mikestaub

They have made genuine progress on 'reading' the mind of the model, but 'writing' perfectly is not mathematically possible due to the gauge freedom in the residual stream.

In your plan is an implicit assumption that those that cannot work do not deserve to have housing and food. This is a political choice made by a society.

I think a more interesting experiment would be to train an LLM on a dataset that had zero references to "first person POV" sentences. And then see if the final model somehow develops that concept.

Very cool!

Seems like a way to justify stock valuations of OpenAI while picking up some good talent to go after vertical stack as the foundation models become commodified. Smart play.

If they are Delaware C corporations and you own at least one share, you have a legal right to demand access to the cap table.

Language in humans is a VERY strange behavior. I think the stoned ape theory may have some credence, as mushrooms do seem to excite linguistic capabilities.

It's impossible to enforce a law like this. The real solution is to create personal AI filters for each person that reads all the data for the user, filters out anything that doesn't enrich the user, and hides the rest. A general-purpose AI-enabled spam filter for your entire digital life.

This is because of the way most software is built. If you are using event modeling, it's very easy to estimate feature costs. The only exception is when you are truly innovating new algorithms so there is not a historical precedent you can refer to. This is not common work in the industry however, most software projects are variations of previous themes.

The problem is that eventually, AI will get so powerful that it will block all ads on the client without anyway of detecting it. Adblockers today are very basic and have no understanding of the mind that is consuming the content. We must adapt and create new business models.

Just add a premium tier or partner with coding bootcamps. Don't give up so easily.