Doubling rate in confirmed cases is closer to 4 days than 3 unless you include only the last week or so. A doubling rate of 6 days in the medium term has also been estimated. This will obviously delay the time to reach one million cases. However, confirmed cases are a lower bound for the total number of cases, as cases only get confirmed when evaluated by healthcare personnel. It is not credible that Iran and Italy have 3000 cases each, when lots of tourists returning from there have been diagnosed. Did they just randomly bump into one of the 3000 in a country of millions?

Uncertainties aside, the model of exponential growth is sound. Disease spread follows exponential growth until it hits an inflection point, either due to running out of viable hosts or due to quarantines and other countermeasures reducing the transmission rate per infected below 1. China forcibly caused this inflection on February 4th, by essentially stopping their economy through strict quarantines.

A doubling rate of 4 days yields 1 million infected globally on the 29th of March, and that's what we'll get unless the transmission rate somehow falls before then. Plenty of assumptions in that statement though.

I honestly have no idea! Can you link to his analysis, I haven't seen it.

Regardless, I'm no expert on this matter, and it would be nice if people stated this more often than defer to which ever thing they heard as being correct, which I think is what people are doing when they say it's 'just a flu'.

Ideally we should prepare for the worst of possible outcomes, hope for the best of possible outcomes, and expect the actual outcome to be in the middle. Those possible outcomes change as more information gets released.

Can you refer to his model? Is it just that the number of cases doubles in 3 days?

I think it is an overestimate but not more than by 1 order of magnitude.