Great Stories Will Find a Way
Why we backed Jamit—and Ike Orizu's bet on African voices as the next frontier of AI-native storytelling.
LAVA is, more than anything else, a bet on African talent. It is not necessarily that there is more talent in Africa, or that it is qualitatively different than the rest of the world, but rather that the talent in Africa has been systematically undervalued and overlooked.
While we are committed to our chosen investment thesis: “simple finance” and “trust infrastructure”, we will occasionally step a little outside of this if we find founders who exhibit the sort of world-class talent we know exists in Africa. Ikenna Orizu is one such founder.
Ike led engineering teams at Roku and NewsCorp. He has a depth of experience in content creation and media tech (both hardware and software) that is difficult to equal anywhere in the world, and he is uniquely positioned (both in terms of skills and personal network) to produce something interesting, different, and valuable.
Jamit revolves around an age-old human practice, with deep roots in African soil: story-telling. This is one area in which Africa might actually have qualitatively better and quantitatively more talent than anywhere else on the planet, but that talent is only realised–only made really meaningful and valuable–when it is shared.
Ike has built an AI-native storytelling platform, powered by a custom AI model called Magic Producer, where anyone can create studio-quality audio stories and recordings with one click and at the same time scale their original IP globally. It’s like webtoons for audio, powered by intuitive tools that understand how you speak and how to make it sound captivating and beautiful to your audience.
Originals
As engineering lead at Roku, Ike had a front row seat as Netflix Originals was built after Disney pulled their content. This is the other side of Jamit’s play: longer form, fictional “Jamit Originals” with well-known actors voicing the content.
He has already seen surprising success with this content format in multiple global markets. While long-form, non-fiction podcasts with celebrity hosts (think: Joe Rogan Experience, Call Her Daddy, etc.) have already saturated their market, there is an appetite for long-form fictional content that people listen to while commuting, doing menial tasks, or when relaxing. This is especially true in the Indian and Asian markets.
Licensing the IP for this kind of content, if it is successful, is an attractive way to extend Jamit’s reach beyond the app, enhancing discoverability and enabling Ike to achieve success even if people remain in their current favourite places when listening to content.
There is a potentially potent flywheel effect here: find a great Jamit Original in your current favourite place for listening, get curious about what Jamit is, and discover a whole world of user-generated content (that is studio-quality) in the app, including using the best of web3 features to power global payments to creators, making it easy (and rewarding!) to engage more.
Augmented
Magic Producer–Jamit’s proprietary model that edits audio–performs on a par or better than the best commercially-available ones from studios Ike has worked with, who specialise in this. But models are not a moat, and the exponential curve we are on will likely make a lot of this obsolete by next month. What we’re most excited about is the data. While we can improve the programming, mathematical, and scientific skills models exhibit by using synthetic training sets produced by previous generations, it is unlikely we can get LLMs telling great stories with only the same synthetic training sets. Our contrarian take is that telling great stories matters more to widespread, economically valuable use than novel mathematical proofs.
The logic here is roughly the same as E. O. Wilson used in Consilience: the “social” sciences (everything from economics through the rest of the humanities) are much harder to do than the “hard” sciences. Consequently, if models are to do more than programming, math, and the physical sciences, we need better data. In fact, the need for better data has been clear since the Chinchilla Paper from DeepMind in 2022. We feel this is something African founders can meaningfully contribute to, and we feel that Ike is the best-placed one we have found so far.
We are more bullish on AI tools that augment human intelligence, connection and cultural practices, rather than those that seek to displace the need for thinking, reasoning, or consulting others. While Magic Producer can produce stories from a prompt, the best results we’ve seen in our own testing revolve around cleaning up recordings of human speech to make it sound professional. Watching how people collaborate with Jamit’s custom models to refine their stories, and then make their recordings sound that final 10% smoother (which generally takes 90% of the effort), will be fascinating.
With over 500M global podcast listeners and double-digit annual ad growth, the opportunity for a new storytelling infrastructure is massive.
Secondary Orality
In addition to the technical skills and background Ike has, he has also been involved in the global media scene for his entire professional career. He knows the story behind the story of IrokoTV because, as the teenage founder of Truspot, he was raising funds at the same time.
Having been humbled by these early experiences, he knows how to avoid similar errors; who to deal with and how to do so fairly; and how to create technologically-enhanced communities that actually empower creators and help people get their voices heard.
The rise of podcasting is one more proof of Walter Ong’s notion of “secondary orality”: a term he coined to describe the communitarian effects that electronic media can induce. Jamit turbocharges these effects with powerful tools for augmentation, and a world-class level of crypto abstraction to make the boring financial stuff (payments, subscriptions, profit shares) and the rewards for listener loyalty easy to navigate for anyone, anywhere.
We’re excited to invest in Jamit and learn more about how to create successful consumer applications with frontier technologies that augment what humans have been doing here since time immemorial. From Africa, for the world.