I'm only saying it's useful in the innovation process. Whether I enjoy it is something else. (I do). But another thing is that I would be searching online resources for answers to my questions anyway. Is the LLM in this context different than that process?
You're wrong and it innovates nothing that hasn't already been thought of and already done. It's for lazy people who don't care about the work of gaining knowledge or the destruction of the environment. Take it back to hell wirh you.
In terms of resource usage? Yes it's at least an order of magnitude less efficient to use an LLM.
Don't disagree. But again my point is whether the costs of LLMs in this context outweigh the benefits of accelerating innovation?
Innovation isn't neutral in its impacts. What us being accelerated might itself make the world worse. Sometimes going slower gives us more time to think, assess, & mitigate.
I would say so. Especially since this is a situation of an externality* where the positive effects are concentrated and the negative effects are spread out, similar to pollution. Especially given what we know about global warming and LLM's impact on it. * en.wikipedia.org/wiki/Externa...
yes, of course it is, search results are indexed variously using more modest computing resources spread over a much longer time, and nothing about them is being recompiled in realtime for you in the usual modes of use
and of course, pragmatically, you know those aren't the same because you're choosing to do the LLM route. so comparing them is a strawman from you too. you prefer something about the human-facing appearance of these models so that's what you choose. no need to be disingenuous.
I prefer it because it saves time. I don't give a shit about whether it sounds human. In fact I find that annoying. My original point was that they can serve to foster innovation. And that has been good for society. The environmental cost does trouble me.
What specifically have they done that’s been good for society?
again, that's a pragmatic reason that you choose LLM. so asking someone to justify it to you using a hypothetical where you use google instead is dishonest in both directions.
Not asking for justification but whatever. Thanks for your thoughts.
you're lying again. "Is the LLM in this context different than that process?" yes, it resoundingly is. i'm answering you. that's what you asked and what i answered.
and i'll just point out that you recombining existing knowledge about existing things is not innovation. dr. t is an expert on aspects of this subject, and i myself am a journalist who talks with actual innovators in science who explain why that's the case. something with nothing new can't make new.
It’s a glorified remix machine.
Disagree Innovating is in the concept, not how it is done. I describe a set of requirements and i get input in how I might implement it. There is nothing different in this process because I'm using an LLM. It's just faster.
"innovating is in the concept" lol okay. you're dreaming now, have a good one
if you aren't in the group of people whose discoveries contribute to new knowledge, if you instead are someone who's learning on your own to do things whose answers already exist, then you aren't an innovator and that isn't innovation. you are learning or something. not innovating.
Innovating with the infinite automated plagiarism machine is an oxymoron.