@timdayipper@JavierBlas China didn’t reduce imports until the conflict began. Jan and Feb and last year are irrelevant because China had full imports over that period. What is your point?
@TukiFromKL Why do you need to exaggerate something is impressive without exaggeration? $20/month will get you roughly 30 minutes of recursive coding with opus
@Austen Let’s talk about “incentives”. Both of the private frontier companies are in a foot race to ipo in 2026. If they are truly 6-12 months away from agi/obsoleting white collar, why cash out?
I actually built my own app for this called EBITDAI. It's powered by Claude Opus 4.6 and Gemini 3, trained on my own FP&A modeling approach. It uses a more structured build process, which is why it handles balance sheets and statement linkages more reliably than generic AI in Excel. I'll send you a DM.
The speed is real but the quality gap matters. Generic AI gets you 70% of the way. The last 30% (proper debt schedules, working capital logic, circular references) is where most models break. We solved this in EBITDAI by baking the statement linkages into the model architecture. Speed plus accuracy.
Balance sheet balancing is the #1 issue with generic AI in Excel. They don't understand the accounting logic natively. We saw a massive improvement with EBITDAI when we built the retained earnings and BS check logic into the framework itself. Goes from broken to perfectly tied every time.
Have to respectfully disagree here. I build 3-statement FP&A models in Excel with AI daily - full income statement, balance sheet, cash flow with working capital dynamics. The trick is structured prompting and knowing what a good model looks like before you start. The tools have gotten significantly better in the last 6 months.
@GptMaestro Exactly right. A year from now nobody will be impressed by a DCF in Excel. The bar is full operating models - revenue builds, unit economics, integrated debt schedules. That's the real benchmark.