Yes, they said on Lemmy. I solidly understand what it is to be in many tiny, insulated ideological minorities from decades of experience, namely that my experience is often profoundly different from most peoples’. Corporations are obviously inducing demand for large AI models, but I do speak to many people outside my anti-LLM bubble and rub up against the reality that people fucking love to use LLMs to (as an example from the OP) summarize everything for them. So many people I speak to are so fucking brainrotted from using LLMs as a crutch that they’ll whine that like 300 words, for a thing they asked to have explained for them, is too long and needs a summary.
Chrome is literally worse than stock Firefox. You seem pretty out of touch with what we’re talking about rn.
Yeah, I’m not going to waste my time with this one.
Also non-generative translation has existed for years just fine, just saying lol.
I have no idea what the hell you mean by “non-generative”. Do you even know whay that word means? I see now that one was introduced in late 2023 and I was unaware of it. But what makes this one “generative” and that one “non-generative”? It’s all tokens to tokens, regardless of if it’s an LSTM (I don’t know if that’s what that one used) or a GPT (which I vaguely assume Firefox uses now, and is a generative pretrained transformer, but that’s not really the job it’s doing here; maybe a BERT? Maybe an LSTM still?)
deleted by creator
Yes, they said on Lemmy. I solidly understand what it is to be in many tiny, insulated ideological minorities from decades of experience, namely that my experience is often profoundly different from most peoples’. Corporations are obviously inducing demand for large AI models, but I do speak to many people outside my anti-LLM bubble and rub up against the reality that people fucking love to use LLMs to (as an example from the OP) summarize everything for them. So many people I speak to are so fucking brainrotted from using LLMs as a crutch that they’ll whine that like 300 words, for a thing they asked to have explained for them, is too long and needs a summary.
Yeah, I’m not going to waste my time with this one.
I have no idea what the hell you mean by “non-generative”. Do you even know whay that word means? I see now that one was introduced in late 2023 and I was unaware of it. But what makes this one “generative” and that one “non-generative”? It’s all tokens to tokens, regardless of if it’s an LSTM (I don’t know if that’s what that one used) or a GPT (which I vaguely assume Firefox uses now, and is a generative pretrained transformer, but that’s not really the job it’s doing here; maybe a BERT? Maybe an LSTM still?)