Espressif chips are not "hardened" to the extent that you could use them, e.g. in a bank, medical or defense environment (to my knowledge anyway). As a consequence, a dedicated adversary would likely always be able to extract your firmware, possibly by means of an electron microscope, laser attacks and tungsten probes sniffing the processor in operation, etc. etc.
But a number of differentiated tools are provided that allow for solid protection for a number of different use cases. For most practical scenarios, the tools provided in the IDF are able to prevent firmware extraction. These tools necessarily make support and diagnostics more difficult, so there is a tradeoff to using them you need to understand.
I guess the short answer could be: the built in mechanisms should prevent a knowledgeable adversary like an ee graduate student with access to typical electronic lab equipment like an oscilloscope from extracting a usable copy of the firmware. Of course they could still come up with a clever side channel. I couldn't estimate how difficult it would be for a specialized lab, e.g. a hardware security research group, the FBI or if the folks at Intel or NXP etc. were really interested in your firmware, but I assume probably not very given the funds.
For a more nuanced answer, I'd start by exploring the different firmware protection mechanisms available. A good place to start would be reading about fuses e.g. here:
https://docs.espressif.com/projects/esp ... efuse.html another topic could be flash encryption, e.g. describe here:
https://docs.espressif.com/projects/esp ... ption.html
Firmware protection consists of more than just setting it to "really protected". You'll be able to get a more qualified answer once you understand the different aspects and can ask more specific questions.