Are those goals noble? This is the same guy who also said "with AI we are summoning the demon" and whose self-justification for getting a trillion dollar Tesla bonus deal involved the phrase "robot army"?
If you build a system explicitly designed to have no content boundaries, and it produces CSAM, that's not a failure of execution - that's the system working as designed. You don't get credit for noble intentions when the outcome was entirely foreseeable.
Deciding to place no limits on what an AI will generate is itself a value judgment. It's choosing to enable every possible use, including the worst ones. That's not principled neutrality; it's moral abdication dressed up as libertarianism.
When I was young it was considered improper to scrape too much data from the web. We would even set delays between requests, so as not to be rude internet citizens.
Now, it's considered noble to scrape all the world's data as fast as possible, without permission, and without any thought as to the legality the material, and then feed that data into your machine (without which the machine could not function) and use it to enrich yourself, while removing our ability to trust that an image was created by a human in some way (an ability that we have all possessed for hundreds of thousands of years -- from cave painting to creative coding -- and which has now been permanently and irrevocably destroyed).