4 ms·

> Even if AI doesn't take your job, it almost certainly will take most of the value it has. I don't need a $175k dev when I can use a model and a $60k dev to do

> Even if AI doesn't take your job, it almost certainly will take most of the value it has. I don't need a $175k dev when I can use a model and a $60k dev to do the same thing. Current trends in writing & editing appear to be that you need a rare and expensive person to get the kind of whole-number multiplier out of AI tools that people initially expected. And it's still not a large whole number. Give the tools to someone who's not got exception taste and skill, and you're lucky if the multiplier's greater than 1 at all. [EDIT] Exception if your output is expected to be crap and you don't care about accuracy. It's a huge boon to spam of most forms, and astroturfing.

Workaccount2 3y ago
I think it's a mistake to use current AI in current times as the status quo in ~5 years from now. AI will likely improve dramatically, as will the general persons ability to work with it. I wouldn't be one bit surprised if typewriter operators outpaced nascent word processor users in the beginning.
- hotnfresh 3y ago
  > AI will likely improve dramatically, as will the general persons ability to work with it. I'm not so sure about ability to work with it. Middling Excel skills still make you a wizard able to do things nobody else can, in most offices (though not most tech offices). Hell, I've seen knowing how to use the Windows file explorer at a very low level of competence put a person in the top five percent by computer ability in an office. Within the last couple years. Most folks can barely use the tools we have, and aren't going to be able to get good output from anything like the current set of AI tools. The ones who can, won't be cheap, and may be hard to replace like-for-like—I think a lot of companies are going to settle on more, cheaper, more-interchangeable workers than replacing them with a few AI tool experts with hard-to-measure-or-quantify skills. The tools themselves need an added ability to reflect and distinguish reality from imagination and suited-to-task from unsuited-to-task, to get as-much better as they need to be to deliver what some are supposing they surely will, and AFAIK that's not the kind of thing we're going to get with "LLMs, but more and bigger" and related techniques, but will require some new breakthroughs.
  - danShumway 3y ago
    > Hell, I've seen knowing how to use the Windows file explorer at a very low level of competence put a person in the top five percent by computer ability in an office. Within the last couple years. This is a really interesting point, phrased in a way that I wouldn't have thought to vocalize but that resonates with me. Tool-based and infrastructure-based changes that vastly improve efficiency and output based on competency with the tools/infrastructure are common; there's a lot of low-hanging fruit that is nevertheless uncommon for people to grab and use. Assuming current AI generation improves in exactly the way its proponents are expecting (which is it's own separate conversation), it is still probably worth spending a little time trying to figure out whether it's a tool that will actually see mass adoption or whether it'll end up in the same category as many of the other excellent tools that exist today that have remained inaccessible or unattractive to most people. Some tools do get mass adoption and people become competent with them at a mass-level. But only some, and it doesn't seem to strictly map to the quality of the tool. I'm not sure I can easily tell/predict the difference and where a theoretical future AI would fall; I would need to think about it more. But I just wanted to chime in and say that it's a good insight.
  - DoingSomeThings 3y ago
    > Middling Excel skills still make you a wizard able to do things nobody else can, in most offices (though not most tech offices) Do those skills translate directly to a job/pay, or are they simply nice-to-have competencies inside other jobs? I've always worked as an ops/analyst/spreadsheet guy but I've found over the years that those formula manipulation skills never actually translate to "advancement". They enable me and my company to do things easier than they would have otherwise, but appear as nice-to-have rather than valued in it's own right. Put another way - Is there actually a career where the primary hard skill is "I'm good at excel"?

Workaccount2 3y ago

I think it's a mistake to use current AI in current times as the status quo in ~5 years from now. AI will likely improve dramatically, as will the general persons ability to work with it. I wouldn't be one bit surprised if typewriter operators outpaced nascent word processor users in the beginning.

hotnfresh 3y ago
> AI will likely improve dramatically, as will the general persons ability to work with it. I'm not so sure about ability to work with it. Middling Excel skills still make you a wizard able to do things nobody else can, in most offices (though not most tech offices). Hell, I've seen knowing how to use the Windows file explorer at a very low level of competence put a person in the top five percent by computer ability in an office. Within the last couple years. Most folks can barely use the tools we have, and aren't going to be able to get good output from anything like the current set of AI tools. The ones who can, won't be cheap, and may be hard to replace like-for-like—I think a lot of companies are going to settle on more, cheaper, more-interchangeable workers than replacing them with a few AI tool experts with hard-to-measure-or-quantify skills. The tools themselves need an added ability to reflect and distinguish reality from imagination and suited-to-task from unsuited-to-task, to get as-much better as they need to be to deliver what some are supposing they surely will, and AFAIK that's not the kind of thing we're going to get with "LLMs, but more and bigger" and related techniques, but will require some new breakthroughs.
- danShumway 3y ago
  > Hell, I've seen knowing how to use the Windows file explorer at a very low level of competence put a person in the top five percent by computer ability in an office. Within the last couple years. This is a really interesting point, phrased in a way that I wouldn't have thought to vocalize but that resonates with me. Tool-based and infrastructure-based changes that vastly improve efficiency and output based on competency with the tools/infrastructure are common; there's a lot of low-hanging fruit that is nevertheless uncommon for people to grab and use. Assuming current AI generation improves in exactly the way its proponents are expecting (which is it's own separate conversation), it is still probably worth spending a little time trying to figure out whether it's a tool that will actually see mass adoption or whether it'll end up in the same category as many of the other excellent tools that exist today that have remained inaccessible or unattractive to most people. Some tools do get mass adoption and people become competent with them at a mass-level. But only some, and it doesn't seem to strictly map to the quality of the tool. I'm not sure I can easily tell/predict the difference and where a theoretical future AI would fall; I would need to think about it more. But I just wanted to chime in and say that it's a good insight.
- DoingSomeThings 3y ago
  > Middling Excel skills still make you a wizard able to do things nobody else can, in most offices (though not most tech offices) Do those skills translate directly to a job/pay, or are they simply nice-to-have competencies inside other jobs? I've always worked as an ops/analyst/spreadsheet guy but I've found over the years that those formula manipulation skills never actually translate to "advancement". They enable me and my company to do things easier than they would have otherwise, but appear as nice-to-have rather than valued in it's own right. Put another way - Is there actually a career where the primary hard skill is "I'm good at excel"?