Do Coding Boot Camps Make Sense in an A.I. World? | Hacker News Reader

Do Coding Boot Camps Make Sense in an A.I. World?

nytimes.com

To me they never made sense in the first place - outside of ZIRP era that is which was very much a fluke. Something that someone from the street can pick up in a couple months isn’t likely to command an enduring market premium.

Doesn’t help that most of them taught precisely the things LLMs are good at. Boiler plate front end and a sprinkling of glue together back ends.

To the extent that they could be a somewhat reliable pipeline into a programming job, they made sense for people with reasonable programming skills (or perhaps capacity, if we're generous) and a lack of credentials.

Can do the work, but can't get hired? Find a reasonable boot camp (hard), do the time and get access to their placement assistance. From there, now you've got work experience and will have an easier time getting through hiring pipelines.

I think they were definitely oversold as the solution to everyone's lack of a good job, and some of them were outright scams.

I've hired a couple of quality developers from bootcamps. Especially if you're looking to hire Junior devs, I highly recommend keeping an open mind about these bootcamps if you're a hiring manager.

I went to a 4-yr university for a CS degree. I mean I did learn a lot and I don't regret it, but tbh I didn't learn any web dev languages or most things I use at my job today through my program. I learned C, C++, etc. which was super interesting at the time but it just doesn't translate into JS/React-world super well. I think there's a place for legit bootcamps that focus on what you'll be using in your day-to-day and connecting you with some potential hiring companies. They just need to be careful to not guarantee anything and I think they could benefit from almost an internal hiring round after the program to see if you have the skills to even be recommended for a job. That way hiring managers could build more trust with the bootcamps.

> I learned C, C++, etc. which was super interesting at the time but it just doesn't translate into JS/React-world super well.

The technology you learn is not really the point though, it's all of the fundamental underlying principles

Have you ever worked with someone who has no concept of Big O? I have. They wrote almost woefully inefficient code, and actually had no concept that code efficiency was even something you could reason about

This is not a terribly difficult thing to teach to a Junior, but it's just one example of things you learn in a degree program that bootcamps generally gloss over

I've met some excellent bootcamp devs and I've met some absolutely horrendous 4 year college CS grads.

It also highly depends on which bootcamp you hire from as some do do push on CS fundamentals using Javascript as the language to teach.

The top bootcamps filters out a lot of unserious, low quality candidates for you already. In addition, it's not a 3 month learning experience. It's 3-6 months of prep work, 3 months of actual bootcamp, and then likely months of practicing leetcode for interviews. It can take a year.

Holy fuck, hope I never come to depend on your company's products.

To back up your point, I taught myself sans boot camp. I got my first job 4 years later. Although I was ready much sooner, it was tough, and I lost motivation a few times.

When I finally got hired, somebody applied who had started coding the same time i did, only she had gone to a bootcamp. All the time I had been looking for work, she had been working. Not saying I’d do anything differently, but no question bootcamp would have shortened the journey.

You are missing out on the huge world of self-taught developers who run many of the massive companies out there. Many of them dropped out too.

Learning informally, or formally, in person, online, in groups is all valuable.

Someone who is self-taught is likely to have an interest in the topic of programming. Most bootcampers I've come across have no interest in understanding the fundamentals.

In my experience, the former produces much higher quality developers and it's not even close.

Bootcampers can absolutely learn the fundamentals.

I'd put that on the material being taught, how it's organized, and being presented.

Why? Teaching adults is very different than teaching children, and universities mostly teach how to teach children as a degree program.

In terms of teaching adults, there are ways that have existed for hundreds if not a few thousand years at least: apprenticeship, learning with and beside someone who knows how.

Pro-tip for all bootcampers: run thru something like javascript30.com first.

A bootcamp can be someone’s best entry into the field if they’re later in their life where a 4-year institution would be difficult to swing.

The average lower bound for the quality of a bootcamp graduate is likely lower but the upper bound is limitless.

