You lose power/weight RATIO as you get larger no matter what you do. Weight is proportional to volume, which is a cubed figure; strength is proportional to cross section of bone and muscle, which is a squared figure. Double your dimensions, and that factor of two gets cubed for volume and weight and only squared for cross section and strength, you'll be eight times heavier and only four times stronger; you'll have cut your power/weight ratio exactly in half.
Obviously, you can only halve your power/weight ratio so many times and still stand up and walk and since muscle tissue is basically the same for vertibrate animals, we can compute a rough limit for the world from what we know about weight lifting sports.
The strongest human athletes are the top unlimited weight category power-lifters who compete in the World's Strongest Man competitions which you've seen. Take Benedikt Magnusson for example, who holds the world's record for the deadlift.
Magnusson weighs around 380 and the record lift was 1015 lbs. The deadlift pretty much uses every muscle in the athlete's body to a maximal extent or at least comes closer to that than any other exercise.
For all lifting events, you compensate for the effect of the square/cube problem by dividing through by 2/3 power of the athletes' weights, i.e. that lets you compare the lifts of the champions of the various weight divisions. When you divide the championship numbers for a particular event by 2/3 power of the competitors' weights, the numbers almost line up and become the same number; one will stand out a bit from the rest and that guy has basically done the best pound-for-pound lift overall. This works, at least up to the point of the super heavyweight division, because the athletes are all roughly built along the same lines.
However for the thought experiment of scaling the SAME athlete to different sizes, this isometric scaling is perfect since the symmetry at different sizes is perfect. The idea is to answer this question: at what point in size does it become the same effort for Magnusson to simply stand up and lift his OWN weight, as it is for that 1015-lb lift at his normal size of 380. On the left side of the equation you want the 1015 for the bar plus the 380 for Magnusson divided by the 2/3 power of 380, and on the right side of the equation you want x/(2/3 power of x), i.e. the guy just lifting his own weight:
1395/380^.67 = cube root of x
x = 17,718 lbs
I.e. at around 18,000 lbs, it would be everything in the world Magnusson could do just to stand up.
If you put a large sauropod dinosaur next to Magnusson, you're looking at one animal at the top of the food chain and the other near the bottom. Magnusson's body is mainly muscle and that terrifyingly well trained; the sauropod's body is mostly gut and digestive system for processing leaves and grass. If Magnusson couldn't stand and walk at 20,000 lbs, the sauropod sure as hell couldn't. Magnusson is very much stronger than any possible quadruped herbivore his size. The only thing any 400-lb quadruped herbivore could do with a 1000-lb weight is be crushed by it. The first quadruped herbivore which could do anything at all with such a weight other than be crushed would be an elephant.
They're finding dinosaurs now which were 150' long or thereabouts. Some (brachiosaurids) held their necks upwards, others (diplodocids) held their necks outwards. The seismosaur was one of the kinds which held his neck outwards, and that neck would have been 40' - 60' long and could easily have weighed 40,000 lbs. If the center of gravity of that neck was even 10' from the shoulders, you'd be looking at trying to hold 400,000 foot pounds of torque with flesh and blood on a 24/7/365 basis in our gravity.
In real life, the only thing there is on this planet which figures in the ballpark of a half million to a million foot pounds of torque would be the combined max torque of all of the engines of one of our largest ships. A seismosaur in our gravity would be trying to hold that much torque with flesh and blood on a 24/7/365 basis.
The problem for the other kind of sauropod (brachiosaurids) which held their necks upwards is that the heart it would take to get blood to their heads would not fit in their bodies, that problem is well known.
There are similar problems with the 500 - 1000-lb flying creatures of past ages; in our gravity, the largest birds which can take off or land are around 30 lbs.
Scientists aware of these problems keep trying to lowball the weight estimates for sauropods. Christopher McGowen "Dinosaurs, Spitfires, and Sea Dragons", claimed a volumetrically derived weight of 180 tons for the ultrasaur and despite the grief he caught for that I suspect that number is reasonable. If you assume an equal level of effort to stand for the ultrasaur in his gravity and the largest possible elephant (16,000 lbs) to stand in our gravity and solve, you get a necessary attenuaution in gravity of about 2.8 - 1 for the ultrasaur to function.
I don't see how anybody could look at that and want to claim that gravity was some sort of a geometrical thing, particularly when we're now getting rqadiocarbon dates of 20K - 40K years for dinosaur remains, finding soft tissue in dinosaur remains, and finding accurate depictions of known dinosaur types in native American petroglyphs, i.e. since there is no longer any respectable way to claim that the Earth has had 65M years to triple its mass.
Best current theory on what gravity actually is, is that of Ralph Sansbury who describes gravity as an electrostatic dipole phenomenon. Starting from that, one can believe that a change in the surface charge of the planet would produce a change in gravity.