Ask HN: What will be the most valuable technology skills in future?

55 points by kobotigo ↗ HN
Young person, a bit lost, struggling to get anywhere. Not really qualified for any work or able to afford degree. Had to drop out of college and have a year long unexplainable work gap after mental health decline. Have a generalized knowledge in tech from linux use, tinkering with networks and have a programming admiration but no idea how to translate this to a marketable skill set.

If you were in early 20s and had 6months to a year to try and get on track. what would you focus on? IT, programming, cloud, what is the optimal vehicle and will be valuable in 5 years? Is IT or programming safer bet?

57 comments

[ 0.26 ms ] story [ 73.5 ms ] thread
I'd do a mix of sales and coding
> translate this to a marketable skill set.

Don't. Learn soft skills. Put all that tech in the cupboard.

How? Man, I don't know, but I bet there's plenty of material out there. Maybe get a job in a service industry that kind of overlaps with your interests.

Get to know people. Skills can be outsourced. Lovely, empathetic, and patient people (with a backbone) cannot be outsourced.

I understand where you are going with this comment and I partly agree, but it depends.

For example, I'm in a startup producing hardware products for industrial use: who do you think got fired in the last few years, 1) the one guy who implemented all the low level stuff and knows the product inside out, helps sales and support with technical knowledge, fixes bugs at any time, or 2) the nice sales people who could not really deal with the complexities or were not persuasive enough to make enough sales?

Food for thought.

Fair enough.

However, I was assuming OP would balance their soft skills with their tech skills and aim more for management and eventually a CO position. IMHO softskills(tm) mostly leads to sales only and, funnily enough, a lot more opportunities than a ossified tech stack.

Resign yourself to taking the abuse of cold calling, develop your salesmanship, build a network, then try to get a real job.
If you like tinkering, admin, and generally lower-level or systems things... you might look into DevOps, Cloud, SRE, and Kubernetes topics. There are lots of certifications and educational content online. That particular area is in much higher demand,.fewer people enjoy it

Knowing some python and bash is very helpful

The age-old certs debate: Have you noticed tangible benefits from a cert? I am considering getting one, but purely to increase my earning potential. I tinker on AWS all the time for work, but have 0.
I'm not a hiring manager but I think if you're early in your career, a cert wouldn't hurt and may distinguish you from other candidates. If you have 5-8+ yr exp, it probably matters less.
Yup, this for sure. Though all you really have to do is demonstrate what you know before & during the interview. You can certainly put together a hire-able portfolio without certs or job experience
Agree. Learning AWS and especially scripting things with CLI / using terraform is a good start.

AWS is a huge area with a bazillion of ingredients but you could start with learning: S3 buckets, permissions, lambda functions.

I'm a Cloud/DevOps eng, and feel like DevOps is at least somewhat vulnerable to AI/automation. I think "architect" types may be safe, but all of the hands-on knowledge about wiring up different systems seems to be ripe for automation. My guess is in 10 years time, DevOps is just absorbed into backend engineering at most companies.
I wonder if your optimal path would be writing the stuff that's going to replace your job?
I just transitioned to doing MLOps for our deep learning team, so yeah I'm kind of already on that track :)
There is a lot of value in having dedicated people thinking about operational efficiency, cost controls, developer productivity, build correctness, security posture, what technologies to even use...

"DevOps" can be a lot more than configuring AWS or Kubernetes and SRE'n those systems. I don't think it wise to fully absorb it into the development teams, the ops team enables them to do more and has cross team visibility to create opportunities that aren't seen at the team level.

I would expect backend development to experience as much automation as operations. There is a lot more training data for backend compared to operations as well

Yeah, I think this'll be a case where like >50% of it gets automated and salaries drop. Like you said, big cos are always going to need people thinking wholistically about this stuff, but I do think we're exiting an era where every small-to-midsized startup needs a dedicated devops team. I also think availability concerns as a whole is a bit of a bubble. Most news sites don't actually need 4 9s of availability.
If you are really interested in programming, C# is not a bad way to go. You could do desktop, web, or mobile apps. Work in the Cloud. Learn data access. There's a lot more to know that just C# syntax though, and many possible paths.

