You have to tell that to a number of companies where I have supplied high-security devices. Some who have been sent through specific certification processes because of the humans afety requirements.

And yes - several of them are using Atmel AVR or ARM chips. So despite not being 8051 chips with the 8051 legacy of a weak pull-up, they still use active low signals to activate relays our outputs.

Only a processor with specific programmable "fuses" would be able to descide state of a GPIO pin directly on power up, while the processor is still in the reset state.

If a processor pin is floating (high-impedance input or with a weak internal pull-up or pull-down), then you can use an external resistor to force the signal to VCC or GND. That would result in the signal holding whatever state you want until the application gets into control, and forces the pin to a given state.

If designing hardware with a processor that can not force a pin high, then it's 100% the hw designers task to make sure that the external logic is designed in a way that this doesn't result in any problems. In this case, you seem to have gone for a hw design that does not fulfill this requirement. It does not represent an error with the processor, but an error in the understanding of the requirements. The more you show your anger, the more you tell the world how you failed to understand the used hardware components. So all the blame is actually reflecting back onto you. That should be a good reason why you should consider stopping your blaming of the processor and instead starts to do the smart thing.

8051 chips exists in fire alarms, pacemakers, strong radiation cannons, ... What important product do you work with that have a higher security requirement than that?