Is it still worth it to learn to code? (I think so)
ฝัง
- เผยแพร่เมื่อ 20 ก.ย. 2024
- I try to answer the question in this video of "Will I still be able to get paid with the skills I'm learning while I learn to code?". It's a new style of video for me, so let me know what you think. I hope it was enjoyable for you.
Also, if you like videos with zero cuts, fun presentation and low stim, feel free to leave a like haha and subscribe of course! ;)
Links
Twitch: / teej_dv
Twitter: / teej_dv
Github: github.com/tjd...
Neovim: BTW
#ai #programming
i love how professional this is compared to prime lmao
He explains things like this often lol
It’s up to us in chat to bring it back to the middle ground. Eight equals D, code deeznuts, so on and so forth.
I like that he doesn't talk like he drank a gallon of coffee.
Prime was much more digestible at first. He got down the Twitch rabbit hole of bait and crazy ADHD-inducing content. Can't watch any of him now. So sad.
Skill issue@@Marco9603
I get paid to code but even if all of it was automated away, I'd still code cause it's enjoyable ❤
You know what isn't enjoyable? Losing the ability to feed your family.
lol@@travis8106
What sort of work will you do if you lose your job to AI?
Man, this is refreshing to have a great explanation without dozens of cuts in the video.
Thanks!! It takes a bit longer to script and record, but I'm really happy with the results this way (although it's painful failing at minute 13... Haha)
@@teej_dv It's SO MUCH BETTER. Those cuts are just terrible for my focus. I love how you slightly stumbled at the end but you just....paused and started over. So refreshing. Thank you!
15 minutes of wisdom and empathy. A much needed balanced approach, thank you TJ :)
Glad you enjoyed it!!
I really like the "problem solving" and "solution copying" comparison. At my previous job, the former brought joy and excitement to my work. The latter however, made my work feel hollow and led to decrease of my productivity. So even when I was tasked to "copy a solution", I would always turn it into problem solving. This would at times lead to higher cost of resources, but I just wanted to have fun :D
Same. From this point on, I'm gonna politely "copy" TJ's analogy whenever someone ask me the same question.
All I know is that I'm going to code until the robots pry the keyboard from my cold dead hands.
Mwahahahahaha 😈
As a developer for longer than most .. I love digesting other programming languages and learning things I might not really ever use for a paid job. I enjoy learning what makes all this stuff work in regards to server hosting, repositories, docker containers, neovim, lua, go, php, js.. etc and within all this learning I get as frustrated as a newbie..
Sure my experience does provide me the ability to logically see similarities between these stacks and build quickly or grasp something faster.
Any every level there is a learning curve.
With all the talk of AI, people are forgetting why they got into coding for wanting build something fun. There’s honing to be technical advancement that might even make AI useless. Think about if Quantum Computers can be miniaturized down to a smart watch size, the computing power and core functionality might just be faster than AI. There’s tons of software solutions unimaginable because we haven’t been shown the next challenge.. I doubt AI will just magically evolve and construct solutions for something beyond our grasp of the universe..
AI will be dependent on use to imagine the impossible first..
Great summary TJ, I've enjoyed listening to it and it might keep my thinking busy for the next couple of days. Interesting conclusion...
Thanks 😁
The pacman Joke caught be off guard, love it ❤
First i start getting in my feed this guy with a mustache screaming at the camera (Prime) then, a video of this other guy with a mustache as well talking about frontend development appear (Theo), lastly i dont know how, this guy named tj who i though do pretty interesting things ended up being friend of prime, and now, i watch all these guys creating good content and giving some really good advices, thank you
As someone who's had this thought in the back of their head, this was a very inspiring watch
As always, I really appreciate your input. Love your videos 👍🏼
Bro putting out an absolute banger on April Fools day.
I literally didn't realize this until about an hour after I posted the video and felt big dumb
Thanks, dude. Reading the friendly manual (or listening to it, hehe) really helped me on so many things.
13:58 "I'm just a regular mortal, I put my pants on two legs at a time"
Love it teej
The whitepill about this AI hype is that the people that got into coding solely for money are just gonna move on to the next gold rush.
Programming is once again for genuine tinkerers/problem solvers!
That's a good way to look at this.
Couldn't agree more. I came to coding through science. I chose to focus on it because I realized how grossly inefficient a lot of our work is. I'm just now learning what merged sort is, yet I'm viewed as a programming expert because I learned to fix real world problems. If there's a point where you should stop learning code concepts, I haven't found it yet.
I just started learning web development a few years ago but I have been working as a mechanical engineer for around 15 years. Just lately I started doing things for other people and noticed the huge overlap in the two fields. Problem-solving starts with problem definition (requirement definition). Also communication is hugely important in just about everything.
Really great, well formulated message! Hats off to you!
