I don't think we have. Semantic symbolic computation on natural languages still seams like a great way to bring reasoning to computers, but LLMs aren't doing that.
> Semantic symbolic computation on natural languages still seams like a great way to bring reasoning to computers, but LLMs aren't doing that.
But they do close a big gap - they're capable of "understanding" fuzzy ill-defined sentences and "infer" the context, insofar as they can help formalize it into a format parsable by another system.
The technique itself is good. And paired with a good amount of data and loads with training time, it’s quite capable of extending prompts in a plausible way.
But that’s it. Nothing here has justified the huge amount of money that are still being invested here. It’s nowhere near useful as mainframes computing or as attractive as mobile phones.
I don't disagree with the conclusion, I disagree with the reasoning.
There's no reason to assume that models trained to predict a plausible next sequence of tokens wouldn't eventually develop "understanding" if it was the most efficient way to predict them.
The evidence so far is a definite no. LLMs will happily produce plausible gibberish, and are often subtly or grossly wrong in ways that betray complete lack of understanding.
The reasons why companies hire or don't hire someone usually have very little with the candidate themselves. From my experience, whenever this machine needs another cog, almost any will do - usually the first one within reach. And when it doesn't, not even the shiniest one will be of interest. So it's probably nothing personal OP
Nah. Every company has its lore about what makes a good candidate and they try to test for that. The lore is often rubbish (as in: there's often little correlation between interview performance and on-job performance), but there is still a process and that process rejects most applicants.
Or maybe it has everything to do with the candidate. They author recognizes they have spent much of their life being an unlikable jerk. Past actions can come back to bite you.
I tend to agree, which makes it all the more amusing that companies brag about being so selective. It seems like largely artificial and random selectivity.
Finding the right balance between "zero dependencies, building everything from scratch", and relying too heavily on third-party code. Both extremes are full of misery and frustration, and it's so hard sometimes to get it right. But most of the time these days I find it easier to start with fewer dependencies and probe for balance from there
I think it would be nice to have some UI when you first open the app, so you can at least learn how to play it. It's kind of hard to figure out what you're doing wrong when you tap and what success should look like.
Also, are there any difficulty levels? It feels like the very first sequence is kind of hard.
From what you are saying, I am assuming you bought the app (if so thank you!). When you first open the app (and every time after you completely close it from the background), there is a "?" at the top right. If you click it, it has a guided tutorial that explains in detail how the game works.
Please let me know if you have any other questions though and I can answer them here. If you are struggling a lot with a specific sequence, just close the app completely and re-launch it and the sequence will change.
I tried to set it at a difficulty that would not be too hard or easy. If it is consistently too hard though please let me know and I will make changes :)
Oh, have you ever read this book?
You make a fair point. I agree that by conventional definitions, this book may not be widely considered a 'classic'. Maybe just'personal classic'. After working for 8 years and finding myself at a career low point, this book completely transformed my mindset.
Keep an eye on this topic so that I can explorer which book I can read during the flight! Thank you
Wow, so many great responses, thank you all for that! I'm going to throw in my personal favorite: Siddhartha by Hermann Hesse. Somehow it touches me in a way that no other book does. It's also quite short, I think I've read it at least 3 times in the last few years and found something new and insightful each time
If the money is good, and the job doesn't take all of your time and energy, then you can just ignore the dysfunction and focus on your work. You can use your free time to look for something better, or to work on your own thing if you're ambitious. Being a cog is actually even liberating as cogs are only responsible for a very small part of the whole, so you can just do your part and not worry about the rest.
So far, I haven't found anything better than collecting individual RSS feeds that I like. Mostly personal blogs, a bit of HN of course. It's not a "platform" of course, but it's very future-proof
That looks pretty cool. But as far as I know, the biggest challenge is getting through the ATS systems on the company's side. Even in the old days, every job posting used to get ridiculous hundreds and thousands of applications, most of them garbage. With modern AI tools, it will probably get even worse.
So we're getting into this weird reality where bots are applying and bots are reading. And maybe some lucky 1-5% of all applications will get a human to look at them.
Have you tested your solutions with some of the big ATS systems? I wonder how well it would do.
what if we have chosen a wrong metric there?