Whatever we do, we must be careful.
I am uncomfortable with the Portuguese here.
First why is seja used instead of sejamos?
Is the English translation really correct? The Google translation is “We need to be careful what we do” which isn’t really a sentence.
If I use google translate from the given English the result is
Seja o que for que façamos, devemos ter cuidado.
I only have a passing knowledge of Portuguese, but I can tell that the corresponding expression in Spanish is “lo que sea”, which means “whatever” or literally “what it may be”.
The sentence seems not to make much sense, but this would be the translation if I would translate the english sentence “We need to be careful about what we do” back to portuguese: “Nós precisamos ser cuidadosos com o que fazemos”. “Nós” in this sentence is optional and could be omitted as the pronoun is implied in the verb conjugation, meaning that “Precisamos ser cuidadosos com o que fazemos” is also correct.
Well, Precisamos ser cuidadosos para fazer o que sejamos would translate as “We need to be careful (in order) to do whatever we are.” That doesn’t make much sense coming from a native speaker. Google’s translation of the original sentence isn’t all that great, it actually means, “We need to be careful (in order) to do whatever (it is)/ anything.” The English adaptation Clozemaster offers is clearer, although translations that mirror syntax between languages generally best for teaching.
Now, the sentence Google spat out for you from Whatever we do, we must be careful is also dubious. First, in Portuguese, for is mostly used as a form of ir. While some speakers borrow English for, it doesn’t seem very common. Second, for doesn’t even make sense here under either explanation. Third, for que, even as a typo of por que or para que is nonsensical in this context. A better translation would be [Seja] o que seja que façamos, devemos ter cuidado.
Google’s translation software is a large language model, and there’s evidence suggesting that this kind of software will become less effective as more people use it. (One reason is that feeding machine-processed data to a machine leads to model collapse.) We might be reaching that point.
Thank you for the detailed response.