It doesn't look good.

The key impact of easily available AI — call it AI as a Service — is augmenting productivity for comparatively low-skill tasks. This means that the same tasks can be done by fewer people with a lower barrier of entry. Both things lead to lower wages regardless of any increase in productivity. Remember:

Your salary is not your productivity but the cost of your cheapest replacement.

While you play with easy-to-use AI, remember that your potential replacement is finding it just as easy to use.

Do keep in mind that this happens at the task level, so you'll feel an impact on your career even if your job can't be fully automated. Easily automated tasks allow lower wage employees to take more of your tasks; this won't turn into an easier life for you, but rather into a lower high-wage/low-wage workforce ratio, therefore fewer career opportunities and/or lower wages along the path.

If you're at the beginning of your career I don't know what to tell you, but maybe what's below might help you too.

Basic economics suggests that the way to push for higher wages (other than collective negotiation, which is an entirely different discussion) is not just to increase your productivity but to increase your productivity in a way that helps you negotiate with your company.

Now, there are two aspects to this related to AI:

• Increasing your productivity without using AI is a losing proposition over the long term. AI on its own doesn't always do much, but whatever skill you have it'll be nearly impossible to improve it as far or as fast without it.
• Increasing your productivity using your company's AI or publicly available AI is a losing proposition from day one. You won't get paid more for using things your co-workers can use just as easily. You're just avoiding getting fired because you don't.

This is a point I've made elsewhere and often, and it applies also at the level of competition between companies, but it's worth repeating because it's something a lot of career advice gets wrong. You are often told to learn to use the AI tools provided by your organization or publicly available, and, yes, you need to do that to get and keep increasingly lower-paid jobs.

For better wages you require a complementary skill that's (unsurprisingly) less talked about: the ability to find, teach, and use AIs that none of your competitors have, just as you are taught from the early on that you have to learn to find, learn, and use skills that none of your competitors have. Any programming or technical skill you have will of course be applicable for this — even without AI, it's surprising that not all programmers code personal tools to make their own job easier — but lacking the know-how to develop an AI doesn't imply you have to fall back on using only the same tools everybody else is using. You can and should be able to understand and map the cognitive architecture of your activity (after all, this is part of the definition of expertise) and then look for specialized tools to augment your cognitive abilities in key areas of it.

Just as it's a given that any dedicated professional can find rarely-used sources to improve their let's say on-brain capabilities, it's a given that any dedicated professional can find rarely-used tools to improve their off-brain capabilities. A cognitive advantage is a cognitive advantage whether it comes from something you know how to do your competitors don't or from an AI you have access to that your competitors don't know exists. Thinking that something like ChatGPT will help your career is like thinking that you can get a better job for knowing how to access the Internet. If you were the only one? Sure. When most everybody can? No.

It's all essentially the same phenomenon: the supply-demand curve of cognitive skills. To thrive in a changing economy the first skill is to think not just with your combined brains and software but also about them. Much of your personal career strategy is your personal skills strategy, and to compete with other AI-augmented professionals much of your personal skill strategy has to be your personal AI strategy. You better have one and it better be yours.