[dead]

Are you implying that if I propose a "brilliant" solution here in the comments, Anthropic will DM me, inviting me for an interview?

Wait, I meant it sarcastically, but now I actually really wonder if you believe it (not insulting you! genuine question).

Only to lay me off next year after the IPO, no thanks.

Slow down, I am not throwing shade or even making any conclusions here :D I feel like you got irritated, perhaps specifically because I appealed to money - I will expand on it below. Nevertheless, I just presented my internal model and nudged people for some deeper context that could help reconcile it, that's all. And yes, your perspective did give me some new useful insight, thank you. I'll respond to some things.

> So developers who get paid more than $X,XXX,XXX per year don't need to pay attention to their customers. Got it.

But the point is that the specific discussed behavior is mostly technical. It's a mostly under-the-hood behavior, and mostly something that doesn't involve tradeoffs for the user based on the implementation chosen. So it's not really something that you would just go around and ask users, especially when you have all the session statistics from CC, skilled systems engineers (where I partially used money as a proxy) who are paid to spend months on this problem and can iterate ideas among each other, polls and questionnaires, thousands of experiments from within different sandbox configurations, insight into the training of the model itself, etcetera.

In all of that, six random people from a forum, even from one such as hackernews, seem to me to be a totally statistically insignificant, comparatively under-informed dataset.

Something like that is where intuition leads me.

Now, the reason I appealed to money is multifaceted.

- It's an investment from the company into a brainpower that is directly tasked with solving this. An investment so absurdly huge, that that team delegating such fundamental problems to six people on a forum for ideas seems hysterical to me.

- They are being paid absurd sums of money. For-the-value-they-provide. I am not, and neither is anybody in this comment section (probably). Not even close. Yes, I expect them to be able to solve these problems much, much, much better than the average HN commenter. Otherwise - why the hell are they getting paid that?

- The guy getting 100$m for joining the company rattled me, I admit.

Now, those are all the reasons behind my intuition. Of course, I understand that a guy optimizing a model can easily bring 100$m of additional revenue to a company of OpenAI's scale and money turnover. I understand that these people are all still just humans, like you and me. It just... rubs me the wrong way. I concede my irrationality.

Anyway, I think your point about "listening to users" is... weird, because I don't think that's the same thing as what I was talking about at all.