The Cursor for PMs and Marketers Doesn’t Exist Yet. Here’s What We Actually Need.

There’s a moment that happens at least twice a week for me. I’m deep in a research thread, I’ve got context built up across three conversations, a scratchpad full of half-formed ideas, and a clear direction forming in my head. Then I switch tools, and it’s gone. I paste a summary into a new chat, try to reconstruct the train of thought, and spend the next twenty minutes getting a different AI back up to speed on work I’ve already done.

If you’re a product manager, marketer, or strategist who relies on AI, you’ve been here. You know exactly how this feels.

The Context Collapse Problem

The modern AI-powered knowledge worker is quietly managing something nobody warned them about. Call it context collapse: the constant erosion of shared understanding every time you move between tools.

Here’s what a typical workflow looks like right now. You start a research thread in Perplexity. You take notes in Notion. You draft a strategy doc with Claude. You refine messaging in ChatGPT. You fact-check in a different tab. At no point do these systems talk to each other. At no point does the work you did in one place carry meaningfully into the next.

Every session starts from zero. Every model gets a cold introduction to your project. Every context switch costs you time, precision, and cognitive load you simply cannot afford.

This isn’t a minor inconvenience. It’s the defining friction of doing strategy work with AI in 2025. And the cost compounds. The more complex the project, the more context you need to maintain, the more energy you spend managing the pipeline instead of doing the actual thinking.

What Engineers Already Have That We Don’t

When developers hit this same wall, someone built Cursor.

For those unfamiliar: Cursor is a code editor that keeps your entire project in context. It knows your file structure, your past decisions, your codebase conventions. You don’t re-explain the architecture every time you want to write a function. The AI understands where it is and what came before. The result is a dramatically faster, more coherent way to build software.

That environment doesn’t exist for the rest of us.

There is no equivalent workspace where a PM can walk in, pick up where they left off, and work across Claude, GPT, Perplexity, or Gemini without starting over. There is no shared project memory that carries your competitive research, your positioning hypotheses, your audience insights, and your campaign brief into every session. There is no unified interface that lets you choose your model the way a developer might choose a compiler.

The question I keep asking is a simple one: why not?

What This Tool Actually Needs to Do

The problem isn’t that AI tools are bad. They are genuinely powerful. The problem is that they are isolated. Solving this is less about building a better model and more about building a better environment around the models that already exist.

The workspace that PMs and marketers actually need would do a few specific things.

It would maintain persistent project memory, so that your research from last week is still accessible and alive when you come back to it today. It would support multiple models under one roof, so you can use the right AI for the right task without switching subscriptions or losing your thread. It would let you describe what you want to build in plain language and have it actually happen, rather than translating your intent into a format each tool can individually understand. And it would treat strategy, research, growth, and marketing work as first-class outputs, the same way Cursor treats code.

This isn’t a fantasy. The technical infrastructure to build this already exists. What’s been missing is someone who has thought seriously about the workflow problems of non-technical builders and designed around them.

Why This Matters More Than Another AI Feature

There’s a tempting shortcut that many tools are taking right now. They add an AI button to an existing product and call it an AI workspace. That isn’t what I’m describing.

The distinction is architectural. A real solution puts the AI environment first, and wraps the product experience around how knowledge workers actually think. That means understanding that a growth marketer’s context is not a single document, it’s a web of connected ideas, hypotheses, experiments, and audience insights that live across time. It means understanding that switching between models isn’t a bug in your workflow, it’s a feature you should be able to do deliberately and without friction. It means understanding that the output isn’t always a piece of content. Sometimes it’s a decision, a framework, a prioritized list of bets, a positioning statement you’ve been stress-testing for two weeks.

What 8080 Is Trying to Solve

I’ve been watching what the team at 8080 is building, and it’s the closest thing I’ve seen to an honest attempt at this problem.

The core idea behind 8080 is that you should be able to describe what you want to build in plain language and have it actually materialise as a working digital product, whether that’s a website, a dashboard, a SaaS tool, or a workflow. No code required. No stitching together five separate platforms. Just your idea, described in natural language, and a system that understands how to execute it.

What makes this relevant to the context collapse problem is the underlying architecture. When a tool is built around the idea that plain-language intent should drive outcomes, it naturally has to solve for memory, continuity, and coherence. You can’t deliver on that promise if every session forgets what came before. You can’t build something useful if the system doesn’t carry your intent forward.

The early results are genuinely interesting. People are shipping dashboards, internal tools, landing pages, and product prototypes without writing a line of code. The promise isn’t just speed, it’s coherence: your idea, maintained and built out, without the usual fragmentation.

That matters to me because it points toward something bigger. If you can build a product from a plain-language description, you can also build a working environment from one. The workspace I’ve been describing, the one with persistent memory, model flexibility, and strategy-first thinking, is a natural extension of the same core idea.

The Opportunity Is Real

I posted a version of this frustration publicly not long ago, tagging some of the sharpest product and growth thinkers I know: Lenny Rachitsky, Andrew Chen, Shreyas Doshi, John Cutler, and others. The responses I got were immediate and consistent. Almost everyone recognized the problem. Almost no one could point to a solution that fully addresses it.

That’s a signal worth paying attention to. When people who think about product and growth for a living all share the same workflow frustration, and nobody has a satisfying answer, the gap is real.

The builders who close this gap won’t do it by adding another AI feature onto an existing productivity tool. They’ll do it by rethinking the environment entirely. By treating context as infrastructure. By designing for how strategists actually work, not how developers work or how writers work, but how people who spend their days moving between research, insight, and decision actually think.

That tool is coming. And when it gets here, it will change how this kind of work gets done in the same way Cursor changed how engineers write software.

I’m watching closely. And I’m building in that direction.