But that's the point. Coase says that transaction costs create the need for firms, because they make it more expensive to engage contractors for every job. You reduce transaction costs, and you consequently lessen the need for firms in certain areas.
Lessen, but not eliminate. Even the best approaches currently imaginable for hiring freelancers have rather hefty transaction costs - in the economic sense, not as fees.
Finding, validating and qualifying a good freelancer is a serious transaction cost; Risk of poor delivery for the buyer and risk of nonpayment/partial payment for the worker is higher than in the employment scenario; those risks are serious transaction costs.
As you say, it can lessen the need for firms (more exactly, make the optimal firm size smaller) - but as the parent states, it can't eliminate the need for firms, there will remain an area where firm-like cooperation is optimal, better than freelancing.
