4 ms·

>I am turned off whenever there is a handy new tool for us to use and then some people come up with reasons why using the new tool is a harmful thing to do. Sa

>I am turned off whenever there is a handy new tool for us to use and then some people come up with reasons why using the new tool is a harmful thing to do. Same, but it turns out its complicated. There is something materially different learning French and using your phone to translate French. Or learning to play piano vs listening to a recording of someone playing the piano. The danger is that these tools give make you feel that you can do something when you can't. > For some tasks LLMs are very useful and for other tasks they don’t offer much help Yes, and the difference is that I don't care about bash so I let it write scripts for me, but I DO care about other languages so I generally write my own. I'm basically curating my own skillset according to value and capacity. In that way AI can be a big win for teams of approx 1 who only need pro forma help in certain areas, but deep expertise in others.

chii 3y ago
> The danger is that these tools give make you feel that you can do something when you can't. i dont think people are so stupid as to feel this way. It didn't happen with cars - you could travel hundreds of miles in less than an hour with it, where as without it, you can barely do 5. And yet people have never made the mistake that they could travel that fast themselves. LLMs as a tool is not more harmful than what could be done with existing tools today prior to the advent of LLMs. And the benefits of the LLMs should outweight the harm imho.
- voidhorse 3y ago
  > i dont think people are so stupid as to feel this way. It didn't happen with cars - you could travel hundreds of miles in less than an hour with it, where as without it, you can barely do 5. And yet people have never made the mistake that they could travel that fast themselves. But cars are in a totally different category insofar as they do one very specific thing and are bound by very obvious physical limits. Our own physical limitations are very obvious to us, our intellectual limitations, I think, are less so. Further, companies (I think knowingly) market AI (read: llm chatbots) as though it's some general, infallible solution to all intellectual and digital content problems. This is obviously not the case, and if you use a chatbot to work in a domain in which you have any expertise whatsoever, you'll quickly uncover this. The problem is people without expertise using these tools and being fooled into thinking they've plumbed the depths when they haven't tread past the shallows. Physically speaking, your capabilities are very apparent when you attempt to run 100mph and simply *cannot*. Realizing you cannot do something or don't understand something correctly intellectually is a far different scenario.

chii 3y ago

> The danger is that these tools give make you feel that you can do something when you can't. i dont think people are so stupid as to feel this way. It didn't happen with cars - you could travel hundreds of miles in less than an hour with it, where as without it, you can barely do 5. And yet people have never made the mistake that they could travel that fast themselves. LLMs as a tool is not more harmful than what could be done with existing tools today prior to the advent of LLMs. And the benefits of the LLMs should outweight the harm imho.

voidhorse 3y ago
> i dont think people are so stupid as to feel this way. It didn't happen with cars - you could travel hundreds of miles in less than an hour with it, where as without it, you can barely do 5. And yet people have never made the mistake that they could travel that fast themselves. But cars are in a totally different category insofar as they do one very specific thing and are bound by very obvious physical limits. Our own physical limitations are very obvious to us, our intellectual limitations, I think, are less so. Further, companies (I think knowingly) market AI (read: llm chatbots) as though it's some general, infallible solution to all intellectual and digital content problems. This is obviously not the case, and if you use a chatbot to work in a domain in which you have any expertise whatsoever, you'll quickly uncover this. The problem is people without expertise using these tools and being fooled into thinking they've plumbed the depths when they haven't tread past the shallows. Physically speaking, your capabilities are very apparent when you attempt to run 100mph and simply *cannot*. Realizing you cannot do something or don't understand something correctly intellectually is a far different scenario.