Live data from Hacker News

Ask HN: Why is machine learning easier to learn than basic social skills?

news.ycombinator.com

1–10 of 182 posts

Ask HN: Why is machine learning easier to learn than basic social skills?

#1
(and any advice on acquiring these basic social skills?)

A year of haphazardly watching YouTube videos and reading papers and I learned enough to start contributing to real research. But 18 years of human interaction and I'm still missing out on social skills apparently. It's like everyone else has a degree in all these unwritten rules that I'm just supposed to know.

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#2
I don't know why it is and whether it is true or not. But I felt the same. And after several years of individual gestalt therapy and group therapy I have developed some social skills which I have been missing. Before that I tried reading books about developing social skills and they didn't help me almost at all.

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#3
My theory is that everyone starts off with the assumption that other people work like they do and, in the vast majority of cases, this is true so learning is easy. e.g. if you care about these annoying unwritten rules yourself, then it's easy to follow them. If you're one of the people who finds this annoying, you probably also don't care when other people break them, and then you have to learn by observation alone, which is much more difficult. In machine learning terms, their models are pre-trained, to a degree, out of the box, and yours isn't. What I don't understand is why this seems to so often coexist with a mind that can learn machine learning off of Youtube.

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#4
It's not. I think you just need to go out more. Social skills are not improved by spending all your time indoor in front of your computer playing with the latest cool new framework, or learning about ML from youtube or grinding leetcode for that next fancy job or playing videogames or watching Netflix.

IMHO social skills are easy to pick up, and that's coming from an introvert, you just need to put yourself out there as much as possible in as many new situations as possible, even though that means you'll miss out on the latest tech developments/opportunities or that latest Netflix show.

Basically you need to constantly keep venturing out of your usual routines and comfort zones and you need to be OK with failing at a lot of those new endeavors since there's no one size fits all formula for success here. You need to find your own way and pe prepared to feel exhausted a lot.

I think tech workers are most affected by this because of the SW dev culture both in workers and in employers that have normalized this FOMO driven, "carrot dangling on a stick", need to keep up with the latest tech developments or grind leetcode or you'll face the risk of becoming outdated and be overtaken by more ambitious new-grads, or be offshored, or just miss out on the more lucrative career opportunities, which is not helped by ageism and old-tech-stack-ism discrimination when hiring, plus crazy long multi stage interview sessions, take-homes assignments and what have you, that massively eat into your free time whenever you switch jobs, so these people are most likely the ones who have spend their prime years mostly indoor in front of a screen working away on code and algos, chasing that top dream job, instead of "out-there" working with people.

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#5
Well, youhave managed to ask a question on a public forum. So consider it as "Asking questions" achievement unlocked.

What do you consider basic social skills? You still have to unlock achievement "I don't expect other people to read my mind.":P

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#6
Keep trying. If you’re like me, the first 18 years will be the hardest. It will feel like trying to learn a game where the rules are continually changing. The gap between how a 12 year old communicates and how a 15 year old communicates is massive. Once you reach adulthood, the rules stabilize. This gives you a chance to catch up.

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#7
Comfort yourself in the knowledge that for most of the population, it's the other way around. We're all gifted in our own ways, and we need every "type" for a functioning society.

Plus, if things get too wild, and we're facing societal collapse, are at the brink of a robot uprising - at leas you are able to program a backdoor for the robots to ignore you. No amount of social skills will persuade a robot ;-)

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#8
The whole premise of this question is wrong, because it's not. Pretty sure a machine learning person would have an easier time learning social skills than it would be for a non-technical business person to learn the ins and outs of how a neural network works (and by that I dont mean learn how to use a pytorch api, but actually learning the theory and mathematics behind it).

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#9
Serious suggestion: have you tried watching YouTube videos and reading books on acquiring social skills? You did, after all, make some effort, however haphazard, to pick up ML. You might start with Charisma on Command's free videos [0] or How To Win Friends and Influence People by Dale Carnegie. [1]

[0] https://www.youtube.com/c/Charismaoncommand [1] https://en.wikipedia.org/wiki/How_to_Win_Friends_and_Influen...

Re: Ask HN: Why is machine learning easier to learn than basic social skills?

#10
How easy is it for you to make "Machine learning" into smaller subproblems? What are basic social skills. Do your methodic differs?

For me it's that I'm more likely to see my error with a topic like CS. While basic human interaction is not that manageable. It's interaction with Human A and Human B. These are distinct as Computer Science and biology. Finding the common denominator between them takes time for me. And don't forget that you interact with a lot of entropy. Communication is not easy.

How do you tackle these two problems ?

Post reply on HN