I'm not necessarily saying "TSLA Will definitely run out of money in 3-quarters". I'm saying "TSLA is currently, on a linear-approximation track to run out of money in 3-quarters".
Obviously, the Model 3 rampup will improve profitability. But the Model 3 Ramp-up also requires more resources (more humans in the assembly line. More robots in the assembly line. etc. etc.). So that still requires money.
A linear approximation is basically a quick-but-shitty-approximation. Its really easy to calculate and full of flaws, but full of flaws in a way that almost everyone can immediately understand.
Exponential approximations (ie: Musk's approach) might be more realistic... since production lines can exponentially grow to some degree. But if you are off by even a little bit, then you end up being ... oh I dunno... 8-months to 10 months late with your 5000-cars per-week estimate.