Tbh I have yet to see something that would severely impact SWEs.
But the pace of the industry makes me feel a bit uneasy. I mean just look at DALL-E 2. The improvements over the previous version were very impressive. OpenAI also has Harvard AI researchers working with them to solve the super alignment problem.
I want to turn a blind eye to all of this but that feels ignorant. GPT5 could severely impact jobs or it may not. You just can't say for sure.
I agree that there are lot of issues to solve as well. The main ones that I feel are currently slowing adoption are compute cost and prompt injection.
Hardware could be solved by building ASICs. Groq is already doing it. There was also the trillion $ demand by Altman to build the hardware required that garnered headlines. Also, when he was fired, he was actually routing funds from Saudis, VCs etc. to start his own AI hardware company (Project Tigris) with Jony Ive. I'm pretty sure there are already startups that have begun tackling these problems.
So you never know.