I recently watched a video describing a decent learning path: https://www.youtube.com/watch?v=hnjly9iBHso

And Tim Corey has a lot of good, free, videos on YouTube. But like he says, you really have to sit down a practice. You can't just watch videos all day long. It's one thing to watch an NBA game to see how basketball is played, but it's another thing when you get on the court and do it yourself.

And by practice, you don't need to come up with an idea for an app. Just create a project to test the one skill you are trying to learn. Want to understand foreach? Create a project for it. Want to understand switch statements? Create a new project just for that.

Good luck!

I am interested I think. It is hard though to differentiate occasional fascination and making/editing very basic python programs, From an actual willingness to do the meticulous process and the aptitude to contribute to complex programs.
Before I even clicked the YouTube link, I knew it had to be Tim Corey.

#1 Youtube channel on C# out there imo.

The fact that I knew before just goes to show how well known he seems to be in the C#/.NET community, is what I'm saying.

Security isn't going anywhere anytime soon. If you think about the timelessness of war - how those creating weapons always seem to profit, it's kind of like that. Basically an internet defense soldier, and the demand right now is pretty much sky high. Places seem desperate enough in this space to actually train people unlike the software space.
Just look off Indeed or anywhere specific? Asking for OP.
Cosigning security as a stable job, but also noting that security at the vast majority of companies is tremendously boring. Usually, it means you're a SOCII check-box checker. Pays well though and, like the parent says, it's only becoming more important.
security, absolutely. I moved, partially, out of SE by title because of it.
I will say entry level security positions can truly suck, and you may be treated with a lack of respect in the org. That said the pay is good, and as the demand goes up I think the positions will become better.
Actually just getting a job helps immensely. Look for intro jobs like Help/Service desk positions and do the IT/programming on the side.
What are the odds of landing help desk job without any tech credentials? Making it past the resume automated culling I mean.
It's possible, if you can speak knowledgeably about computers and know how to troubleshoot, build, or fix them. If you know how to troubleshoot a "no video condition on a pc" and can speak from memory about the troubleshooting path you would take. You have better odds than someone who can't.
IT support -> write useful internal tools -> SWE is a path I have seen many take. The first role would be best for a small/medium company that hires based on what you can do rather than your credentials. Has to be big enough for a FTE but small enough that the interview isn’t too barricaded by Hr.
By valuable, do you mean what is actually valuable to building a project, or do you mean marketable in the job market?

If you mean actually valuable, I think the most valuable skill in the future might be the same as it is today: being good at considering the tradeoffs to different technology choices and design decisions and putting it in context with how you actually solve specific business problems intelligently. The best Java developer in the world is not as valuable if he doesn't understand business at all and doesn't question and even push back on the requirements and spec he's given once in a while and come up with some simpler ways to solve problems other than sitting down and churning out the thousands of lines of code he's told to.

If by actually marketable, in a 5 year timeframe, something like Blockchain programming might end up being in very high demand. As interesting as all of these AI things are, I wouldn't trust them for anything except toys for quite a long time. In 5 years, I'd think centralized digital currencies might emerge, banks might use some form of crypto for lots of transactions between themselves, oracles might be more mature and smart contracts might be established in some business domains, etc. The amount of money that might be secured by crypto in the future could be significantly more than it is today.

Blockchain programming? It's already packaged up into tons of accessible libraries. To me that's roughly similar to knowing how to use string functions or OOP: just part of any programmer's set of tools.
What part of blockchain programming? Knowing how to use the web3 API through Infura is just an ordinary programming task; knowing the finer points of smart contract security and all the corner cases in the EVM or Solidity is a lot deeper area.
I don't feel like you can accurately predict the details 5 years out, and they will probably matter.

My suggestion is if it's not too late, focus on something totally different: effective networking with people and integrating yourself into communities. A lot of success seems to be about who you know, what community you are in and who knows about you and your services or products.

