How we made v0 an effective coding agent
vercel.com
vercel.com
The "obvious" path forward for frontend assistants is to move away from raw code generation toward some domain-specific language. UI is inherently structural - it should be expressed through component hierarchies that implement pre-defined design guidelines (with colors, margins, border radius etc being defined outside of the core model, and applied as a kind of theme to the model's output).
If we define the UI as a composition of VStacks/HStacks and predefined components, we can use diffusion models to generate the layout, and consistently apply the 'theme' afterwards. It's a much cleaner abstraction than asking an LLM to hallucinate valid CSS classes.
> Before we invoke the LLM, we replace the long URLs with shorter versions that get transformed into the proper URL after the LLM finishes its response.
So, instead of fixing your blob URLs, you effectively run a regex to search and replace on the prompt before you feed it to the LLM? That does not seem like high quality engineering.
> In addition to text injection, we worked with the AI SDK team to provide examples in the v0 agent’s read-only filesystem. These are hand-curated directories with code samples designed for LLM consumption. When v0 decides to use the SDK, it can search these directories for relevant patterns such as image generation, routing, or integrating web search tools.
Curated or created? If you "curated" them what is the copyright license on those examples? Are they just copied into the project?
Remember when programming was about creating useful libraries of code? Now it's about sequestering them inside of an LLM and then charging insane amounts of money to convince a machine to sometimes copy them into your code for you.
Just.. wow... what are we even doing? Prepare for a very fragile future.
Given that they’re talking about their own SDK that they created in this article, the examples are probably from their own docs and engineers.
> Are they just copied into the project?
No, v0 doesn't copy and paste snippets into projects. LLM coding tools don't work like that.
Providing example code or patterns for LLMs to follow is a very effective technique. They don't copy and paste or clone the code. They're good at identifying patterns, especially when you give them hints. What they did is not new or novel, it's just a well-known and effective technique for accomplishing various things with LLMs. Providing similar or generic examples goes a long way.
That’s not how I understand his post, instead "read-only filesystem" seems to reference the LLM working directory and not it’s outpost directory. He’s asking what’s the licences of the code sample which is fair.
It’s forbidden for Toyota to use a process protected by BID even if they only sell the result of that process (the car) and not the process itself.
I might have misunderstood the original article or themafia message though.
Even better than anything I’ve seen from Claude Code.
So I use it a lot for inspiration for screens and even have used it for proposals.