Machine Learning/Data Sciences/R

Searching for a Job in the Age of Hype!
This month, we’re trying something different with a change of venue. I’m hoping for more audience participation, and I’d love to add your voice to our discussion.
There has been a lot of hype surrounding AI, particularly in machine learning (ML), ChatGPT, and even quantum computing. I’m interested in hearing about your personal experiences with job searching in AI, ML, and ChatGPT. While many blogs discuss this topic, it would be valuable to survey people directly and gather their insights.
One job which seems to have been automated with AI (although not much AI required) is upgrading and releasing new versions of software, e.g. Java releases. I have seen a metric which states how many man hours it replaced. Well, I agree that the man hours count is good metric, I just don't like the AI label on the task. Sort of like claiming Romba robot vacuum cleaner is AI, but it is really just some feedback control system with memory and sensors. I do think that word processors, PCs and portable printers have improved productivity of office staff.
Lately, I’ve come across posts claiming that anyone with some ML experience can easily find a job in the field. Or management drinking the AI KoolAid about how you can now replace half the programming staff with AI. However, recent graduates with degrees specializing in machine learning often share their struggles, including having to settle for positions far below their expectations. There are high paying positions for Director level roles with years of industry experience to back it up. Companies want employees who can deliver on the hype.
This kind of hype isn’t new. In the 1990's, Japan launched their 4th generation computer initiative, which was going to turn all programmers into specification writers, QAs and testers. I guess that IF you can specify everything correctly, then maybe. I think that you are going to spend a lot of time as a prompt writer. Anyone who has spent a week in a software design planning committee is convinced that no amount of specification will ever get detailed enough to write workable software. You just have to work and test the problem enough to get a workable solution.
We can recall the internet implosion of 2001. I remember 1992 when DEC launched the Alpha chip—a technical marvel that didn’t sell as expected. Or Arvind’s DataFlow architecture, which was supposed to revolutionize CPU design by addressing the von Neumann bottleneck, yet Motorola lost $50 million (at a time when that was a significant amount of money). And, of course, we’ve been “20 years away” from sustained nuclear fusion for the past 40 years.
Remember, to err is human, but to really foul things up takes a computer. I think we are in "The Uncanny Valley" and will be there for quite a while

See AI Snake Oil ( https://www.youtube.com/watch?v=UeSy8GVBs9A )
MIT economist says that AI will only replace 5% of jobs ( https://www.youtube.com/watch?v=unrvuUsM5vk )
Another example is how one ML grad took a job where he was suppose to fulfill all the hype and was being fired since it wasn't going to come true on the scale advertised. This reminds me of a joke, "A successful salesman is one who lies to make a sale! "A successful engineer is one who makes the salesman's lies come true."
The funniest job post was from a company that is pushing GenAI, but warns applicants not to use it for resume and cover letter generation.

NB. The top image has been generated with ChatGPT!

Please fill out the survey
https://www.surveymonkey.com/r/63P66J8