I think we may have different definitions of a kettle. I mean something like this:
Which you put on the stove. I can’t imagine that having tea in this is a problem at all. It’s just glass.
I’ve also done this with something like:
Which I could imagine keeping more of the taste/being a problem.
I assume you mean something like this by a kettle?:
Apparently I’m committing all the tea sins. I definitely make tea in a kettle. But if I do that, I boil the water before adding the tea bags. Isn’t that pretty standard? I’d only do so if I’m making a lot of the same tea (or iced tea), usually for a group of people
I often do this. With loose leaf tea too. The quality of the result highly depends on the tea and whether you get the timing right. I know my microwave pretty well and can hit boiling or just before boiling by changing the time for a black vs a green tea.
When boiled appropriately, I can’t really tell the difference for most bagged teas, so maybe I’m just tea uncultured?
The earl grey loose leaf I have I actually like better when it’s kept boiling for longer (about 15 seconds of boiling), and the microwave allows me to easily do this.
The loose green tea I have changes its flavor a lot when heated for different amounts and to different temperatures. The microwave also let’s me easily control this in a way that I would struggle to with a kettle. I suppose I could add the tea afterwards and just get the water a bit hotter to compensate, but I’m lazy and I always forget about my tea in the microwave so it’s easier if it already has the leaves in it so I don’t have to re-steep
It definitely depends on the microwave. In my office, one boils a mug in 2:20, while the other requires over 3 min
This is a really fantastic explanation of the issue!
It’s more like improv comedy with an extremely adaptable comic than a conversation with a real person.
One of the things that I’ve noticed is that the training/finetuning that’s done in order to make it give good completions to the “helpful ai conversation scenario” is that it flattens a lot of the capabilities of the underlying language model for really interesting and specific completions. I remember playing around with gpt2 in it’s native text completion mode, and even with that much weaker model, it was able to complete a much larger variety of text styles without sliding into the sameness and slickness of the current chat model fine-tuning.
A lot of the research that I read on LLMs is using them in the original token completion context, but pretty much the only way people interact with them is through a thick layer of ai chatbot improv. As an example for code, I imagine that one would have more success using an LLM to edit your code if the context that you give it starts out written like it is a review of a pull request for the code, or some other commentary of a form that matches the way that code is reviewed in the training data. But instead of having access to create that context directly, we have to ask for code review through the fogged window of a chat between an AI assistant and a person discussing code. And that form of chat likely isn’t well represented in the training data.