Creative machines
A few thoughts on AI + the future of creative tools:
AI makes creative tools easier to use, which fuels user growth.
Lots of potential users fail to adopt creative tools like Figma or Replit because they lack some set of skills. For Figma, it may be familiarity with the tool itself. For Replit, it may be coding competence. My expectation is that new forms of AI assistance are going to flatten learning curves and help more people get started with these tools.
AI makes creative tools more powerful, which will yield more creative output. When creativity gets easier, we’ll see more of it. I expect the innately talented to become more prolific and the less naturally gifted to discover that many creative domains are no longer off limits.
The arc of creative leverage: AI will work to diminish the gap between thinking and making. As AI assistance becomes more and more adept at producing high quality content, we will see that surfaces which are upstream in the creation process become strategically more important. In other words, the tools we use to think about what we want to create (e.g. Github, Figma, etc…) may end up preempting downstream creative tools. As we commoditize the process of creation, we should expect value accrual to the surfaces that facilitate our thinking.
Because creative tools are poised to become easier to use, we should expect to see viability of new product categories. Many categories have been ill suited to mass markets because the technical hurdle was too large or because the audience with requisite skill and/or talent was too small. For the reasons I’ve laid out above, I think new categories are poised for viability. In that light, I continue to be very bullish on building design tools for domain-specific categories. E.g. design tools for furniture, interior design, fashion etc….
Playgrounds capture increased creative demand. As demand for creative output increases, users will seek tools that can deliver it for the lowest cognitive cost. While generally used for experimental, non-production use cases, “Playground” apps (e.g. Replit, CodeSandbox, val.town, Streamlit etc…) fit this mold.
The magic of playground apps is that they miniaturize the atomic unit of creation1. JSFiddles are miniaturized apps. Jupyter cells are miniaturized scripts. When the atomic unit of creation becomes smaller (like tweets vs blog posts), there’s a lot less cognitive friction involved in the creation process. Playground apps will likely be very attractive to users who find themselves increasingly creative and eager to experiment with new capabilities.
The dominant meta-question in AI seems to be whether value accrues to incumbents or startups. What I’ve laid out above pertains to a specific category of software but reveals some useful patterns that may shine light elsewhere.
Skill + talent barriers disappearing makes new product categories viable.
Rich-gets-richer effects for incumbents which own workflows/user-activities that are poised to increase in strategic significance (e.g. creative surfaces stand to gain from increased ease of creation).
Business processes (practices + tools) are inherently human-centered. But as machine intelligence becomes a more reliable collaborator, we may end up redesigning business processes to better leverage machine intelligence.
Playground apps have lots of strategic advantages.
They’re a natural nexus for collaboration. Micro-units of creativity are easy to collaborate on.
They have extremely fast time-to-value, which is helpful for activating new users.