Dan R., a loyal MR reader, poses this question:
I would be curious to know what you consider the biggest flaw in the labor theory of value to be. Also, would you say that it is disproven, unnecessarily bulky, or simply marginalized?
There is a simple model in which the labor theory of value is true. If inputs are homogeneous and constant returns to scale hold, the proportions of labor input will indeed be proportional to price. If not, labor inputs will be reallocated until this proportionality holds (Much ink has been spilled on whether this is what Smith, Ricardo, and others had in mind; it is one way of reading Smith's deer-beaver-hunting example.)
One problem is that we need labor, capital, and land for production, not just labor. The so-called "transformation problem" tries to square this circle. The simplest response, however, is to give up the labor theory of value.
Another problem is that inputs are heterogenous. They have to be valued in dollar terms, and that requires imputation, a'la Friedrich Wieser, and that in turn requires information from the demand side. Price determines cost of production at least as much as cost of production determines price.
Compared to Marshallian supply and demand scissors, the labor theory of value is at best awkward and most of the time it is wrong. There are some economic sectors where constant returns to scale hold and thus demand has little influence over market price. But those are special cases, even if some Cambridge-U.K. linked economists promote them as the main show.
Addendum: David Henderson comments.