2) You're right that you almost never want to produce at a loss based on true market prices and, for that reason, auto-supply's current behavior can be seen as just a layer of tedious micromanagement. As you've mentioned, though, current market prices don't represent true values due to stockpiles, so there will be cases where producing at a loss makes sense. Producing at a loss when you have a monopoly (or near-monopoly) on a raw resource and a stockpile of it is the clearest example, as you'd want to keep the raw resource price high to prevent others from competing in those markets and you'd actually be making money from the manufacturing. Perhaps making auto-supply "smart" by default would be correct, with an extra button allowing you to tell buildings to continue to produce at a loss, though I do wonder if there might be technical drawbacks from this approach.
Another thought about the monopoly or near-monopoly mentioned in this quote and the other situations involving withholding resources from the market:
Unless someone can point out others, I believe the list of fundamental resources that don't require an input resource (other than initial cost of construction for the production building of course - and the opportunity cost) in-game is limited to these: power, water, iron, silicon, aluminum, and carbon. I can manufacture power, water, and carbon via constructing my own power plants and building ice/dry ice condensers or getting my carbon from the air. Your firm control of the limited number of resource deposits will possibly give you competitive advantage in these markets, but not a monopoly.
Regarding iron, silicon, and aluminum, let's assume there is a single high deposit of iron on the map. You control it and, as taught by socialists and according to the FTC, you have a monopoly with which you can drive up prices to seemingly limitless levels and keep your outrageous profit margins on the iron and steel markets. In-game, just as in real life, there are always alternatives to obtaining or replacing the monopolized resource. Once the map clears I can click 'Toggle' and compete in the race to claim the limited resource tile. Being slow, I probably wouldn't win. Using my time for other purposes I would, however, probably have enough money and time to perform - and win- the following race (any money spent or debt accrued would be evaluated against the benefits of my winning):
Build an Expansive HQ (thus reducing my need for your steel), spend $300 for steel to immediately upgrade my HQ to Level 2, Spend $3,400 on glass for a Patent Lab, spend $7,400 on chemicals and race for the slant drilling patent.
Unless you plan on using debt to somehow corner the markets for steel, glass, and chemicals immediately, I will create competition to the iron market via slant drilling, and use auto-selling consistently at the level that will bring me the maximum profit over time, not shortsighted unsustainable sums.
Also if you decided to glut the market with iron instead of using it in your steel mills and the resulting price reached $1, I would just buy it below my costs of producing it to give me a competitive advantage against you in the steel market.
If I were to lose the race and you took the slant drilling patent as well, I would simply have to use a hacker array to buy up the item at a reduced price, reducing demand for your item at your price in the process, or wait for a new resource discovery tile of iron to bid on, and overall just give you a very bad day to have to micromanage during all that time freed up by not clicking iron's sell button during its accumulation. I suppose I could also dynamite your iron mine if all else fails.
At the risk of getting kicked to the General Discussion Forum, I will mention this: As far as alternatives to resources in real-life, a nice example is when employers' labor costs get too high because the minimum wage jumped to $15/hour. At that point I would recommend reading about the rapid pace of advance being made toward inexpensive robotics and automation processes for insight as to a cheaper alternative, and how higher labor costs will only expedite the process. Also, setting prices above market level will create a surplus - in this case labor - with no way to correct by decreasing the price. Compare this to the trope that people earning more money will have more disposable income to spend, thus driving up demand and creating more jobs. The market reality only allows one correct answer.
I also haven't had time to thoroughly analyze this, but I believe, given my limited experience with playing the game on snail and pausing, that in general selling consistently to meet demand over a given time period, with profits you can keep track of with the aid of auto-selling, results in a more reliable and higher level of profits than hoarding the resource and attempting to sell at inflated prices. It would take control of variables over a relatively long period of time to determine this.