And I’m saying that as a electrical engineering graduate who knows some basic things that some CS grads don’t know that they should really know better than I know.

That's ridiculous. If you can't afford actual education, settle for a scam? Wtf?

> If you can't afford actual education

For working people it's not necessarily the cost of a degree, it's the opportunity cost of time out of workforce. Rent/mortgage still needs to be paid.

> settle for a scam?

Yes, boot camps are a "scam" in a sense, but when done right they are a happy one. The service they are providing is legitimacy washing. It is working around credentialism in the hiring process by providing a baseline credential or ticket to participate. A company that wouldn't hire a "self taught programmer" will often be more inclined to give a "boot camp grad" a chance. Really it's the same person, just with 100% more rubber stamp.

It is VERY rare in my experience that you're going to take someone with no aptitude or experience in software development and make them employable in a few weeks or months. This part is oversold. But targeted training on in-demand skills (e.g. the modern JS horror show) can be quite beneficial for people who already have some foundational skills.

Its still an mostly American phenomen. I've met plenty of people in my career who didnt learn CS and changed in the field later, mostly trough self education and interest.

In Switzerland if you make a second degree it's usually shortened and better payed. Plus often sub-financed by the state. I made my programming degree with people between 19 and 50.

Never did I meet anyone proclaiming doing a private certificate bootcamp thingy.

> Something that someone from the street can pick up in a couple months isn’t likely to command an enduring market premium.

There's a large gap between your local market value and that in a poorer country. On the Balkan, a nurse will have a terrible salary, but if she goes through a bootcamp and learns enough to be useful in QA, she'll quickly make 2-3x what she would earn working in a hospital. There's a pretty hard ceiling when you approach Western income levels, but everyone I talked to is super happy hitting that ceiling, because it'll mean they have great income compared to the national average.

Nah, I advocated for hiring someone out of a bootcamp who was previously a dancer. She ended up being a great addition to the team, was an excellent communicator, and was a significantly more clever and harder worker than others I’ve worked with who came from elite schools.

I sometimes think this attitude that you can’t teach good programming practices quickly on the job comes from folks who have never tried. We literally have code review practices for this reason.

That’s a great point I hadn’t thought of. There are a lot of things built into the coding process that are designed to support and develop junior engineers. Why shouldn’t bootcamp developers be successful?

As someone who is at management level, for most dev roles, give me someone intelligent with a really good attitude over a dev who is full of himself but highly technical.

"Between the time Mr. Rendon applied for the coding boot camp and the time he graduated, what Mr. Rendon imagined as a “golden ticket” to a better life had expired. About 135,000 start-up and tech industry workers were laid off from their jobs, according to one count. At the same time, new artificial intelligence tools like ChatGPT, an online chatbot from OpenAI, which could be used as coding assistants, were quickly becoming mainstream, and the outlook for coding jobs was shifting."

The big question for me is how much of the reduction in available job for boot camp graduates is because of AI-assisted programming productivity boosts, and how much is because they are now competing with 100,000+ more experienced people who got laid off in the last couple of years (due to a retraction in market size after an over-exuberant hiring period by tech companies during the pandemic).

Anyone seen any credible studies about that?

For experienced developers, copilot and friends seem to be a slight win. (Maybe 2x one day out of five.)

That’s as not as big a boost as, say, IDE autocomplete.

However, for entry level Lego style coding, I can imagine it’s a bigger improvement.

I’d be curious to see a study that breaks productivity gains out by task. (Including learning new languages/frameworks/code bases).

My prediction is that the sorts of jobs coder camps target will get hit harder than most, because most productivity gains will be there. That probably implies that type of development training will need to focus more on higher level tasks, like incorporating requirements or making tactical design decisions.

I 100% agree that this is the right question. My intuition is that higher interest rates, lower multiples, and layoffs are the cause. AI is just what we point to to give ourselves comfort that our smaller teams can accomplish as much as our bigger teams used to.

