From what I can see chat gpt can produce code snippets and has quite a bit of surface-level knowledge of technologies and their implementation. It's very useful to grab snippets from, ask basic questions, etc.
I can imagine it will be more useful as the base model is trained more and as companies are able to train it specifically for their use in building enterprise applications

What I do not understand is why people think it will take software engineering jobs. It still requires very explicit instructions to do even basic things and those instructions need to keep in mind building clean scalable systems. If the idea is "well it can do this now, just imagine with it can do in 5-10 years", then my main question would be what's the timeframe for this?
2023-2027: general adoption by the public
2028-2035: adoption for building enterprise applications, integrating into enterprise systems, use by developers at major companies
2036+: ???

The most I can see this doing is removing some bloat CRUD API developers. Any developers that care to continue learning or those who are already in senior positions will keep on designing scalable systems and use this as a tool to help them do it more quickly.

Thoughts?