In my school projects I've found them super useful for working with libraries. In the past it felt like the theory I learned was pretty low impact. 80% of my time was spent learning the quirks of a library. Now it feels like I can take theory and iterate over a ton of different solutions without having to worry about learning whatever library the professor requires. Basically I'm I feel like it lets me spend more time learning the thing I want to learn rather than all the busywork around it. Would be curious to hear your thoughts. Thanks!

Like at some level, yes, transformers are trying to emulate a human brain but the second you ask folks if they do a good job of it, I think most rationale people would say no.

I'm not sure I understand this. Why not?

I don't actually think the commentator you responded to is arguing for either of these narratives and I thought it was a pretty useful way to look at some of these arguments.

Anyway, I've found that if you want to get a coworker into reading technical books, the best way is with a novel or three. I've had good success with The Martian. The Phoenix Project might work too. Slip them fun books until they've built a habit and then drop The Mythical Man Month on them. :)

I had a similar experience showing her Skyrim. She never quite figured out how to walk and look at the same time. Made for an absolute berserker of a barbarian.

In any field, when you're surrounded by competent people, you'll begin to take that baseline competence for granted. I think especially so in ours due to virtual forums. I can work with my peers all day, go home, and talk with more online. It's enlightening to walk a curious outsider through your day (and probably also a great test of the systems you have in place).

fixed it for you! now it’s in a casual, laid back tone.

It feels analogous to complaining about how Michelangelo painted the Sistine chapel on the ceiling instead of on a canvas where we wouldn’t have to crane our necks to see it.

This single sentence probably took so many man-hours. I completely understand why they’re trying to integrate ads but this feels like a generational run for a company founded with the purpose of safely researching superintelligence.

Conversations like this are always well intentioned and friction truly is super useful to learning. But the ‘…’ in these conversations seems to always be implicating that we should inject friction.

There’s no need. I have peers who aren’t interested in learning at all. Adding friction to their process doesn’t force them to learn. Meanwhile adding friction to the process of my buddies who are avidly researching just sucks.

If your junior isn’t learning it likely has more to do with them just not being interested (which, hey, I get it) than some flaw in your process.

Start asking prospective hires what their favorite books are. It’s the easiest way to find folks who care.

Cosplay that role long enough and you become it. I’m still learning but it has been a great signpost for me over the last couple years.

Cheers and keep crushing it!

For many classes this is still the case, and I lump these folks in with the great students. They still care about learning the material.

My experience has been that these students are super common in required undergrad classes and not at all common in the graduate-level electives that I’ve seen this happening in.

> ... this has never been true in my experience, as a student or hiring manager.

Good to know. What’ve you focused on when you’re hiring?

As someone studying CS/ML this is dead on but I don't think the side-effects of this are discussed enough. Frankly, cheating has never been more incentivized and it's breaking the higher education system (at least that's my experience, things might be different at the top tier schools).

Just about every STEM class I've taken has had some kind of curve. Sometimes individual assignments are curved, sometimes the final grade, sometimes the curve isn't a curve but some sort of extra credit. Ideally it should be feasible to score 100% in a class but I think this actually takes a shocking amount of resources. In reality, professors have research or jobs to attend to and same with the students. Ideally there are sections and office hours and the professor is deeply conscious of giving out assignments that faithfully represent what students might be tested on. But often this isn't the case. The school can only afford two hours of TA time a week, the professors have obligations to research and work, the students have the same. And so historically the curve has been there to make up for the discrepancy between ideals and reality. It's there to make sure that great students get the grades that they deserve.

LLMs have turned the curve on its head.

When cheating was hard the curve was largely successful. The great students got great grades, the good students got good grades, those that were struggling usually managed a C+/B-, and those that were checked out or not putting in the time failed. The folks who cheated tended to be the struggling students but, because cheating wasn't that effective, maybe they went from a failing grade to just passing the class. A classic example is sneaking identities into a calculus test. Sure it helps if you don't know the identities but not knowing the identities is a great sign that you didn't practice enough. Without that practice they still tend to do poorly on the test.

But now cheating is easy and, I think it should change the way we look at grades. This semester, not one of my classes is curved because there is always someone who gets a 100%. Coincidentally, that person is never who you would expect. The students who attend every class, ask questions, go to office hours, and do their assignments without LLMs tend to score in B+/A- range on tests and quizzes. The folks who set the curve on those assignments tend to only show up for tests and quizzes and then sit in the far back corners when they do. Just about every test I take now, there's a mad competition for those back desks. Some classes people just dispense with the desk and take a chair to the back of the room.

Every one of the great students I know is murdering themselves to try to stay in the B+/A- range.

A common refrain when people talk about this is "cheaters only cheat themselves" and while I think has historically been mostly true, I think it's bullshit now. Cheating is just too easy, the folks who care are losing the arms race. My most impressive peers are struggling to get past the first round of interviews. Meanwhile, the folks who don't show up to class and casually get perfect scores are also getting perfect scores on the online assessments. Almost all the competent people I know are getting squeezed out of the pipeline before they can compete on level-footing.

We've created a system that massively incentivizes cheating and then invented the ultimate cheating tool. A 4.0 and a good score on an online assessment used to be a great signal that someone was competent. I think these next few years, until universities and hiring teams adapt to LLMs, we're going to start seeing perfect scores as a red flag.

I'm hoping the text editor + project directory approach helps force ML projects away from a single file and towards some sort of codified project structure. Sometimes it just feels like there's too much information in a file and it becomes hard to assign it to a location mentally (a bit like reading a physical copy of a tough book vs a kindle copy). Any advice or thoughts on this would be appreciated!

One friction point I keep running into is how to handle logging and evaluation of the models. Right now I'm using Jupyter Notebook, I'll train the model, then produce a few graphs for different metrics with the test set.

This whole workflow seems to be the standard among the folks in my program but I can't shake the feeling that it seems vibes-based and sub optimal.

I've got a few projects coming up and I want to use them as a chance to improve my approach to training models. What method works for you? Are there any articles or libraries that you would recommend? What do you wish Jr. Engineers new about this?

Thanks!

Comments URL: https://news.ycombinator.com/item?id=45929871

Points: 3

# Comments: 2

> Get great, if not perfect grades.

I think this was great advice five years ago and just no longer works for the AI era. I'm back in school to get my masters and every single one of the best students is struggling to break into the A- territory, let alone get A's.

Cheating has simply gotten too easy.

I think it used to be that any class was separated into four groups. Best grades (A's) went to the best students, the good students got B's, and then there were the folks who were struggling. Some of these struggling folks would cheat, and sometimes do well, but for the most part they were in the high C-low B range.

AI has turned that on its head. Curves do not exist anymore. The cheaters get straight As on every assignment (tests included), the great students get Bs and the good students struggle to pass the class.

A few weeks ago I had a professor tell me that I did amazing on a test. My final grade was an 81% (a failing grade in a masters program). When I asked them what they meant saying I did well they told me that, of the people they didn't suspect of cheating, I got the highest score.

My advice is to do all the things that she listed and, whatever you do, don't focus on grades. It's a sisyphean task. Find what you enjoy in your courses, outside of the too, and spend time doing it. Crush any presentations you get. Find what makes you happy. Just, for the love of god, don't focus on your grades.

> 2. His resume is designed poorly… This is the world of TikTok and Instagram reels

Imo this is exactly the problem. We’ve constructed a system where brilliance doesn’t shine through. The idea that someone as thoughtful as OP needs to tiktokify their resume to even have a chance at getting hired is ridiculous.

I’m young, so I have no clue, but surely the job market didn’t always work like this?