That's possibly plausible, but it's a lot more complicated a claim than just those industries being bad. To take the example of tobacco, you'd need to quantify the enjoyment smokers get from it, the harm to their lives, the harm to other's lives, the benefit to everyone in the supply chain (from farm workers to farmers to advertisers to executives), the harm to everyone in the supply chain (especially the exploitation of those who are do the actually picking of the tobacco and assembly of the cigarettes), all the politicians who get benefits from, all the harm they cause, and so much more.
It's also the wrong question to ask. It's asking if sum(negatives) > sum(positives). But not all those harms and benefits should be equally important under any rational more framework. For example, I don't think any reasonable person would count the benefit corrupt politicians get from tobacco companies slush funds to be an argument in favor of keeping tobacco legal.
In other words determining if what you said is true is extremely difficult and extremely pointless.
sum(negatives) > sum(positives) is perfectly legitimate -- it's not extremely pointless, nor is it particularly difficult to assess. It's (quite literally) a "pros vs. cons" analysis. It can be complex to quantify, but we do so all the time through modeling. The function needn't be linear in the parameters, and many parameters can intangible (e.g. societal, moral, etc) that are subject to jurisdiction, norms, culture, etc.
In fact, we do similar comparisons all the time: Annual employee performance reviews, asset appraisals (stock, realestate, bonds, etc), and internal rate of return (IRR) or risk-adjusted rate of return. The models are just models -- they can easily suffer from incomplete data, history, sentiment, and bias, intangibles, sentiment -- but the world makes the decision based on these models all day, every day.
You're right that assessing a model like that is perfectly legitimate, but all the models you mention are much more limited and well-understood than multiple sectors of the economy. The original post in this chain claimed "a good chunk" of the modern economy was net-negative. I don't believe we make similarly hard to quantify calculations every day.
It's also important to note that all the examples you give have a lot more data supporting them than any of these internet comments. If the original poster actually built a model of a "good chunk" of the economy and proved it was net-negative and zero-sum I wouldn't have been dismissive.
It's also the wrong question to ask. It's asking if sum(negatives) > sum(positives). But not all those harms and benefits should be equally important under any rational more framework. For example, I don't think any reasonable person would count the benefit corrupt politicians get from tobacco companies slush funds to be an argument in favor of keeping tobacco legal.
In other words determining if what you said is true is extremely difficult and extremely pointless.