In this post and your linked post you come off as quite confrontational while exaggerating the dissent and underplaying the support. It’s a great way to get comments that strongly disagree with you. In all of that commotion, it sounds like you’ve missed the point: You’re regretting the AI disclosure, when really people were voicing their opinions in hopes that you’d regret/rethink the AI use.
Personally, I really like the AI disclosures. I like seeing how the sausage is made and what it’s made with. If the sausage is stuffed with hot dog “meat” I want to know. I’ll also avoid it like I avoid junk food, because that’s what it is. LLM-generated code is the junk food of programming. It’s a waste of resources, it’s bad for your (and everyone’s) health, and it can’t be trusted.
Some people really like junk food and hot dogs, and that’s fine. We just need to be clear that that’s not healthy and encourage healthier habits like eating fruits and vegetables (or at least real meat). People like ignoring the externalities of junk food and LLMs; both are responsible for a lot of pollution and societal harm (increased healthcare costs, dark patterns, misinformation, etc.). That’s enough reason for some people to completely remove them from their lives and be understandably vocal about how others should too.
The occasional sweet treat isn’t bad, but making sweet treats at home is always better. At home, I have more control over what ingredients I use and where I buy them from. Though I know I’ll never be able to use zero resources to make a good cinnamon roll, I can at least make the most optimal choices for myself. Ask yourself this: what value does text I generate provide? Could someone else, given the same information, generate the same text? Could someone else write functionality similar text without boiling the oceans and massive copyright infringement? I think any LLM-generated content that anyone could reasonably generate and use themselves is worthless and a net negative impact on the world. If it’s LLM-generated content that can’t be immediately/easily used, then you’re at least providing some real value by doing some work to make it usable.
AI disclosure doesn’t protect you against other people’s opinions. Some people will buy and consume the product, some people will try out the product and decide it’s not for them, some people will discover the product is made very unethically and decry it publicly in hopes that others won’t enable the ethical violations.
Meanwhile…
Anyway:
I think you’re finally getting to the realization: doing things without considering the moral or ethical implications is inherently, uhhh, immoral and unethical. Do you do more good than bad with your smartphone? Does it use thousands of watts (hours) every time you use it? Do you replace your phone every year with the new model, maximizing your environmental impact with the manufacturing? What do you think makes using a smartphone just as bad as using an LLM?
Look, I get it: considering all of the consequences of your actions all of the time is tiring. I don’t like doing it either. Clearly companies don’t do it at all (at least not for external consequences). Look where that has got us. Apathy is not an option.
Local models are better in many ways, yes. The local user has control over how much power they use (for the query, not on training), other resource usage (water, electrical power sources), what model (and by extent the training data it uses), and exactly what documentation they get out (depending on the query/ies). The best option is still obviously to write the documentation yourself, as the subject-matter expert.