Hacker News Reader

The top student in any grad class is probably still going to make a ton of mistakes compared to even an average dev with 10 years of experience though.

There really is no substitute for actually making the mistakes and learning from them

The average devs with 10 years of experience also make ton of mistakes, even when developing the 10th project. Projects are not all the same. I've met people with 10+ years of experience, labeled as seniors who write horrible code.

I'd take a smart junior who can think logically and who is passionate about the job over someone who has 10x 1 year experience and doesn't want to learn, or is just not smart enough. Chances are a good junior can beat such senior after a few months (sometimes even weeks) of experience on the project. And the junior typically won't fight with the tech lead over something they "have always been doing" (even when it was an anti pattern).

Anyway, it is good to have at least one senior dev who is an expert on the tech stack / architecture used.

It doesn’t depend on the student. Every developer starts out as a dumbass. There are an infinite number of sharp edges in software and we all cut ourselves on all of them.

Couldn't agree more. The only reason i'd call myself employable as a developer is because i've taken on many self-projects and reinvented many wheels, which resulted in lots of sharp corners and hard lessons learned.

It's a definite time sink to learn these lessons, but it's of real value to me personally.

"starts out as" 10 years in and I'm still a dumbass... I work circles around most and I'm still a dumbass... I'm close to being honestly a full stack developer - server, database, api, windows services, websites (I know Asp.net MVC... but not deeply... finally learning Angular)...

... I'm still a dumbass...

I wonder... if after 10 years... will I stop being one after 20? 30?

Honestly the feeling hasn't completely gone away but I've slowly stopped thinking of myself this way as intensely as I did before. What it took was strong mentorship from a really good boss and exposure to all levels of the stack and strong architectural and technical principles to building robust applications. Taking some interviews I felt much better about myself than I had in years cause I was actually able to answer things that I couldn't have a couple of years ago

I hope you find that place in your life too. It's a lot of relief when it happens

> ... I'm still a dumbass...

While that may be true, you're less of a dumbass now, thanks to experiences. If you were to guide a junior developer now, you'd probably walk them back from a dozen sharp edges, thanks to your experience, but we tend to discount that.

i feel also with time, you get better from recovering from these stupid mistakes faster and also come up with better solutions/fixes faster.

This will have taken your younger self , twice or more to understand the problem and come up with solution

You also getting better at owning mistakes without making them sound like a big deal. You just say stuff like “that’s an interesting edge you’d only learn by querying a table with millions of rows” instead of anxiously shitting your pants in shame.

Last Friday learned that Postgres (and apparently MySQL) will effectively ignore an index in a column when combining “ORDER BY … LIMIT …” and “WHERE some_column IN some_array”. I had introduced a query that timed out on a critical prod table , fixed it by removing the limit, and turned it into a teaching moment to others who could have made the same mistake: https://stackoverflow.com/questions/8909163/avoid-full-table...

No. 25 years in here. You may think you're a dumbass even then. Other people may think you're a dumbass. You will still make stupid mistakes. You will probably make fewer than other people, and there will be many mistakes you don't make any longer, because you've already made them and have learned your lessons.

But the feeling of being a dumbass may never leave you entirely. You will need to learn to control that some, and not always let on to everyone around you how dumb you feel at all times. But... honestly, the folks I've worked with who never ever admit to feeling dumb or doing dumb things - they STILL make the same mistakes, often more, and seem to have no capacity for self-reflection or skills improvement. So... understand your "I'm a dumbass" feeling may just be part of you pointing out that you can get better, and help you find opportunities to improve.

Well, I honestly use the feeling of being a dumbass (AKA: imposter syndrome) as fuel to learn more and do better.

As you say, self reflection and skill improvement.

I honestly think it's why I'm a good programmer...

That... and being lazy. I'll work for weeks to make processes easier. IE: a release process that used to be manual and fraught with mistakes is now, more or less, button clicks and powershell scripts.

I'm almost two years into self-employment and I'm wondering when I'm going to either stop cutting myself on things or land a position with some people who can show me their cool scars for a while.

This forum is the closest thing I get to coworkers and I don't even understand what half of you are saying. I'm fortunate for the opportunities that I have, but my antisocial tendencies and lack of professional peers can make progress a bit of a bear on occasion.

It never ends, unless you stop advancing/keeping up. That's the frustrating thing about this field, you are never an expert because the technology advances so quickly.

It's just as important to develop the soft skills as the hard skills, and I think you've just quite nicely summarized why that is.

I'm in a very similar situation, just a few years further down the road. I would gladly show you my cool scars.

I have often thought about trying to build a network of like-minded individuals. Of course, it doesn't come naturally. If it did, there wouldn't be a need for it!

My contact info is in my profile, feel free to reach out.

I'm somewhere between two and three decades in the field.

> when I'm going to either stop cutting myself on things

Never. You (should) get better at recognising some of the themes and danger signs, but the field evolves so fast that at some point you are going to work with something so new that literally nobody knows what the sharp edges are.

> land a position with some people who can show me their cool scars for a while.

