Taleb only clearly pointed out that if you use pricing formulae based on the assumption of a time-inavariant Gaussian distribution of returns, all small gains

Taleb only clearly pointed out that if you use pricing formulae based on the assumption of a time-inavariant Gaussian distribution of returns, all small gains you accumulate in the good states will be wiped out and more by losses you will suffer in the bad states. But this is because you have underestimated the downside risk. If you used a havy-tailed distribution (e.g. something from the stable family, which was suggested by Mandelbrot (1963) and Fama (1965)) and/or allowed for time-variability of moments and/or accounted for systemic risk you would see different prices of assets that better reflect that downside risk.