Beginner lifters don't know what their max potential is and might overtrain by accident (because they are naturally strong and might do more reps to try and get to failure) or might not do enough volume to trigger the musculoskeletal adaptions needed to lift more weight (because they are not naturally strong and might tap out early).
> As an example (bragging a bit), as a ~50 year old, 165lb, 5' 7" guy, I'm able to lift 100lb dumbells on flat bench chest press, one in each arm. I can hit around 10-12 reps with that, most of them BS cheat reps where I'm powering with my shoulders.
I personally never recommend cheating reps.
Bodybuilders do them to localize hypertrophy in specific regions, like the upper pecs or shoulders, to improve physique during competition. Most people aren't bodybuilding though and would not benefit from this style of training.
At best, you're short-changing the development of primary and secondary movers this way in the name of moving more weight (like not engaging the lats enough when you do a bent-elbow pull up).
At worst, the risk of injury goes way up (shoulder blowouts are super duper duper common in bench presses; doubly so for dumbbell bench presses, since there isn't a bar to stabilize your arms).
In my experience, I've found that it's better to go lower in weight and use an assistance exercise and/or increasing rest/recovery time and, if needed, decreasing volume instead of cheating reps when you're stuck.
Anecdotal example. When I was struggling with getting to 225x5x3 on the bench, I dropped back 20%, added weighted dips to my rotation and decreased bench press volume. Dips engage the pectoral and tricep muscles more directly, which contributes to forward progress on bench. I can do 250x3x6 now at around 9.5 RPE.