Meetups, slack, discord, IRC, forums. Find the one that works for you and be a regular. Be an expert in one subject and be modest and willing to learn when it comes to others. Nobody has it all figured out.

I like your spirit. Welcome to the party.

They may have extraordinary potential, extraordinary problem-solving ability, extraordinary communications skills, and extraordinary critical thinking skills. They may be tremendously resourceful and almost entirely independent.

But all of these qualities only accelerate the learning process that happens in the real world. They don't replace it.

Edit: To be fair, given the choice, I'd rather have an MIT CS grad than some schmuck who barely squeaked through the MIS program at Big State U. But I'd take a good experienced dev over either of them.

I don’t think it would take long for the MIT CS grad to start out performing this “good” experienced dev. The experienced dev could be a principal engineer by 20 years if they cared more about CS.

Most projects don't require some deep CS knowledge so even if the MIT CS grad was great at getting up to speed they still lack experience of edge cases and gotchas. Most companies don't have Principal level roles.

Might still make sense because they're cheaper.

What makes you think so?

Not everything is written in books or taught in classes, even the most accomplished genius is going to have gaps. Depending on the person and they're penchant for studying outside of the curriculum, those gaps could be relatively smaller between people, however it's not going to erase them. Someone who has already built a live system, potentially multiple times, is going to have experiences you cannot just intuitively reason through, you must live them. You don't need a majority of folks like this on the team but one or two could save you a lot of wasted time and effort.

I agree that recent graduates can be extraordinary, but a team with no seasoned developers is still a major red flag to me.

I've worked on many projects, some successful, some not. The difference was often that the successful projects had a critical mass of wise people who had learned from past failures and could anticipate the pitfalls and avoid them -- it seems more rare that actual project success comes down to having yet another brilliant programmer.

I'm not sure why this is being downvoted. There's no evidence that the kind of instruction that e.g. Facebook gives its junior developers is meaningful for performance on average. However I bet hiring from top universities is definitely meaningful for performance on average.

It isn't the instruction on the job that is important, it is the experience of actually doing things in anger

This is probably true, top university admissions is a function of innate ability, and programming is too apparently.

A junior developer is going to have a lot of blind spots because of unknown unknowns. A team with both talented junior developers and experienced senior developers can utilize the raw ability of the junior devs while having the senior devs cover the blindspots. A team with only talented junior devs will likely make some great things but they’ll also wind up having crucial problems that will only late into the project.

I know exactly why you’re being downvoted: most developers identify as the average experienced developer and can’t handle the fact that they can be outperformed by an extraordinary rookie (although such people usually have a lot of experience from working in their free time).

I don't think that was the point. Rather, it's that many rookies, including bright graduates, think "meh, I can run this SQL script without making a backup first."

"most developers identify as the average"

That is what average means

Let's make this more obvious: Imagine that in you have to produce Nitroglycerine, explosive set off by the slightest of mistakes.

Would you rather it done by someone superbly clever, but straight out of uni, or someone who's average, but has done it for 20 years and survived?

Well if you are happy that you found someone that can follow a checklist then fine. I doubt the top tier grad would want that type of job anyway.

Do you know how many hundred top tier chemical and nuclear scientists died while "following the checklist"? Look up Demon Core

1, you are now really exaggerating. 2, this analogy is really straying far from software engineering. 3, what is your point here?

IIRC the work with Demon Core was described as "tickling the dragon's tail", not "following the checklist". Wikipedia also mentions "Slotin's own unapproved protocol".

Also how could there be a checklist? They were literally pioneers

An easy way to create one would be to ask "what could go wrong" and "how bad would it be", then develop procedures to minimize the risk.

You are right, it's a different point - just trying to illustrate that being some of the smartest people on earth doesnt mean you wont make a terrible mistake or will have good judgement.

Like they knew it was super dangerous, there wasn't new science discovered since then about this particular setup. Today rank and file engineers know not to do that

Not calling you out, but where did you get the number deaths > 100 from?

https://en.wikipedia.org/wiki/Criticality_accident

Problem was, there was no real checklist. They were part of a huge number of people who have created the checklists, partly as a result of lessons learned after making mistakes resulting in death and destruction.

Out of interest with your name Google234, are you a student or early employee? You seem to be very certain about things in your comments here but I'm unsure if they come from experience or speculation.

Most senior engineers expect to be out paced by juniors in terms of raw throughout.

But what the seniors have over the juniors is several years of experience making mistakes.

“Raw throughput” like what? Mashing on the keyboard? Then - sure. Sorry but that’s just not true ime. On some tasks jrs are even infinitely slower than srs because they just can’t make any headway at all

On some tasks that’s true. I was making a generalisation rather than defining an absolute.

I’m having troubles imagining how more experienced eng could be slower than junior provided their code is reviewed and tested to the same standard. Unless the task is so soul crushingly rote that your senior is looking for another job rather than working on it ;)

Raw throughout isn’t the same as speed.

The senior might spend more time doing PRs (juniors might not feel as comfortable), more time in meetings discussing new projects or architecting around arising problems. A junior is generally more enthusiastic so might voluntarily work longer hours to close of tickets, or just take fewer coffee breaks than a senior. A senior might also spend a portion of their time training juniors.

