Dr. Hansen from 2008:
“Given the solar constant that we have today, how large a forcing must be maintained to cause runaway global warming? Our model blows up before the oceans boil, but it suggests that perhaps runaway conditions could occur with added forcing as small as 10-20 W/m2 (Watts per square meter – a 60 watt light bulb provides 3 to 6 times more forcing per unit of area than is required to turn the Earth into a Venus)”
“There may have been times in the Earth’s history when CO2 was as high as 4000 ppm without causing a runaway greenhouse effect {the Mesozoic period – time of the dinosaurs}. But the solar irradiance was less at that time. What is different about the human-made forcing is the rapidity at which we are increasing it, on the time scale of a century or a few centuries. It does not provide enough time for negative feedbacks, such as changes in the weathering rate, to be a major factor. There is also a danger that humans could cause the release of methane hydrates, perhaps more rapidly than in some of the cases in the geologic record. In my opinion, if we burn all the coal, there is a good chance that we will initiate the runaway greenhouse effect. If we also burn the tar sands and tar shale (a.k.a. oil shale), I think it is a dead certainty.”
It was not just speculation, here is the 2012 update to the 2008 paper that those quotes discussed: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3785813/
Here is a relatively comprehensive look at runaway greenhouse modeling, although it is a bit outdated (2012 - cloud modeling is slightly better now): https://sseh.uchicago.edu/doc/Goldblatt_and_Watson_2012.pdf
You’ll note that it is not written from a catastrophizing perspective, yet it confirms the possibility of runaway warming with the release of just the CO2 in known FF deposits as expressed by Hansen. Hansen isn’t much of a catastrophizer, but I suppose it doesn’t hurt to present external validation of his work.
This is all a little overkill, though! No runaway feedback loops are needed.
All that is needed for the situation described by the earlier comment is for plant carbon fixation pathways to fail, which occurs with just a few degrees of warming.
If runaway feedback loops are within reach, then just a few degrees is obviously within closer reach. Projected warming from existing emissions using an updated ECS of 4.8C is 10C, reduced (temporarily) to 8C by aerosol masking (https://academic.oup.com/oocc/article/3/1/kgad008/7335889). Warming will average around double over landmasses (warming numbers are globally averaged and land temps are much more volatile than ocean temps). Photosynthesis efficiency for plants is already past the peak in the summer months in most places with current warming, as demonstrated by this paper (from 2007 so the climate assumptions are very out of date): https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3040.2007.01682.x
That 8-10C warming is already “in the pipeline” and as you can verify here, FF consumption is still trending upward despite increased renewable infrastructure: https://ourworldindata.org/fossil-fuels
As for burning that much FF being impossible, at current rates we are estimated to burn through known FF deposits in 30-50 years. Whether or not it happens, burning that much is certainly possible to do.