The thing is, at least in the first example, the way I read it isn't "autistic people can't admit when they are wrong". From your own telling of the story it seems like a manager asked you "Should we rewrite the checkout page in React?", and you responded with "Should we rewrite the checkout page in React?". You failed to realize that all of the questions you asked them are the question they were asking YOU. They've been reading articles about technical teams gaining things they don't really understand from switching to React. Maybe increased velocity (they know what velocity is but don't know how React increases it), reduced bundle sizes (what's a bundle? why do I care?), faster rendering times (faster is a good thing, right?), etc. But they don't have the knowledge to judge if React will do this for them.
Answers to this question might be "No, we already use Angular which provides similar benefits to React", "Yes, switching to React from MooTools (I'm old) would simplify the design, increasing our velocity and probably resolve some annoying rendering bugs", or even "I'm not sure. Perhaps we should meet and discuss the weaknesses of that page and figure out if React is the right tool to resolve them".
Instead they got back the questions they were hoping you'd answer for them.
Many long years ago the CIO at a company I worked for asked if we could use machine learning to improve our data analytics. I responded with "I don't know, what kinds of questions are you hoping to answer?". We went back and forth on this several times, with me insisting you couldn't just throw machine learning at data without some idea of what your hoping to get back. This wasn't resolved until they brought in a vendor claiming to provide turnkey machine learning solutions, they ran it against our data and came up with the stunning insight that the biggest predictor of dollars sold was.... units sold. To be fair they did provide other data points, but it was all things we either already knew, or didn't care about and if we ever did could find out quickly.
I failed to recognize that the question I was asking was a big part of what I was being asked. If I had realized this and provided some insight into the kinds of questions machine learning could answer, and why we didn't need sophisticated machine learning algorithms for the ones the business cared about I could have saved everyone some time. Instead I kept rephrasing the question back at him, with neither of us having the insight to realize we were actually asking the same question.