Vishal Bhargava

We're rolling out a feature that lets you vibecode Excel models, the same category an a16z-funded startup has been getting attention for. Building it out, I realized we don't just match what's currently considered best. We beat it.

Here's our output for the prompt "build a detailed LBO model for ExxonMobil": https://docs.google.com/spreadsh...

Run that same prompt on any other platform and the gap is hard to miss.

How we did it: we generate Excel server side instead of client side, using openpyxl as the engine instead of SpreadJS. We built on top of openpyxl to support pivot tables and the rest of what a real model needs. The reasoning behind the bet is that openpyxl is heavily represented in training data, so the model actually knows how to drive it. That's the difference between a spreadsheet that looks right and one that is right.

Hey Product Hunt

We built Associum for professionals and firms who are tired of AI tools that look impressive but stop short of actually doing the work.

Here is what Associum does:

There are established players in AI slide generation and AI Excel generation. There are almost none that generate long-form reports. That gap isn't an oversight, and it isn't accidental. It's because the problem is genuinely, structurally harder. Here's why.

The slide problem vs. the report problem

AI-generated presentations work because the format is constrained. A slide has a title, a few bullet points, maybe a chart. You know roughly how much content goes on each slide before you start. The structure is fixed, the length per unit is bounded, and the output is predictable.

Long-form reports don't work that way. A section on "competitive landscape" might be two sentences in one report and four pages in another, depending entirely on what the research finds. There's no fixed length. There's no fixed structure. The content determines the form, not the other way around. That single difference makes the engineering problem an order of magnitude harder.