unless you’re top 1% (rich or technical) it won’t matter for you when AGI is achieved *in a lab*
the point of inflection for you is when AGI becomes *generally available*
for at least the next decade, you should be looking to combine “AI” with traditional rules based software
what do I mean?
Consider a project I worked on recently.
We had a nice little prompt that created some neatly formatted output.
The moment we introduced a new requirement to the prompt, the output collapsed.
I mean, the output literally stopped making sense.
the solution was to use TRADITIONAL programming methods (JavaScript and CSS) to format the output
it’s good separation of concerns - LLM for the content, JS/CSS for the display layer
but also you should learn that trivial (but predictable) software PAIRED with machine learning
…can create products that are 1000x more complex while remaining reliable than what is possible right now using purely machine learning / artificial intelligence models. ✌️