white collar

Imagine you are starting at your new career job, but you can’t use spreadsheets or calculators or computers. everything is manually done. you still know the general concepts and ideas of what you need to do, but the way you do it is very different. I think this is how white collar will be affected but in reverse. they will gain tools that will accelerate their work and productivity. think about how productive someone with programming skills is when they can automate their work tasks. now this automation capability is democratized. but not only that, there’s vast data, content, and knowledge repositories accessible, queryable, and actionable by this human-mecha hybrid worker.

Accessible is possible today if the human knows where all the information is. At merely this level, it is very tedious going to various different knowledge bases and going through all the records to find what you need and cross-referencing relevant content in other datastores. Think of it like poring through printed physical library books and records. Queryable is enabled if it is all indexed and all aggregated and manipulated from a central interface. Think federated search. Actionable is when there is easy access to vast and deep knowledge sources and it is trivial to ask sophisticated complex questions to it and receive a custom generated response which may also have the option to take action on the human’s ask.

The white collar knowledge worker will very likely still be valuable if s/he does more than just a repetitive task that can be replaced by a program. So the question to the business is, does the top line move up? or does the bottom line move down? do we keep the existing workforce size and take the gains in productivity? or do we settle for status quo work output and reduce the workforce? I think if you’re a company who is confident in your core business value and mission, you will want to raise the top line and and dominate your industry before someone else does it.

over reliance

But then there’s a down side to a lot of this. People will become over-reliant on this technology. One day in the (near) future, someone will say, “AI said this was the right thing to do.” Here’s where the spreadsheet analogy breaks down. You might use a spreadsheet to organize data and auto calculate figures. It may produce incorrect results, but this is typically because it had bad data put in or it was not structured properly. It is a reliable tool. Generative AI, in its current state is prone to hallucination.

Imagine feeding the entirety of all the world’s tabloid magazine content into a “thinking” machine and asking it for custom responses to your questions with its limited and skewed corpus it has been trained on. The generated responses would likely include wacky things like an alien Elvis giving birth to a batboy baby. This is an extreme example, and there is alignment and tuning and chain of thought strategic actions to ‘normalize’ responses, but the possibility for error is still there.

Here is where the role of the human is critical, s/he needs to be well versed in their area of expertise and fact check what the AI is doing. If it were me, I’d probably use AI to help fact checking since that would be so tedious, but I’d probably try to make sure I still do a thorough job.

blue collar

Now what about blue collar jobs? or jobs more in the physical world? once these technologies are perfected, I think these jobs will simply be replaced by machine. Similar to how machines revolutionized agriculture; tractors working the earth can be a lot more productive that using large teams of human labor. Uber/Lyft drivers will be affected. People working at quick serve restaurants will be affected.

After the initial investment of the robotics systems into cars, restaurants, package delivery, barbershops?! massage services?! chiropractic clinics?! … the net effect is cost reduction. There will still be a need for a manager at a Popeye’s chicken to help the customers who complain or somehow got a wrong order. But still, many of these jobs will be gone. So, will they reduce the prices on their extra crispy spicy chicken sandwich? probably not…

What’s safe then?

so then, what jobs are safe? I think teaching jobs will be safe, for the most part. Especially with younger children, I doubt that parents would want a computer program to be the only source of instruction. Or if it were a robot teacher, I can’t imagine how it would handle managing a rowdy classroom. And in the upper grades, the human connection and relationship is important in the learning process. So, teachers who get their kids motivated and interacting and being human themselves, I think those are skills that will be highly sought after

Also, comedians. I think AI is really bad at making jokes. And I have a low bar because I love telling “dad jokes”. So for now, I think comedians are safe. speaking of which here’s a bit from Ronny Chieng that I thought was really good - www.instagram.com/reel/DD-M… - I don’t know if AI will be able to come up with comparable content.