Also don't sell yourself short. Having some Linux and networking experience is more than many young people can say.

Generative AI is the obvious current hot topic as far as a marketable skill. I think it will continue to become even more important to know how to leverage these tools. But also for people to find your products or services you need to be part of their network so people skills are also very important.

I am building a website that integrates multiple generative AI APIs with the goal of being able to replace humans doing basic server configuration, making simple programs, editing websites and generating content for customers.

Within say three years we can anticipate many such systems, some built in yet-to-be created AI systems like more advanced video generation.

Robots are getting better. There will be opportunities for people who build services that train and the rent out robot platoons for different tasks.

We don't know what's going to happen with Neuralink but the rapid approach of AGI may accelerate their schedule. If we get high bandwidth BCI in the next 5-10 years that, combined with AI, creates new types of collective socialization, entertainment, problem solving etc. with their own opportunities.

A job say 7 years from now (maybe 17 not sure) might be "VR Role-Playing Game God" which requires a high bandwidth brain-computer interface. Your job could be to watch advanced magic users in the game who are trying to create new spells, and judge whether they meet the requirements to be allowed and setting some technical parameters for their use. After which the AI would determine whether the spells were invoked properly. The requirements for invocation could even involve certain specific visualizations if the BCI were integrated into the visual cortex.

Personally I don't think generative AI will be a marketable skill the same way mobile development was a decade ago or web development two decades ago.

Using an existing generative AI model is relatively trivial, just an API call, and so it doesn't have the complexity needed to create barriers to entry that'll justify high wages. Training a generative AI model does have that complexity, but it also requires compute & data resources that are beyond the means of an indie developer or small startup, and so you pretty much have to work for a well-funded company that can afford oodles of servers and petabytes of data. That limits the pool of potential employers, which again keeps salaries down.

I think it might be much more like the economics for say aerospace engineers or petroleum geologists, both of which are very difficult technical fields, but do not have the bargaining power to demand high wages the way software engineers do. Hell, even like embedded software engineers, which arguably is harder than web/mobile development but paid significantly less because there are fewer businesses competing for those developers.

Robotics potentially has a lot of potential, particularly if standardized hardware platforms emerge that could put the hardware aspects within reach of a lot of startups. It definitely has the domain-specific complexity, and yet will likely be structured as lots of small firms for individual markets searching for a common skillset rather than a few big firms trying to hire a very specialized skllset.

Thanks, I find this motivating for some reason. Although your example feels hyperbolic, nobody can see future well. It is all coming very fast and I don’t want to be left behind because I was lazy. I don’t have many more chances to launch a change.

I have great intelligence, but not the credentials, resources or network. I have to construct some sort of plan and start now or risk being bitter and aimless till the end. I’ve been following GPT since beginning and the amount of progress language models has really astounded me. I feel as if the world is about to change very quickly because of things yet to come, and fear more than death being stuck mindlessly working Walmart type jobs, just to have food.

I have great intelligence

"I can handle things! I'm smart! Not like everybody says... like dumb..." -- The Godfather Part 2 (1974).

Maybe signaling ones dumbness will be the more common thing one day when thinking and working with abstract ideas is just the uncool taxing thing machines do much better.
Forget safety. Live where you fear to live. Destroy your reputation. Be notorious.
I had an analogous experience at your age, but about 40 years ago. I spent that time learning skills that most programmers of that era found difficult, but that I liked. My thinking was that without a degree I'd need something marketable enough that they'd overlook my non-education. (At that time the things I learned were assembly language and compiler writing.) It was a very successful strategy.

Don't take these examples literally, but maybe I might try to be the best Rust programmer I can because it's thought to be a hard language to learn. Since you like networking maybe low-level Posix networking in C, or websockets. This are probably bad examples because I'd do a seriously deep dive into what jobs are being advertised that look interesting to you, which ones pay more than average, and which companies you might prefer to work for.

