Yes, it would be correct not to pursue a technology that has that chance of wiping out the world. But where are the people who are suggesting these probabilities getting their numbers? I personally can't imagine where, and I'm an engineer with a specific technical interest in LLMs. And I haven't heard one clear description of the methodologies used to calculate these chances.

The much more likely explanation to me is that people are just spitballing, either because they've watched too much sci-fi, or they have some weird counterintuitive agenda (e.g. Anthropic and OpenAI trying to position themselves as the amazing, trustworthy keepers of this dangerous technology before their IPOs).

I think what the parent poster is trying to convey, is that let's say the apocalyptic stuff has a 5% chance as you say, but the non-apocalyptic stuff (social-economic chaos, total centralization of power, eradication of social mobility, total information/trust collapse) might have a good 95% chance as we are seeing it starting to unfold already.

But the discourse is dominated by paper clip experiment discussions and not let's say by the fact that new grads have an unprecedented difficult time getting jobs. Unsurprisingly one of those is a sexy hypothetical beneficial to power and the other one is not.

Multiple problems can be important, pointing a different one out doesn't invalid or take away from another one.