Modern AI Discourse is Talentless Business Students Trying to Give Young Engineers PTSD

Since the AI hype took off in 2023, a specific archetype has emerged on the internet. Especially on LinkedIn (which is not a great cohort...), there has been a dramatic increase of people giving lectures on how "AI will change everything" and how "Ai WonT tAkE YouR JoB, SomEOnE uSiNG Ai wILL tAke YoUR JOb".

One particular topic that tends to emerge with these people is that coding is dead. Now, no one can predict the future, especially not the people who say coding is dead (2 years after advocating for "everyone should learn to code"). As a result, we could simply ignore these claims. However, this never-ending doom scrolling does have a very negative effect on younger engineers that want to get into coding. Hence, I want to write a little blog post ranting on this topic.

Characteristics of an AI enthusiast

Prior to the release of ChatGPT in 2023, AI was considered a niche topic. However, since that moment almost everyone seems to have an opinion about it. Did we spawn millions of knowledgeable computer scientist in the meantime? No. Instead, a bunch of business students that are have never actually worked with tech (and therefore jump on every tech-bandwagon, like AI, crypto or web 3.0) have found a new pastime. Namely, to lecture people in tech on how tech will make them obsolete.

A central narrative pushed by this cohort is that coding is dead. Now, after all the correct predictions about crypto currencies and web 3.0 (wink wink) I do take their perspective seriously. And of course, they could technically be right at some point. However, because scary posts like these tend to get more clicks than balanced posts from AI researchers, this narrative is becoming very prevalent online. As a result, young ambitious engineers and computer scientists are becoming more insecure about their job prospects. This is a very problematic side effect.

Note, this article is not meant to debunk the prediction about software engineers being replaced. In fact, there's no 100% certainty that most jobs we do now will exist 10/20 years from now. So yes, maybe there will be an advanced AI that can do the job of a programmer one day. However, most professions related to business studies can be replaced by nothing right now, and those jobs seem very secure. Hence, it's probably a bit premature to start worrying about our jobs. Apparently, we can become obsolete, but still be employed.

Some comments on AI taking jobs

Obviously, the recent advancements in AI are very impressive. And (as someone that has presented at multiple (...2) conferences on AI in the past) I know it can add real value to business processes and services. However, this value is typically added when the AI is aimed at removing the robot from the human. For example, in my second AI related publication, we used BERT to collect evidence in the legal discovery process. Typically, such tasks are done by legal interns that are forced to go through thousands of emails (if not more) one-by-one to find evidence that could potentially be used in a case. Hence my point, a robotic task that's done by a human.

In contrast, more artistic roles are a lot harder to effectively replace by AI. This is due to two reasons. First, the risk of plagiarizing. Note, I'm not a lawyer, so I can't say anything about the legal validity of the copyright claims put forward by the NYT, Authors Guild, etc. However, to me, the whole concept of a loss function in machine learning is to plagiarize data points. As a result, being impressed by the art produced by AI, is a bit like being impressed by a student that copies homework. It doesn't have true originality, which is required to be considered a creator.

The second reason AI is less fitting in art-related domains is due to board rooms. Contrary to popular belief, large commercial AI projects don't just "follow the data". In fact, they have to adhere to limitations set by people (with biases). These limitations are often there to manually filter out controversial results. However, innovative creativity centers around controversy! In fact, most great artists (Picasso, van gogh, Michelangelo, etc) progressed art by creating data points that were very controversial (sometimes illegal) and impossible to predict.

In conclusion, the likelihood of a job being automated by a robot, depends on how robotic the job is. Now, this doesn't mean AI can't do programming jobs per se. But, it does mean that creative ideas do require things that the lawyers in corporate board rooms won't easily allow an AI to do: take risks. Please remember that for your next coding project.

An optimistic final note

If you are still worried about AI coming to take your job, good news! A meteor might fall on your house and kill you! That might not sound like good news, but hear me out. Life is uncertain, and no one knows what's around the corner. So how can we truly worry about what happens 30 years from now, if we can't even predict it?

Hence, live your life step by step. Do you have a job now? Try to enjoy it and do it well. Are you in college now? Try to enjoy the experience. You'll be fine.