I would put gross overhiring (and overpaying) which caused a positive(/negative) feedback loop and fomo ahead of the other factors. The pace and scale of hiring wasnt even justified by the rates / multiples which had been low (rates ) and high (multiples) for the best part of a decade.

Seems hard to do a credible study on that. You'd need hiring managers to answer "how many more bootcampers would you have hired in a counterfactual world without AI", which seems very hard and subjective.

Intuitively you'd expect AI to be a big win for bootcampers. Their productivity is typically low because they're constantly hitting roadblocks and getting stuck, even on simple tasks. Unsticking them is an AI strong suit so they should be able to plough through tasks much more effectively despite still asking for the same low starting salaries.

Problem is, AI also benefits the more experienced types, even if less. And if you can pick between a desperate senior or a bootcamper, and the price gap isn't that big, why would you go for the bootcamper? The senior will still be much easier to work with even if they both would have access to top of the line models.

So it's got to be the layoffs (ending of ZIRP and COVID stimulus).

Microsoft is not going to subsidize AI forever, they'll need to make a profit, and last quarter the spent $19B and only earned $2.5B from it. People are cheaper in the long run, probably forever.

it will be people + machine as it always is. For instance, modern robotics hasn't made factories unmanned, instead it's just a small fraction of what was once needed. Same thing is true for mining. Over 90% of the jobs will vanish and they will be different. In the 1800s, the professional photographer replaced the portraiture painter. Different pricing point, different skill, comparable product.

But none of these 100% disappeared. Ai is not a buggy whip to programming. It's more of a camera to painting.

I feel like "AI taking programming jobs" is just yet another in a long list of hyped up things that so many people seem convinced about, but then never ends up happening.

In X years from now are the people that pushed this idea going to reflect and change their approach to things? I doubt it, they'll just have moved on to hyping up the next thing.

Off the top of my head, hyped up things that everyone "knew" that have yet to happen:

- Driverless cars. 10-20 years ago they were all the hype, my friends thought that in a year or two they'd have a driverless car, and now it's 10-20 years later and they can still not buy a driverless car.

- Bitcoin hype. All the people that were so convinced that bitcoin would take over, and people would be paying for groceries with bitcoin, and everyone would switch to bitcoin for everything and it'd be a huge revolution .. and it has yet to happen.

- NFT stuff. I feel like I don't need to explain this one.

- Another current one is "just keep scaling LLMs and AGI will magically happen, we're really close". Related to AI programming I guess. Check back in in 10 years on this one, it ain't going to happen.

- Various short-lived hypes where a bank closes down and everyone on HN and elsewhere is spreading doom about how it'll be a chain reaction and the whole economy will crash and .. it never happens, and everyone just forgets what was said.

It's just hype after hype after hype, it's exhausting.

printing press was hyped radio was hyped tv was hyped planes were hyped the internet was hyped

things arrive slow, big things take time

btc is almost 100k today i can take a waymo right now

i agree in between a there are letdowns and scams, but not all is hype

hype after hype is not exhausting, it’s exciting

Today: "We can save money! We don't have to hire all these junior devs!"

2035-2045: "Why does our software suck? Why is it so hard to find good senior devs and architects?"

"Why are we having to pay 10x what we paid a couple years ago for a senior developer???"

I hope so.

Also possible: "Wow people are still paying good money for our sucky software. I guess we never needed good software nor good developers."

"Why has our built-with-AI code base become unmaintainable?"

Non-paywalled link: https://archive.ph/yYVus

Opinion:

I am extremely tired of this same old bs about “almost 60% developers use AI”. Yes, 60% developers use AI because the previous glorious Google/Bing has drowned in slop and now my search for documentation of a particular library or an arcane question returns 100% SEO spam, so I need to ask claude(or my local llama running on ollama) and get a decent enough(albeit outdated) valuable result from which I can pursue more from somewhere. At no time does it mean that, AI is doing a significant part of my work, and no my IDE already generates all the code I need and no I am not going to blindly use the rote replication of the example from StackOverflow that the AI just generated when I said it to generate a particular function, because these examples are often not the best implementations.