A really well put together presentation, thanks!
This is a great take. I find that the more I focus on the process vs the end result, the better of a developer I become.
This was inspirational man more than you believe thank you.
Love it as always Teej, great perspective
many great points, clearly explained and neatly presented !
TL;DW for Rust programmers: Sometimes you can't RWIR because the value is owned by the user and thus can't be modified.
Even the jokes have ownership semantics.
I have worked for a huge top company before and let me tell you they used to lay off thousands of people just to be more efficient on a basis that is much much smaller than an AI revolution. When AI will become perfect in solving code, I am 100% certain they will have no issue whatsoever to lay off 99% of the software developers.
the 2nd part of the talk is quite vim-biased. I personally love to code. And if AI will take it from me, I'm not interested in solving problems via indeterministic text prompts. I'll be switching to woodworking...
Imagine an artist that likes to draw. Not to create pictures, but to be drawing: they love to mix colors, to move their hand on top of the canvas, to switch brushes and hand pick their paint in a shop. I think those people are screwed now.
yes, that is very true. I am a financial auditor with programming skills. I build a cli for myself to work efficiently. :)
I do think the introduction of AI into the fabric of software development and operations is going to be faster than many seem to believe - sure there is a lot of hype, grift and wildly cost-ineffective solutions right now - but the potential is already there for AI to be at the very least a highly useful complementary tool. Still, I find your stance very sensible. Because even a rapid introduction of AI is not going to obsolete software development / coding skills for a long time yet. And if it is slower and follows a more complementary path - that just makes learning to code even more important, as AI will increase the value of what one can do then - but in that scenario will be purely an amplifier and not relevant as a replacement.
I am very bullish on AI becoming a key part of software development - as complementary/productivity tools at first - then as ever more autonomous and agentic system. However - I still think it is worth learning to code. To be a good coder means being good at understanding a lot of domain knowledge - from low-level and highly technical topics like protocols, file systems, OS/system level stuff to algorithms to business domain knowledge. And this includes the skill of translating fluffy requirements into code. A software developer transforms business requirements from poorly articulated human language into better human language into technical design into code. A software developer is a troubleshooter and a crisis manager. It doesn't matter if AI becomes an increasingly big part of the work. Even if/when programming becomes entirely in natural language - the skills of a software developer are still relevant - the coding is rarely the difficult part - figuring out the requirements is. Navigating the wider technical ecosystem is. Operations is. And so on and so forth.
So from a super bull on AI: go learn code. It will make you extremely valuable labor even in a post-code world. If we get to the point where AI can figure out all of the many aspects mentioned above? We'll be in post-labor utopia anyway - because it will be able to replace every other job function as well (including 'business').
But if you don't like adapting to new technologies and new methodologies? Pick your specialization carefully. But that was always the case in the world of coding. It was always risky and kind of dumb to stubbornly learn a very specific skillset and refuse to adapt. Ask the people who got left behind when desktop application development turned into web development. Ask the mainframe developers who didn't stick it out until they became hyper-valuable keepers of ancient COBOL lore.
I enjoy both coding and problem solving, but for me I am mostly interested in learning about how computers work from a fundamental level. That’s why I decided to stick to C, even though it’s clear that Rust is much more powerful. Eventually I will learn Zig but for now I’m sticking with C.
Maybe it’s better to know what you are copying and pasting, in case you get outsourced and have to use your own brain. By potentially not getting stuck over that problem . This quote is my lifesaver:”learning is living”not learning something because you fear it will be useless in the future won’t do you any good. Learning it won’t do you any harm either. You might fear that your brain’s storage will be wasted, but when are you going to use this brain with billions of neurons capable of storing billions of pieces of data?
Amazing video TJ!
I like how at the end of the day, this isn’t even about AI at all. It’s just good advice
AI generated programming code reminds me of the dude in Catch Me If You Can - it has the plausibility of looking like the real thing, but at the end of the day you really wouldn't want the dude really flying your plane or really doing your brain surgery regardless of how convincing his gab is. Have been seeing that around 50% of AI code doesn't even work and often contains out right bogus details. Because AI is witless and doesn't actually comprehend. Fast food industry is seeing about 70% failure rate of AI order processing - which means 70% of orders still have to be done by a human. Well, the AI systems and vendors aren't cheap and will require much investment to overcome a 70% failure deficiency.
I was that person that picked up neovim because it looked cool. Whenever I pick a new editor I get so overwhelmed though, so that's was the reason to google my way through things and just copy and paste (basing off kickstart, btw :3).
Now I've more or less re-done my config, set up my own version of tmux sessionizer, sway, and honestly it's a blast. Even if I don't get to use nvim for my job (yet?), I still don't regret learning it, my cli tools, bash scripting, etc. Always nice to go faster
Where was I going with this? No clue, honestly, but thank you, Teej
Solution copying is a gateway drug to problem solving.
tjdevries uploads his VODs to the YT since 5 sep 2013. Luke Smith since 12 nov 2012.
