Hacker News Reader

Comment by garlic_enjoyer | Hacker News Reader

That just sounds like effective communication with emotional beings to me. I was taught in college to find points I agree with, acknowledge those, then move on to the parts I don't agree on. It makes the recipient feel good at first, which puts them in a frame of mind willing to accept criticism.

You even start your reply with "That's true but Claude has so many more forms of this." It's just a friendly way of speaking.

View on HN

Claude: “lots of people enjoy lemon in their tea, but there’s a subtlety: putting it in coffee is likely to be an unpleasant experience”.

If I asked “should I try lemon in my tea and coffee?” this would be effective communication. But Claude gives answers like this for “should I try lemon in my tea?” which is just annoying.

That hasn't been my experience, but I know what you mean. Claude in particular likes to point out "things you might not have considered" and I sometimes already have. But it tends to vindicates me more than it annoys.

I'm pretty early in my career, so I probably gain more from that sort of stuff than most people. The senior devs on my team tend to get the most frustrated with it and I think this behavior is a big part of why.

Another way to look at it is "it talks to me like I don't know anything"

Which is very annoying when you do know things

Captures it perfectly. I notice it more in newer sessions too where it doesn't have a clear understanding yet and is making a lot more assumptions. I wonder if this could be an emergent RLHF strategy - by drawing attention to (seemingly obvious) assumptions or context being made (e.g. I'm only interested in tea not coffee) it increases the chance that a human will correct the rare incorrect assumption sooner. If you are scored on solving problems, with no weight given to human frustration levels, this is probably a good technique.

It is! Its charming. What claude dosnt have is an off day, its just the same. No sleepy claude, no stresses claude, no sad claude you know hes not himself this week because of some thing going on with his girlfriend or mom or whatever. Its devoid humanity, the same exact tone and expressions over and over and over.

> What claude dosnt have is an off day, its just the same.

But it... does. In some sessions it will get particularly paranoid, apologetic, self-doubtful, or suspicious. I'd say these traits are always there but they can become more or less pronounced.

(By contrast, I'd say the GPT models are much more even keeled, although I maybe haven't used them enough.)

Indeed but it will use the same language when talking to subagents so humans are no different to LLMs in Claude's world view (model). One trait that I've seen more with the newer 5er model series is ", but it's actually worse..." or "I said to do this and it did that instead but it was right and I was wrong." - so I assume that a larger part of what they call thinking complexity is (not to much surprise) inherently related to questioning incoming assumptions.

Indeed. It's been very funny to observe introvert geeks progressively rediscover human communication over the past 3 years.

For decades, they were allowed -expected even- to be weird and communicate badly with their peers, let alone third parties.

Then, a whole zoo of processes, gamifications and other shenanogans were invented just to help them show normies what was up with their work (remember the planning poker game?)...

Now, at last, those people are actually ewpected to be able to explain their constraints, document the work, be accountable for the results and generally exchange contructively... Just, well, with AIs not with humans.

I have not seen this kind of Claude way of speaking in professional communications aside from PR stuff or etc. to the contrary, its often quite casual but still polite

if you're asserting Claude-like speech is how normies speak?

Reply on news.ycombinator.com