Entry-level jobs down by a third since launch of ChatGPT

However I agree that democracies around the globe mostly work, in the sense that voters get what they are voting for.

What you will find is that the agent is much more successful in this regard.

The LLM has certain intrinsic abilities that match us and like us it cannot actually code 10,000 lines of code and have everything working in one go. It does better when you develop incrementally and verify each increment. The smaller the increments the better it performs.

Unfortunately the chain of thought process doesn’t really do this. It can come up with steps, sometimes the steps are too big and it almost never properly verifies things are working after each increment. That’s why you have to put yourself in the loop here.

Like allowing the computer to run test and verify an application works as expected on each step and to even come up with what verification means is a bit of what’s missing here and I think although this part isn’t automated yet, it can easily be automated where humans become less and less involved and distance themselves into a more and more supervisory role.

1. For business stakeholders, they are motivated to find ways to use AI to achieve whatever they want, because business requests are always ASAP and they don't want to wait for downstream software engineers to do the job. This has always been the case since the start of the computing business, I believe. (Maybe not the case during the mainframe/supercomputing era as the "business" sometimes are engineers themselves so it's easier to communicate)

2. Many software engineering jobs are not that technically challenging. For example Frontend, Data engineering, etc. A lot of time is spent on requirement clarification and collecting. Business might eventually figure out that the best way to use AI is NOT to make AI adapt to humans, but to make humans to adapt to AI. I'm pretty confident that if we have an integrated AI and the business stakeholders can ask questions in a way that AI can understand, 80% of my job (as a DE) could be done by AI. But it requires feeding data into AI and training the business stakeholders to use it properly. TBF, a lot of my work COULD be solved by automation, but we never had the time to properly automate the data pipelines.

3. Whatever the outcome, junior software developers are having and are going to have to a tough time. Unless they work in some low level system programming positions, they have to prove that they are MUCH better than AI to justify for a hire. Nowadays companies expect senior developers to use AI to enhance their productivity, so juniors need to learn that too to catch up -- but they also need to learn to program without AI to actually obtain the knowledge properly.

4. AI is going to impact every field, not just programming. Office work is the first to get impacted, and then blue collars too. Unions and governments are going to hold for a while, maybe a long while, but as long as there is one major player (China for example) that is pushing for AI advancement, the others HAVE to follow suite, or they run the risk of losing everything. The Russo-Ukraine war and Israel-Iran war are the bells that toll for all of us, and the existence crisis brought upon all of us by the increasing numbers of hot wars are going to make everyone push for productivity, whether in production or killing.

5. If we can't figure out how to make the world a bit more fair and happier for the average people (you and me) because AI really takes off, a dark dystopian future awaits us. Good luck. We might have some 10-20 years to achieve that.