A related strategy was that I only applied for 2 or 3 jobs at a time, because I looked for companies I thought I'd like, and which I thought I could serve well with my skills. When I got to the interviews I would have a genuine interest in the company and could ask informed questions about areas that intrigued me.

This is great advice and led to success for myself coming from an engineering degree with no programming courses.
Interesting. Thanks. I will look more closely at what these positions actually want.

I thought about rust before, seems like has potential for growth and hype is building. I was discouraged by previous advice that it’s rough introduction language and that it will take longer to make any projects that would convey competence absent credentials or experience.

Heads up: there are notoriously few Rust jobs, and they're mostly in crypto.
I agree. I tried to emphasize that these were just examples and that OP needed to do real research. Although I still claim that if you level up enough you can snag one of those rare available gigs. Back in the day compiler jobs didn’t exactly grow on trees but I did get them with exactly zero formal training.
However, there are many more jobs that can be done with Rust in Linux networking. The company probably won't care much (especially if they don't have legacy code in C) if you want to do it in Rust.
Whatever it is, learn the fundamentals of it. You'll get swept around by quick job market changes if you only learn enough to be productive. The more things you learn from first principles, the more robust your skills will be, and the faster you'll be able to learn whatever you need to learn next.
This doesn’t answer your question precisely, but:

Pretty much every job I’ve ever had came from knowing someone, whether a friend or former co-worker. I think technical people tend to overestimate the degree to which raw “skills” factor into who gets the job. Companies, especially small companies, are far more comfortable hiring a friend with level 6/10 skills than a complete stranger with level 9/10 skills. It’s just how people work.

So, my advice is to attend a bunch of meetups in various technologies (Python, mobile, security, etc.) and see who you get along with the best. In my experience, certain personality types gravitate to different fields. Because the relationships you build will ultimately be the ones that get you employed.

Counterpoint: SF Bay Area, been attending tech meetups for several years, some for years at a time, organized some of my own, go to co-working sessions in cafes, spend a lot of my afternoons at a "hacker space"... And I expect exactly none of this to mean anything for my career. Never has, seems like it never will.
But are you actually building relationships with the people at these events and keeping in touch with them?
Yes? I'm not sure what evidence to offer you.
The way your comment was written, it gave me the impression that you were mostly just hanging out at events.
> Ask HN: What will be the most valuable technology skills in future?

IMHO, the answer has not changed in 100 years: Critical thinking. Knowing what to believe and what not to believe.

If I was young again and had six months to a year to just do whatever I’d definitely get back into hacking on blender and turn it into a real job.

There were a bunch of people who were getting jobs at animation studios based off their work on blender back then and nowadays the Blender Foundation has funds for more than just a couple core devs.

If you’re good at math and can code a bit learning your way around a huge codebase like blender so you can add state of the art stuff from papers will definitely get you noticed.

Spend some time reaching out to companies and making relationships. Take a look at the company page and then ask them what roles are needed like product development, security, ops etc. A few cold calls a day won't hurt and you will develop various business skills really quickly.
Contrarian take: plan for failure? Learn to live on as small of a budget as possible. There is an American man who only eats $1 of food every day. Mention this frugality in job interviews.
Critical thinking and comprehension.

The present itself is changing so much.

Knowing how to adapt and evolve to change is a skill that many people don’t seem to have. Unfortunately they get left behind.

Getting too comfortable doing things one way or assuming things just won’t change is naive thinking.

If someone has solid critical thinking and comprehension skills, they can problem solve their way out of anything. Doesn’t matter if they’re in the IT field or researching history. You need to learn and pick up new skills fast, and be multidisciplinary.

Finally, I think the goal for someone who wants to stay near or at the top is to make as much money ASAP in their early days.

If you can quickly make your way into the investor class, you are basically free from having to keep building new things and hustle like a mofo. At that point you can just invest in some funds or maybe lead them, and honestly just enjoy riding the wave as long as possible instead of swimming against it to catch the next one. Just don’t over extend or get too greedy :p

Not most, not optimal, but leveraging the efficiency prowess of LLMs will definitely keep you alive longer before being replaced by one of those AIs...