Two men walking in the bar and going straight to the bathroom together. Man jumped to conclusions.

Not sure what that’s supposed to help with. I’d be even more uncomfortable if my steak had eyes and made eye contact than when a person does it.

Make a large enough model, and it will seem like an intelligent being.

That was already true in previous paradigms. A non-fuzzy non-neural-network algorithm large and complex enough will seem like an intelligent being. But “large enough” is beyond our resources and processing time for each response would be too long.

And then you get into the Chinese room problem. Is there a difference between seems intelligent and is intelligent?

But the main difference between an actual intelligence and various algorithms, LLMs included, is that intelligence works on its own, it’s always thinking, it doesn’t only react to external prompts. You ask a question, you get an answer, but the question remains at the back of its mind, and it might come back to you 10min later and say you know, I’ve given it some more thought and I think it’s actually like this.

Exactly. As the mandatory sexual harassment and money laundering trainings have taught me repeatedly, if the company knows about it and doesn’t do anything, they’re equally liable (and in many cases even if they don’t know about it). So stopping inappropriate behavior is in their interest.

Remember to look into his eyes

I don’t know if it’s some neurodivergence or if other introverts feel the same way, but that is something I personally find very difficult and uncomfortable and I can’t hold eye contact for more than a second or two at a time. What feels natural to me is to look at a person’s mouth when they talk.

Pear and gorgonzola is a typical combination.

They are effective, but in the other direction. I wouldn’t be surprised if they’re funded by fossil fuel companies.

Somewhere on the vertical axis. 0 on the horizontal. The AGI angle is just to attract more funding. We are nowhere close to figuring out the first steps towards strong AI. LLMs can do impressive things and have their uses, but they have nothing to do with AGI

How about a voting license that needs to be renewed every 30 years? You have to pass a test that checks if you are capable of thinking objectively or something like that.

Any type of criteria that is not absolute (like age), can and will be used to exclude certain groups of people from voting.

I was confused why a package manager would need to import posts from a social network.

Why name a new product the same as a very popular existing product?

Huh. TIL. I always wondered why libraries treat ebooks like physical books.

With the introduction of AJAX, web pages became apps. It was the advent of SPAs and SASS. Which enabled the things you saw as a consumer.

I am curious how you’d deal with the ambiguity of contractions vs. ending single quotes

That’s the thing, nobody even asks this question.

you could just match on /[a-zA-Z]+/

That would already put you in the top 10% of solutions I’ve seen so far on this problem.

That is totally a non-trivial problem, which requires a lot more conception before it can be solved.

Most candidates don’t realize that. And when I say they split by single space I mean split(' '). Not even split(/\s+/).

Does “don’t” consist of one or two words? Should “www.google.com” be split into three parts? Etc.

Yes, asking those questions is definitely what you should be doing when tackling a problem like this.

If I got that feature request in a ticket, I’d send it back to conception.

If I got it, I’d work together with the product team to figure out what we want and what’s best for the users.

If you asked me this question in an interview, I’d ask if you wanted a programmer, a requirements analysis, or a linguist and why you invite people for a job interview if you don’t even know what role you are hiring for.

That would be useful too. Personality, attitude, and ability to work with others in a team are also factors we look at, so your answer would tell me to look elsewhere.

But to answer that question, I’m definitely not looking for someone who just executes on very clear requirements, that’s a junior dev. It’s what you do when faced with ambiguity that matters. I don’t need the human chatGPT.

Also, I’m not looking for someone perfectly solving that problem, because it doesn’t even have a single clear solution. It’s the process of arriving to a solution that matters. What questions do you ask? Which edge cases did you consider and which ones did you miss? How do you iterate on your solution and debug issues you run into on the way? And so on

I always feel bad when I try out a new coding problem for interviews because I feel I’m going to offend candidates with such an easy problem (I interview mostly for senior positions). And I’m always shocked by how few are able to solve them. The current problem I use requires splitting a text into words as a first step. I show them the text, it’s the entire text of a book, not just some simple sentence. I don’t think I’ve had a single candidate do that correctly yet (most just split by a single space character even though they’ve seen it’s a whole book with newlines, punctuation, quotes, parentheses, etc).