I mean, this is also a particularly amateurish implementation. In more sophisticated versions you’d process the user input and check if it is doing something you don’t want them to using a second AI model, and similarly check the AI output with a third model.

This requires you to make / fine tune some models for your purposes however. I suspect this is beyond Gab AI’s skills, otherwise they’d have done some alignment on the gpt model rather than only having a system prompt for the model to ignore

So you don’t have to modify the amount when the recipe called for kosher salt but you only have sea salt. A cup of pasta? Depending on the type you end up with vastly different weight

It’s one thing to just use the software, it’s another to open bug tickets that you expect the maintainer to prioritise. It’s free software, the maintainer doesn’t have to do anything for you. If they want tickets fixed with high priority, they should work something out with the maintainer.

Is the current incarnation beatable, or was that a while ago? I’m not making any progress

You don’t look like you feel bad to me

Christ. I didn’t care too much about the gamers nexus video, but this sounds awful.

I think the main reason OOP has a well-known term and pattern for dependency injection is to differentiate these two (out of multiple) options:

• the constructor of my object creates other objects it depends on itself
• I construct the dependencies of my object elsewhere and pass them in to the constructor and use an interface to make it easy to swap behaviour

However, this becomes less of a pattern in functional programming as you wouldn’t make such objects to begin with. In FP, you pass all parameters where a function is invoked, and DI just becomes using generic parameters. You wouldn’t instantiate a dependency on each function call after all.

As this is such a minor change, it’s not really talked about much and it’s not really a pattern,