Bad Ideas are Good Again
Now that we have large language models (LLMs) and generative AI, as an AI fund, we get asked all the time “what do you invest in now?”
One answer: “Bad Ideas.” No seriously.
Venture has become almost boring. Many categories are played out, crowded, and seeing diminishing returns. It is difficult for a new idea to be truly fresh and to break out. However, generative AI is resetting everything, across the board. All this new technology is forcing us to challenge fundamental assumptions around how industries operate. As a result, many of what we used to consider Bad Ideas are now Good again. We are excited for what’s to come and to invest in seeds today that will shake up entire industries. There are a number of ways this is happening. Let’s dive in.
Bad categories are now Good
Gaming, entertainment, and media have been considered dead categories for venture investments because of their return profile. Gaming studios and media IP franchises are well-known as hits-driven businesses, which makes it difficult to either invest directly or into businesses that sell into them. Generative AI fundamentally disrupts the cost structure of these businesses, which will shake up the industry and create opportunities for big winners.
The disruption
The major cost centers in these businesses — where the most time, money, and resources are spent today — are precisely where generative AI is best applied to. For a gaming studio, it is the labor-intensive process of creating art, 3D models and animation, which grows more content to monetize. For Hollywood, it is visual effects (or VFX), using many techniques to bring to life on screen what’s not present in real life, whether it be fantastical characters, scenes or motion.
AI has conquered 2D imagery (with Stable Diffusion, Midjourney, and DALL-E etc), and the frontier is moving towards 3D models and animation. Today, an animated feature needs 24 frames per second and each frame needs to be created by hand. Tomorrow, such works can be generated in minutes with AI. Today, to make a 3D character look and move with realistic human motion requires the manual workflows of rigging, motion capture, and frame-by-frame cleanup. Tomorrow, AI can generate realistic motion in real-time that abides by the laws of physics and is tailored to each character.
A number of tech players are working on the many aspects of this and selling into the incumbents. To name a few, Ziva focuses on realistic facial motion, Wonder Dynamics focuses on CGI for Hollywood, Uthana focuses on animation in games, Mirage focuses on generating 3D scenes, and Volinga applies neural radiance fields (or NeRFs) to virtual production. All will drastically reduce the cost of production or make the impossible, possible.
Rise of the indies
Disruption will also come from below and we’re seeing this in the rise of the indies. When the cost to produce asymptotes to zero, every kid in their bedroom can create games and films; the limit is only their imagination. These indie creators will be more forgiving of imperfections as the tech improves and they’ll be more willing to try new tools vs professionals in industry. New form factors for entertainment will rise, not limited by today’s constraints in production. The successes of Roblox, Minecraft, and Fortnite have shown us what is possible when you make a game that enables players to create games.
Many startups are already working on this as well. For example, Luma AI applies NeRFs to empower anyone to create photorealistic game assets by just capturing a cell phone video. Versed, Dreamlab, and others allow anyone to make games by generating scripts, assets, levels, and animation. Versed and AI Dungeon are reimagining role-playing games, starting with the fanatic Dungeons and Dragons community. Linum empowers anyone to generate video. Talofa is building the Roblox of fitness gaming for gen Z. We will see an explosion of creation.
Bottleneck shifts will create new winners
In a future where labor for tasks like animation, VFX and 3D modeling are no longer bottlenecks, the plane of competition will shift. When a AAA studio can make 100 high quality games per year vs one and when indies can create high quality games and films without big budgets, the bottleneck will no longer be studio capacity, but rather the capacity of humans to create and consume higher-quality content and new and more personalized experiences.
In 5–10 years, the leaderboards for AAA studios and top Hollywood studios will be reshuffled in a major way. The ones today that cannot or will not adapt to this oncoming revolution will fade. Tech is advancing rapidly on many fronts and while it’s premature to predict exactly where the value will accrue, we can take a hint from history.
No industry is perhaps more acquainted with the implications of technological disruption than the entertainment industry. From the advent of the talkies to streaming, every new technological innovation has created new winners and losers, new dimensions of competition and ultimately stronger content and more optionality for consumers. Today’s shifts will propel new layers in the stack and new platforms to prominence, creating new winners. Technology won’t alleviate work, it will merely elevate expectations and incentivize competition in new ways.
Call for Bad Ideas
Generative AI is changing the game for venture investments, enabling us to look at categories previously full of Bad Ideas with fresh eyes. Gaming, entertainment, media, and many other areas that were once seen as dead categories are now becoming attractive.
We are at a unique point in time where the laws of physics suddenly work differently. The assumed advantages for incumbent players are going away and the rules of how industries have worked can be challenged. By fundamentally disrupting the cost structure of these businesses, generative AI is creating opportunities for big winners and shaking up entire industries. The rise of indies, the creation of unique form factors for entertainment, and the emergence of new bottlenecks will all create fresh winners, leading to a major industry reshuffling in the coming years.
As an early stage VC, we feel so much excitement for what’s to come. We will continue to share our thoughts in the coming weeks. If you’re working on a Bad Idea (that you believe now to be Good Idea), we would love to hear it! Shoot me a note at chang@basisset.com
Many thanks to Matt Aimonetti and Eduardo Fonseca for reading drafts of this post and providing valuable feedback.
