This is an AI “interview”, which is to say you’re reading someone else’s ChatGPT conversation reposted to a different webpage. That is as uninteresting and devoid of insight as you would expect.
The subtitle is “Don't take it from me, hear it straight from the AI”. The author’s conclusion comes from asking the LLM if AI could’ve solved the problem, and it responding “no”. Well, the LLM doesn’t know, it’ll just make up text anyway. Taking their denials at face value is as wrong as doing the same thing for their assertions.
I thought this was going to be something interesting with a modicum of value, like using a physics simulator to reproduce the events of the Apollo 13 and then asking an LLM to fix the problem. As it is, there is nothing to see here. Neither detractors nor fans of LLMs have anything to learn or gain from this post, it’s just more slop.
latexr•7mo ago
The subtitle is “Don't take it from me, hear it straight from the AI”. The author’s conclusion comes from asking the LLM if AI could’ve solved the problem, and it responding “no”. Well, the LLM doesn’t know, it’ll just make up text anyway. Taking their denials at face value is as wrong as doing the same thing for their assertions.
I thought this was going to be something interesting with a modicum of value, like using a physics simulator to reproduce the events of the Apollo 13 and then asking an LLM to fix the problem. As it is, there is nothing to see here. Neither detractors nor fans of LLMs have anything to learn or gain from this post, it’s just more slop.