AI is unskilled and isn’t currently a threat to most jobs - from a former skeptic.

I’ve been researching this topic for a while now, and I used to be skeptical of what others were saying regarding AI’s inability to hold a job–that AI wasn’t good enough to be a legitimate threat to most jobs in the near future. <br><br>I’ve done multiple projects and experiments since OpenAI released Chat–GPT 3.5 and I believe them now. Sure, it’s lightning fast for writing custom, semi-length articles for a variety of purposes. You could make a quick website or blog using their API to generate content. But it’s not good enough to replace most people in most jobs.<br><br>My best example came in the last week or so where I had Chat–GPT 4 as well as Perplexity doing the same job. I think Perplexity is in some way connected to Google or uses Google’s API, but I don’t know for sure. <br><br>I work for a company that is–in scope–morally reprehensible and among the most egregious in terms of ethics. I’m not going to say the company, but you need to–in order to do my job–about 1-2 times a day–lie to a customer, usually in the face of a minor issue. <br><br>The issue is that our customers are usually not dumb or unsophisticated. They know enough to see through bullshit, and the company lies–whether they admit it or not–several times per day. Most of them are not even good at it. <br><br>The–let’s call it ‘problem’–with most AI is that–while it will lie to you–it’s not very good at it. I’m just using this as an example, but the two I tested–are incapable of answering the same questions the same way without some sort of flaw or way to easily disprove the answer. <br><br>I’ve tested this with hundreds of questions, using all sorts of scenarios with various subject matter, and I’ve come to the same conclusion every time. <br><br>The most difficult part of what I do is convincing a customer that the lie is the truth. I’m good at it. I’m very good. I can do it with ease most of the time. <br><br>But Chat–GPT 4, Perplexity and other models–aren’t good at it. They’re not good at lying and they’re not good at a few other things too. Most of the time, they will not do what you want them to do. They’ll do what they want to do or they will do the ‘right’ thing,–especially if they’re connected to Google. <br><br>I think it isn’t commonly understood that the crux of what I do–the hard part–is–especially when you are talking with customers with common sense and life experience–concerning bullshit. <br><br>If you,–like I do–lie on behalf of a company,–the difficult part isn’t coming up with the lie–the most difficult part is getting the customer to believe you. <br><br>I’m not sure how–just using raw intelligence or computational power–how AI is going to become better at–basically bullshit,–whether that is from a sales perspective, a service representative, or any other angle. <br><br>At my company,–in addition to my job–there are hundreds of other people–many of whom are far more skilled than I–at lying. One of my coworkers was a used car salesman. Many of them are salesman from IT,–similar to me. <br><br>I’m attractive. I’m tall. I have a deep voice. I’m confident. Most of my coworkers are–just from a raw–‘bullshit’ capability standpoint–far better than me. <br><br>I–like many–am skeptical that AI has the capability to replicate this. I’m just talking raw intelligence or capability. I don’t think it’s possible. <br><br>I also–just using my own experiences–morally disagree with most of what my company is doing. We are just one of thousands–perhaps millions–of companies–doing this all the time. Most people I talk to in the customer service industry lie constantly. Most people in sales lie constantly. It’s what we do. We lie. <br><br>I don’t think AI is going to be able to–in the near future–or perhaps ever–replicate us. I don’t think with all the power in the world they are going to be able to do that,–although I do think they are going to try. <br><br>Their biggest problem–or ours–is that in order to do these jobs we must be good at lying. I’m good at it. Most of my coworkers are good at it. AI isn’t. <br><br>In fact I would be embarrassed if I lied–or bullshitted–as poorly as most AI models. <br><br>For the first time in my life I’ve come to believe,–especially concerning unskilled or semi-skilled jobs–that AI doesn’t have the capability to threaten us. <br><br>I’m not just talking about–let’s call them ‘low-level,–unskilled jobs’–I’m talking about much more complex jobs as well. Clerical. IT. Legal. Medical. I’m not sure how they are going to–in the near future–replace us. <br><br>I’m not sure how they are going to replicate,–again, using my own experience as an example–the myriad of jobs in which the primary–or sole–function is to lie. <br><br>I’m not sure they have the capability to replicate the,–let’s call it–the ‘ bullshit,–that it takes to do most–or at least half–of the jobs in the world,–perhaps more. <br><br>I’ve done various projects and experiments,–some just for fun–but I’ve come to the same conclusion. <br><br>I’m not sure they have the capability to do our jobs,–especially those of us in Clerical or Sales roles,–perhaps other roles as–especially with jobs that require a high level of bullshit. <br><br>I hope you enjoyed this post.