I’m not saying a senior engineer can’t be a prolific at closing off tickets as a junior, but more seniority means more responsibility and that often means less time writing code.

But there are also occasions when I might genuinely expect a senior to be slower. Like if they’re in their mid to late 30s and have young children who aren’t sleeping. People say you shouldn’t bring your home problems to work but the reality is that we are not robots and if you’re not sleeping at home then you’re performance at work will suffer. Whereas most 20-somethings have fewer commitments outside of work.

If we are talking really senior, like 60s, then a sad fact of reality is our mental faculties aren’t going to be as sharp as they were when we were in our 20s. So it might take an engineer a little longer to solve complex problems than someone fresh out of uni. Particularly if it’s an academic problem that one might have studied at uni but seldom need in your career.

So there are plenty of mitigating factors that might affect the throughput of a senior engineer. But that doesn’t matter but you hire them for their experience rather than their youthful motivation. And it’s also why it is important to have a balanced team of junior and seniors.

The hinge of this analysis is on the word "outperformed". Performance of what you're measuring exactly? If our sole performance criteria was churning out LOC and subsisting on adderall, any new grad can kick ass. But you wouldn't entrust them multi-million-dollar man-hours of headcount to a steer multi-quarter project to actually work and bring in money. Without such steering, any code is mere dead matter, for which the performance of making is irrelevant.

"A stupid person can make only certain, limited types of errors; the mistakes open to a clever fellow are far broader. But to the one who knows how smart he is compared to everyone else, the possibilities for true idiocy are boundless."

Few things as frustrating as coming back a year later to uncommented code written when you thought you were being smart.

One thing that is slightly more frustrating is saying “What dumbass wrote this code? How did this ever work?” And then looking at git blame and realizing the dumbass is year-ago you. (Or even five years ago you.)

Or sometimes six months ago…

There're very few extraordinary students who can start contributing from the start. But those students are already stars of their own, not the kind you can hire. For the vast majority, there's a very long journey from graduation to professional developer. This applies also to PhDs, and yes, also from the best universities in the world (I have those in my team often). A company that only hires junior graduates as devs is a company that doesn't take software seriously. Such companies often collapse under their own weight

There is a reason years of experience matters more than class GPA or school rank. Being a good student in many cases just means being able to cram for exam week, it says nothing about your consistent habits. Software is a marathon not a final exam.

GPA doesn’t have anything to do with being a good student

I've worked with all sorts of brilliant people and genuine 10x developers, and it's true both that students can achieve great things and that anyone can make a mistake - and that low-quality work environments like that of the OP can force anyone to make mistakes.

Others have to fill your gaps. I have no problems telling people their methods are wrong even if they might fire me over it. We either fix it or I'm gone.

> although there is a lot they don't know

I think that's exactly the problem. Re top students I assume we're talking about things like a deep knowledge of algorithms, implementation details of relevant languages, the latest frameworks, being comfortable writing in an object oriented or functional style, being extremely smart, having a lot of working memory, loads of energy and enthusiasm, being able to grasp complicated concepts quickly etc (I've never met a recent grad who is even close to that, but I understand it's at least theoretical).

But experience lets you do things like push back a little on the requirements (because you know that what is being asked for is not actually what is wanted). Or choosing path A instead of path B because you know that path B will come back to bite your team later because you've seen how it played out a few times before. Knowing which things will make the biggest difference to the speed or quality of delivery, whilst costing the least to implement. You know what problems you should ignore and what problems are serious risks. Things like that can sometimes make the experienced developer a 10x programmer, it's certainly not because they're typing 10x faster than anyone else.

I agree with your assessment.

“Experience is what lets you recognize a mistake when you’re in the process of making it a second time.”

Despite the deluge of downvotes, recent grads and dropouts formed many of the worlds largest companies, and many you work for them or for their successors. Linux was started by Linus while in school as well. I don’t know why this is getting so much hostility

Bill gates Michael dell Zuckerberg Jobs Dorsey (Many others)

I’m sorry this makes so many people mad but… stop looking down on people

Also, why are people conflating high gpa with being a top student? Top student means “the best at creating software”

https://en.m.wikipedia.org/wiki/History_of_Linux

> although there is a lot they don’t know

That "although" is the whole point.

Knowing the technical basics (language syntax, getting things to work) is the easy baseline. What's missing in every studente (unless they went back to school after a solid stint in industry) is the experience of understanding the processes and approaches required to build something lasting and to operate within a larger team.

An app like the one the author of the article is talking about doesn't need "extraordinary" coders. The reason for the crash was that someone entered invalid data. That is a very simple, and very common, problem that you only start to understand with a bit of experience. You don't need to be a brilliant genius to understand it. Literally every dev whose worked on a few apps can deal with it. But you do need to have worked with real QAs and real customers in order to understand that people do stupid things that you need to protect against. No one is teaching that in a compsci class. Juniors learn that from seniors. If you have no seniors, demos failing on "obvious" things is exactly what happens.