👋byebye
"Everything besides C will still be legal" now I have a strong urge to write C
relatable lmao
Solutions copier vs Problem solver is 100% correct, I do think its more of a scale though but thats like everything in tech
somebody give this guy an award for spitting facts
I wonder if 5 years is enough to get AI code generation to the point that it needs no human oversight to actually replace job positions. Most likely the workflow just change and the people who are adept at using those tools along side being a human for the reasons TJ mentioned. For example people who operate industrial machinery (not sure what that work title is in English) have been using AI for who knows how long, and that field is still lacking workers to fill the demand (at least where I live that is the case). Creative fields won't also go away, but the production toolset just changes. What I've talked with people who have a degree in different fields of media creation, they use AI to generate references and build something from those. You still need someone to know theory to valide and fix the output. even if you were to go straight from generated image to finished product.
It's hard to automate manual labour that needs critical thinking.
Thank you TJ
"I am a regular mortal. I put my pants on TWO legs at a time!" LOL
Ok stupid question, Background :I'm a senior mechanical engineer and do coding on the side as a hobby hoping it will eventually lead to a passive income SaaS.
I've been a bit confused about the take on this video. I totally agree with you, but isn't it just freaking obvious what you are stating? Or am I special for seeing that because I'm a senior in my own field and maybe have a bit more skill in abstract thinking as a result of it which I can apply in your field?
Totally agreed!
Thanks TJ!
If it was all automated away I would look for other types of problems. You guys might find that there are lots of interesting problems to solve, lots of interesting problems in math science engineering, robotics Internet of things
Machine translation engines like Google translate has been 'good' for several years now, but the number of human interpreters/translaters have been continuously increasing. I would say the same thing might happen to code even smaller scale than translation. Translation by deifiniton has a well defined problem, its just how to express it has some subjectivity to it. In coding, defining and expressing the problem itself is more than half the game.
has it been increasing? my partner worked as a translator and saw constant downward pressure on contracts and clients increasingly using machine assisted solutions and this wasn't even for a language well covered for machine translation in the first place. I cant imagine the situation improves for more accurate translated languages like english to spanish french german or portuguese
great take and good comparison
The issue with predicting AI right now is that no one really knows. I'm a grad student at my University and I know one of my professors thinks that LLMs are pretty much solved and they will not bet getting too much better and another thinks that it's only a matter of time until humans are obsolete. So it seems everything is pretty much up in the air.
Right, but if humans become obsolete it's a little bit pointless to worry about what you're learning 😂 that's why I said for as long as we have jobs, in the video.
The day AGI can replace an experienced developer, it can also replace any other professional who uses computers.
I can't wait to be replaced.
In the meantime, just try to not be replaced by silly Devin instance.
When/If that happens, no intelectual or artistical profession will be safe. Not even CTOs and CEOs 😅 And after that, when robots become cheaper, no manual labor will be safe as well.
I wonder who will consume/pay for all amazing stuff AI will produce if there will be no jobs :-)
If junior software enginner job is largly automated then how will junior se will get the experience to become senior
Kagi's LLM-based summarizer gave a great synopsis of this...but it lost all of TJ's wry asides. I was really hoping for it to mention how all programming languages besides C, of course, will be legal in the future.
I really want to know the opinions of the people who disliked this video, if there are any.
Thanks TJ 👏👏
AI seems really overhyped at the moment. Probably because after many years we have some visible breakthrough.
I still think it will be decades before we have an AI that can actually be a functional member of a software team, making some of the hard design decisions that TJ talks about.
Until then, they are useful coding assistants.
Great take TJ
Is TJ the best TH-camr? Yes, TJ is the best TH-camr.
just take my like tj
will do :)
I hate having to go on Reddit etc and ask for help
I’m too lazy
Would rather sit there and try to figure it out instead of spending time writing a post
Good points touched by JJ here.
C being made illegal is blasphemy. Seriously though, C is a language that's close to my heart. C is both a beautiful language as well as a beautiful tool. I think of C like a rose bush. Beautiful to look at, easy to get punctured if you don't know how to handle it. Definitely my favorite language out of all the languages I've learned so far.
I find AI coding tools to be mostly like a manual on steroids. You still have to know what you're doing to understand what they are telling you.
Great take and 100% agreed
There's a reason they are trying to sell AIs potential to your boss instead of just eating his lunch.
While overhyped, it will cover the 'intern' level stuff: get in now before the low hanging fruit is gone and the ladder is pulled up.
