll.ai enables everyone to film professional-level videos using the device in our pockets: an iPhone. Software simulates a professional videographer with auto-panning, zooming, and wide shots remotely.
Hallway and Hyper make it possible for people to transform into virtual characters and VTube without expensive, cumbersome motion capture equipment.
Rosebud is an AI-powered journal that functions similar to a therapist, helping users explore their inner world and find meaningful solutions to their personal challenges.
ElevenLabs, PlayHT, and Deepgram turn text into human-like voices, in any language, offloading the need for voice actors and recording studios.
Replika and Character.ai replicate rich social interactions with software. Already millions of people use these apps for entertainment and socialization with synthetic beings. That said, real human relationships probably shouldn’t be replaced with synthetic human relationships (or marriage). Fortunately it’s not zero-sum.
Bland.ai, Infinitus, and FleetWorks simulate the human voice to make phone calls, effectively serving as an API layer on archaic infrastructure. They’re saving people time from mundane tasks and making phone interoperability affordable for new use cases.
These companies (and many more not named) have an ability to deliver results at 100x the speed, effort, or cost than their real human-powered counterparts.
But many people have an aversion to the “synthetic”, even harmless fake plants. There’s a tendency to discount low-quality MVPs and feel threatened by new technology that challenges the way they’re used to working or living. Some have justified concern for job displacement whereas others see it as an opportunity to elevate their craft or do things that weren’t possible before.
We’re excited about startups building solutions that:
This shift toward synthetic humans is inevitable and will become a competitive disadvantage for those that aren’t using this technology.
And it’s getting good. Introducing synthetic Ryan (birthed by HeyGen), trying to be funny:
If you’re building something in this space, drop us a note at team@weekend.fund.
P.S. I want to acknowledge the understandable concern some have about job displacement. We view synthetic human technology as an enabler, empowering people to do more (e.g. UX researchers have limited time to conduct interviews, or they can use Outset) and opening access to solutions that were previously inaccessible or unaffordable (e.g. most creative teens can't afford $10K motion capture equipment to VTube, but they can use Hallway).
P.P.S. Read more about what we want to invest in here.
