kapkap

The big corps that provide the llm obviously train them with the objective to make the user addict, it's their business model. If the result was not a galaxy of slop it would be less worrying than it is though.

You probably know about this solution for your tracker, but in case you don't, a pretty much straightforward userscript (in javascript) run with greasemonkey (or similar) should be able to easily count your clicks & submit, save them in indexedb (browser database) and display them in a layer over linkedin pages.

The hero thing, what you can do for the user is probably what most want to know first when they land somewhere. On the other hand, the slop epidemic is gaining traction. People seem to start to realize llm don't think & just predict, and therefore make wrong prediction, flooding the world at every error with more slop (that is then used to train more llm, so expect the prediction success rate...

I am going the full private way. For authentication : To create an account, a couple of RSA keys are created in the browser, one to sign/verify, one to wrap/unwrap (encrypt/decrypt can happen through hybrid encryption with wrapped AES keys). All user data are saved in indexed-db (browser database) and exportable as file. User can encrypt both with a password (the keys are encrypted with...

In my opinion, ads don't have to be invasive, especially in the context of a saas : they can be easily integrated as sponsored messages, as long as they are 100% related with the object of the app (and therefore aligned with the interests of the user). That would require additional work to find those sponsors with activities related to the app, but you would also sell the impression for a much...