You should read the linked Scientific American article before reiterating your points. You've always been very confident that you're correct; however, I think that your treatment is simplistic and doesn't take into account a myriad of climate feedback effects, the type of thing that predictive climate models use.
Read the article and then comment.
First he estimates long ice age forcing as 6.6 watts below now. This is based entirely on changes in atmosphere composition and large surface factors (albedo, in effect). Which we can know for these remote periods from things like ice core data.
Problem - we do not remotely know that the 6.6 watts we can ID are actually all the power sources that changed between an ice age and now. To infer a 0.75C per watt sensitivity from this is not warranted. Other than triggering atmosphere and ice changes, he ascribes nothing to changed solar forcing, ice ages to now.
The physics of it say 6.6 watts can't do that. There is far too much power coursing through the system just to maintain the baseline 285-290 K temperature. A 5 C change requires a 7% power term change. If 6.6 watts were 7% of the real total, the total power would be 94 watts. Direct sunlight at the edge of space is 342 watts. This is wrong, empirically.
Physics does not say you won't get 6.6 watts from changing atmosphere and ice sheet stuff. That sounds entirely plausible. But in the old data, temperatures take off well before the atmosphere responds. This 6.6 watt measure is almost certainly the positive feedback, not the entire signal that causes ice ages.
It might be 20% of the signal, roughly split between 10% land-ice-albedo and 10% air-CO2etc-greenhouse. You'd have a 25 W driver, and these 3 W feedbacks piling on. That would have the driver 35 watts, fitting an estimate of 500 W to maintain temperature now (sunlight plus greenhouse etc).
Maybe it is as high as a third of the signal, 20 W overall, with albedo reducing sunlight etc. But that is a lower bound (it is just sunlight times a .83 albedo - the earth could have that much power with no atmosphere at all). Then the picture of ice ages would be, you get a 13 W driver from the sun, and you tack on 25% feedback sizes from albedo and atmosphere feedbacks.
Those ranges suggest climate sensitivities, reduced to one coefficient, in the range 0.25 to 0.15 C per watt. It is critical to understand that the only thing he has, up to this point in the argument, supporting the idea that instead it is 0.75 C per watt, is the notion that the 6.6 W he can ID from ice core data are the entire forcing for ice ages. This is the old, one would have thought by now discredited, idea, that ice ages are caused purely by atmosphere (and ice sheet) changes. It is just now dressed up as empiricism - since I can measure these factors 100,000 years ago, they must be the driving factors in ice ages, and the same sensitivity should be expected today.
But we also just know this isn't the real ice age story. Because these 6.6 W he can ID, show up after the warming starts. Not only is the amount too small on physical considerations, but it comes at the wrong time to set off the escalating temperature run (low level during, big only after them). Clearly the real power driver comes before them and it is just plain missing from his budget.
Why does this matter? If the same measures are used now, won't the relation be the same? No. If the other factors are follow on feedbacks to the ID-able 6.6 Ws, they might be the same now as then. But if they are solar, they aren't feedbacks at all, aren't driven but driving, and CO2 changes today are not going to set them off.
Then - 13 to 25 W from unmeasured sources, touching off 6.6 W from measured sources. Now - less than 6.6 W from the same measured sources, and no reason to believe the bigger unmeasured one is operating at all.
So, first understand, there is good physical reason to believe the power needed to drive ice ages is higher than his budgeted 6.6 W, more like 20-35W. Which imply climate sensitivities on the order of 0.15 to 0.25 C per W, not 0.75.
Now turn to the modern data. He gives his estimate of forcing factors from roughly 1850 to 2000. He gets 1.6 plus or minus 1W. If you look at the individual error bars of the internals, then going half way high yields 2.9W, and all at their 1 SD lines gives 4.2W. Thus -
His centerpoint - 1.6 W
Within his error bar - 2.6 W
Within 1/2 on all - 2.9 W
Within 1 on all - 4.2 W
Why does this range matter? Because we have an observed temperature change to set against the forcing estimates. It is 0.75C - although some say it is more like 0.6C, the former is the figure he gives. Well, that means the implied sensitivity coefficients run from 0.47, to 0.29, to 0.26, to 0.18 down to 0.14 if you use 0.6C as well. But none of them support 0.75. 0.15 to 0.25 is physically believable territory, not needing large unknown power terms. I'd have no difficulty allow small ones, bumping the range up to 0.3 to 0.375. My range estimate, certainly wide, would be 0.15 to 0.375.
But his own data for this modern period do not support the 0.75 he got from the ice ages - his own centerpoint of 1.6W forcing gives 0.47 on his high estimate of the temperature change. Yet all the later model run comparisons use 0.75. Which is twice to five times what makes physical sense and lies within his own error bars.
Understand just what that +/- 1W contains. It means a range from 0.6 to 2.6 - a factor of 4.33 times. We've also got maybe a 0.25 range factor in the temperature measurement (0.6C to 0.75 C). Put them together, and the estimate of the climate sensitivity is varying over a factor of 5, within the error bars. With only the low end of it compatible with physical considerations (absent giant unseen and unnamed additional power terms).
When we come to projected future changes he sees 1.5 W additional forcing. He tosses in an extra 0.5 W for no strong reason, about an unknown and to get closer to IPCC. Call it 1.5 to 2.0 W. That means 0.47 (his own data) say 0.7C to 0.95C, while his ice age induced 0.75 gives 1.1 to 1.5C.
My range and the same forcing predicts 0.23C to 0.75C. Physical considerations alone say 0.23C to 0.5C - the higher possibility reflects an implicit allowance for as yet unseen feedbacks, but smaller ones than the signal. For his 0.85W forcing optimistic emissions controlled case, I'd give 0.13 to 0.32C with 0.13 to .21 most likely.
Personally I doubt forcing will be that small and do not see reason to seek it with draconian regulation. He thinks for ice sheet and sea level management reasons the temperature change from today should be held to under 1C. But his own estimate of 1850 to now forcing would do that (0.47 sensitivity, with 1.5 to 2 W future forcing). I think the sensivity is half of that and a third of his ice age figure. And I'm well within his error bars for 1850 to 2000 believing it. I'm also compatible with the power needed to maintain a higher temperature.
The epicycles are falling away, but he is still clinging to the ice age theory to justify the 0.75 sensitivity figure. Which is physical nonsense if meant to describe all the power sources needed. Ice age nonsense in that it acscribes the entire change to factors we know lag the actual temperature changes in ice ages.
There is simply no good reason to think the temperature response is so high, and plenty of reason to think it is appreciably lower. His own 1850 to now estimates give a figure 37% lower; his error bars there are huge and accomodate much lower still; physics alone says 65-80% lower.