I'd like to add, I consider myself as a problem solver, but usually I find myself getting stuck due to not finding a certain API endpoint or a library function name, that's when the "solution copyer" skills come, asking chatGPT for the endpoint name if you can't find it, going to stack overflow to search for the 1000th time how to center a div...
I got stuck recently on writing a simple lexer in rust, because I knew I wanted a peekable iterator, just didn't know it had one already. Looking it up properly instead of trying to implement it myself at 1 am would probably be a better idea
Amazing!
This is gold.
that's very well put
Yea this is it. LLM's are just going to be tools that are used by problem solvers and solution copiers in order to solve problems.
LLM's currently are a fantastic tool that everyone should be using. Sure it can write code for you, basic stuff, and you can copy paste it and be on about your day. But, you can also use it as a teacher. You can say, "break that down for me", than you can explain back to it how you think you are understanding it, and it will either tell you you got it, or that your understanding is somewhat flawed. Its like having a personal teacher at all times. This teacher can either give you the answer (up until what it knows, which is quite alot), or it can help you understand stuff so that you can produce solutions yourself. You than can make it your own personal copy paster solution problem solver. Something that works for you. Because you don't need to write every single bit of code. You don't need to be all knowing. You just need to know the people or in this case, the AI, that do know the things that are relavant, and you get to make the decisions on what to do with that. Thats what employers will be paying for.
Great take
Using AI reminds me a lot of Code wizards back in the 2000s when I started learning code. Everything was nice and easy until that point 4 days before delivery when we had to debug that one little flaw that was hidden in a hundred thousand code lines. And 96 hours to go ...
100% agree. but how many of those millions of millions of coders will be needed to all those side tasks while their core job is gone, the job of coding. will companies still keep them and waste huge money on them just to go and negotiate with a stake holder once a week lets say. I can easily see how 99% being laid off.
I gotta be honest here TJ... I didn't get this April 1st joke :/
Awesome video!!!!
Just came back seeing your tweet. I really found this video good. Maybe you could have uploaded it on a different day. But its a shame this video is not doing that good like others on your channel
thanks a lot i really liked the video
Thanks for your comment!!
👏👏
You know the meme of
"your opinion"
- hold this one second
- ok
:/
:)
"My opinion"
That's me right now
Love that meme
I think the real question should be is how is this going to change learning approaches. If the aim is to remove junior devs then what is the new entry level requirements?
What if the internet just becomes a series of float numbers that humans can’t decrypt? Things like Q* might lead to levels of abstraction to the point of human language being the optimal method of communication with computation.
"Pacman user wouldn't understand" 💀
Stardew Valley mentioned
A software engineer is not paid to write software for the same reason an electrical engineer is not paid to wire cables. Maybe part of the work but what they are really paid for is to solve a problem withing their knowledge domain.
I noticed that big companies do a massive laid off. They actively replace developers with AI and prompt engineers who just command AI - do this or that software.
Any sources to back it up?
@@jenot7164 I wish.
it's scary AF that the thought a programming will become illegal to use sometime in the future.... UGH
Let him cook 0:33
LLMs might end up creating more jobs. Optimistically, LLMs have a great ability of expanding a person's ability in more ways than just code generation. I imagine we'll start seeing more video games, more apps, and more businesses. Many of those will be pushed to the fringe of human capacity augmented w/AI, effectively creating more work and greater need for people to be able to maintain the systems on their new edge. And creative endeavors aren't a zero-sum game. People don't only shop at one business or buy one video game on a steam sale. More people will be empowered to learn. More people will have ideas. More people will be empowered by LLMs themselves to explore different problem domains. LLMs aren't wish-machines yet.
I just got a frontend internship and I can tell you that AI doesnt really solve problems for us at my workplace. We are still a ways off
Tech Debt Devin is not taking your job.
if the advice is good enough for your 3 year old son, it's good enough for me. I'm gonna use my roughness and toughness to try again.🤓😎 No need to translate that.
Haha this made me laugh 😁 thanks!
Hot tjake
Peter
My prediction? In five years, an AI coder somewhere will open git blame to find out who wrote that terrible line of code and it will be a previous version of itself.
this 🧑🍳😚 - made me wanna support the channel!
below's an ai/copy-pasta'd chapter suggestions - HTH!
00:00 Intro
01:26 Beyond Coding
03:07 Problem Solving
05:28 Programming Insights
07:37 Curiosity Pays
09:17 Neovim Parable
12:58 Concluding Thoughts
“pacman users wouldn’t understand” KEKW
tell me more about this… …list…en…ing
9,000 applicants per job-No it’s not worth it to code.
is it only me who's waiting to see april fool video
this guy has no pants on and is wearing ocaml undies!!!
i'm a solution copier and wanna be a problem solver...
Man, what a time to be alive. I don't have to learn Data structures and algorithms, and Architecture Design. I just have to want things.