Comments by "John Smith" (@JohnSmith-op7ls) on "NeetCode" channel.

Yeah, LLMs are advanced auto complete. They won’t magically become sapient no matter how much training, memory, and processing you throw at it. It’s just fundamentally the wrong architecture. It’s like how people use to take these vague, nonsense estimates of the raw processing power of the human brain and point out that we’ll soon have super computers with more power. Well, we do, and yet none of them are sapient. The internet as a whole has orders of magnitude more processing power, why hasn’t it magically become self aware? People who don’t understand this stuff pretend it’s just a matter of more data, faster processing, that’s not how biological neural networks operate at all.
57
Oh, all those big tech companies can absolutely be stupid. We’re talking about 4 people here. You don’t think 4 people in related industries can all be dumb about the same thing? If so, you’ve got a lot of history to catch up on. And Google investing in AI was absolutely not worth it if you’re talking about transformers and LLMs. They stumbled upon an advancement at their expense, then did nothing with it, leaving it there for OpenAI and MS to take it and make billions, while,they had to scramble to play catch up, and so far have utterly failed on that front, at much greater expense. So not only did they waste tones of time and money but created a large competitor and heavily bolstered one of their largest existing rivals.
5
Any company that does these kinds of coding interviews is not somewhere you want to work. It shows they have no idea how to find real talent, so you’ll be working with people who just memorized a bunch of useless info to pass their tests. It’s so rare to use any of these algos much less more than a handful that it makes no sense to try and memorize dozens of them just in case. If you ever need one for a specific issue, you can look up the pros and cons of different algos or maybe there’s an objectively best one, then odds are there’s 100 implementations for it already in your language. If not, you can simply learn the algo and implement it. There’s just no need to waste time memorizing things you’ll probably never use when you could be using that time to learn productivity, organization, code and systems architecture concepts and a high level understanding of the most popular design patterns. Things that will actually help you as a developer rather than memorizing test answers that only test how much time you wasted memorizing test answers.
2
Uhhhh, why are these still common interview questions. People just waste time memorizing this stuff long enough to pass the interview, then almost never use it again and when they do, they just spend 10 seconds looking it up.
2
@MatthewKelley-mq4ce But statistics and probability is what it’s doing, we know how it works. We don’t know how biological neural network work anywhere close to the same degree so you can’t go around making comparisons like that.
2
@ItsRyanStudios AI is a nebulous term, it has no agreed upon definition. So AI by definition isn’t AI. AI is a terrible term anyhow. It means whatever you want, therefore it coneys nothing. Better to be very specific about this sort of thing, describing specific capabilities so you’re actually saying something besides, “Just read my mind and you’ll know what I really mean.”
2
@TheManinBlack9054 Playing semantics gets you nowhere. If you’re going to go that low with your bar for what is “intelligence” then every hello world app is intelligent. An alarm clock is intelligent.
1