The AI hype and the desperate lobbying is doing a lot of harm to our industry as a whole. We should see it for what it is, some rich naive dudes invested a lot of money because they feel FOMO from last gold mind rush if AirBnB/Uber/Netflix/Spotify era and is freshly burned from crypto dud hype, so they want to kill the industry by scaring away new folks and make everyone as lazy so we get slow frog boiled to accept garbage from these AI crop.

Also, the market and macro economy is unstable for almost 5 years now due to ZIRP expiring, random war, uncertain geopolitical power struggle, failed regulatory keep up allowing a handful of tech predators to acquire and extinguish new ideas and competitions and so many factors. It is just an uncertain time, money now flows on hype and not so much on good ideas because previous good hype has paved way to too much greed, and of course certain giants built too big of a moat and raised the bar and now the new management is milking the cows until the cows collapse in next few years when the new competition rise once the macro settles a bit.

I am optimistic, from next year, we will see a massive improvement in prospects(despite trump happened). We just need to hold on a bit.

Reminds me of quote Sussman gave when asked about MIT switching away from Scheme:

"Nowadays you muck around with incomprehensible or nonexistent man pages for software you don’t know who wrote. You have to do basic science on your libraries to see how they work, trying out different inputs and seeing how the code reacts. This is a fundamentally different job, and it needed a different course."

>I am extremely tired of this same old bs about “almost 60% developers use AI”.

I don't even believe that is accurate unless you're considering things like IDE autocomplete to be AI or assuming everyone using visual studio user is using copilot.

Or, you can look at the positive contribution that AI is adding for developers. This is true even for curmudgeons like me that tried it early on and had bad experiences. Using AI to look up documentation, and asking the AI to summarize alternatives takes many fewer clicks than searching Google. I also find it generates good code if you ask the right questions. That is, as an experienced programmer, I'm able to generate much better, iterative prompts than a nontechnical person or a junior programmer. With tools and the AI evolving, making iterative changes across your app using an AI is clearly part of the future of programming. Heck, it even chooses good variable names, and that's been the bane of programmers since the dawn of programming.

Learning to code, both in person, and online, both formally, and informally are invaluable.

Agreed. The more unique situations I program in, whether it is new people I am building software for or whether I am building on a new tech stack, the more I realize why so many packages and solutions exist in pip and npm.

Coding new solutions in new ways has helped me learn so much. Getting a career that supports that was one of the biggest catalysts for me (unsurprisingly), however, attempting to code anything is certainly beneficial.

Getting started in any way, keeping going, keeping growing is the key.

Self-taught developers are equally as formidable and can be found at or near the top of any company and required on balanced teams with “formally” educated developers.

Different styles of thinking, both are required on a successful team.

The inherent nature of the technology industry is that we all always have to be levelling up and finding new ways to create value. The better your ability to reason from first principles, learn new tech, etc. the better you will do.

Teaching people a specific approach and tactical skillset without the underlying fundamentals does not set them up for success in a fast moving industry, particularly right now when we're undergoing an enormous shift on virtually every level. But it _can_ be an accelerant for some people, so I'd never discount a bootcamp grad on the basis of their learning method alone.

Do CS degrees make sense in an A.I. World?

No, nor does it matter that it's not A.I. nor that it's not fundamentally useful for replacing humans. What matters is cost reduction for businesses who leverage the technology. Why bother hiring six SWEs when you can hire three and make them clean up a "prompt engineer"'s work?

Of course they do. It doesn't matter if it's independent self-study, a bootcamp, or a formal degree. The goal is to iteratively build on your knowledge, grasp the fundamentals, develop intuitions, and put yourself in a place to continue to build upon that foundation in perpetuity.

It's a starting point. Not a stopping point.

This bootcamp grad now works on AI (LLM augmented services as well as traditional ML models) on a daily basis at a big tech company.

Reply on news.ycombinator.com