DEV Community

Dima Sukharev
Dima Sukharev

Posted on • Edited on

AI is replacing Coders, not Engineers. Learn CS now!

first lets grow my twitter, i will probably change all of the world obsolete education systems, not sure yet, the more people follow me — the easier it is to change the systems


hi

how do you know if you are a coder or an engineer?

not all coders are engineers, but all engineers are coders

if you came in tech from a bootcamp and/or youtube tutorials, landed a job 1~2 years ago and you are to become a Senior when you learn 1~2 more tools — im sorry, i was there, you are going to hit the iceberg now

Image description

But I can already build an app, why bother?

AI is going to build such apps in 2~3 years..

putting together DB with a Cache and setting up microservices handled by a queue — is not engineering, its coding per the best practices

Coders copy stackoverflow by error message, Engineers know what they want to copy from stackoverflow

coding is repetitive task of typing characters as per the best practices set by Engineers

Coders repeat, so do GPTs. GPT does not innovate, it generates what has already been said on the internet

Engineers come up with previously unavailable solutions to problems

Engineers innovate, so they are valuable. Coders copy, so they can be replaced.

you want to continue blind copy-pasting until the error is fixed? GPT-6 is coming for you.

t-1000-running

ok, what do i suggest?

start learning Computer Science fundamentals

it improves engineering thinking and problem solving skills. that's how you make difference and value

  1. learn CS — level up engineering and critical thinking
  2. land Google as a SWE
  3. make difference
  4. deliver value
  5. get more $
  6. go up and play lead roles
  7. CTO something your skills are relevant to
  8. then 2 steps to CEO any stuff you have fun solving
  9. or stay FAANG if stock options are good enough
  10. or be replaced and go make music or sad comedy

the sooner you see the wall — the better. so here am i pointing in front of you — learn Computer Science to solve problems and level up engineering and critical thinking

AI and LLMs like GPT are coming for the repetitive tasks that coders do. the best way to stay in the game is to become the innovator rather than the implementer

1st person immersive experience storytelling

practicing in my capturing attention skills, does it work?

the story goes like this:

coder named Yoo Yoself learns how to make an app in a fantastic bootcamp experience, lands a job and gets respects from parents and friends, good start.

some engineer called Samuel L. Altman comes up with a way of zipping files (some inside-out compression algorithm) so coders can do ~10% improvement on their app TTFR (time-to-first-response).

Samuel then asks GPT-6 to create a website from a pen drawing and so Altman “alts all man" with GPTs. haha sorry i stole this one from Fireship, maaan how did you come up with that..

Yoo Yoself is going home finally noticing ceiling hit, but there is still a great story left to tell friends — “hey dudes, ive learned programming in 6 months and got the money!”, however they already heard it..

Coder repeats — they are great fit for the AI automation. Engineers innovate — they are making difference.

WHY i tell you this

i want people to have more opportunities, the more knowledge we have the more opportunities we see, so im sharing my experience with you

now go and learn Computer Science fundamentals, it does:

  • organize experience you got from practice by building transparent mental models and filling missing picture pieces
  • clean your head from memorizing implementations by moving up to ideas level
  • develop critical and engineering thinking upon solid understanding of how software interacts with hardware and why your awesome code works
  • you dont get replaced

this is so you move out of a sandbox and dont go sansara circles in the tech and life probably..

CS is not learning one more framework or programming languages, it's learning ALL of them at once by understanding computer science ideas all tools are build upon

Selling part

im crafting “Build a Computer Simulator” — 100% practice-first, fun learning experience that you get addiction to — even Israeli doctors cant help

it is fundamental fundamentals of applied computer science.

there is literally no better way of learning CS then assembling ALU, RAM and CPU from logic gates.

you will get a web simulator to assemble a general-purpose machine. then you are to program the machine with language you write compiler for within your self-coded operating system (its gonna be online lego Harry Potter vs Jordan Peterson 3000-deluxe stuff)

it will take you about 50~70 hours to finish (not 4 years of CS major though)

subscribe to my twitter to be the first to know when the Build a Computer Simulator is public!

while im creating the Simulator — go to edx.org and check out MIT and Berkley applied computer science materials

bye


Who the F am I?

im a life-long learner who is obsessed with educating himself and helping other people learn things, so they dont walk circles and align their movements ;)

i got in tech in 2017: engineered programs for Alibaba and now playing startups in Deel — YC alumni, fastest growing startup in history

i created and open-sourced OpenCommit and came up with an accurate way of writing code with GPT-4 via AI-TDD which is going to take the world of AI repetitive coding when i have time telling people about it

check my "Learn ML in 52 seconds"

Top comments (26)

Collapse
 
darkwiiplayer profile image
𒎏Wii 🏳️‍⚧️

I'm still not quite over the fact that this is even a problem. Humanity is well on its way to automate one of the last things that had so far been impossible to automate, and instead of being happy that there'll be even less work left for actual people, we're obsessing over preventing it.

Collapse
 
emi_black_ace profile image
Jacob Van Wagoner

The hard part about programming is never "coding." The hard part is always "coming up with a solution to the problem." LLMs aren't going to replace the latter, only the former. I'm happy about that.

If we're going to whine about stuff getting automated then we might as well (in increasing order of absurdity):

  • ditch checkout scanners and bring back the middle class skilled profession of "store clerk" who tallies up the price of the goods.

  • get rid of the machines that fill pastries with creme and manually inject creme into all our Twinkies.

  • go back to assembly line manufacturing by hand for vehicles

  • eliminate tractors and go back to planting fields by hand

Because the important thing is keeping people in jobs, right? At every single one of those things and more, people protested about job losses.

There will always be more things for people to do. Getting them out of the boring stuff that they don't like doing, while still getting that stuff done, frees them up to do something else. The disruption is always temporary.

Collapse
 
riceball1 profile image
Dana Ng

True, and this will always open up doors for more things humans can do. I feel the fear in a lot of the articles about AI replacing this or that job, but people adapt and new jobs do come out of it. Also things evolve and adapt around the new changes.

Thread Thread
 
emi_black_ace profile image
Jacob Van Wagoner

"BUt tHiS TiMe it'S DIFfErEnt!" - every news outlet ever, which get more views and more revenue when people think there's impending doom, which also said the same thing about every other disruptive automation tech.

Collapse
 
disukharev profile image
Dima Sukharev

same opinion, cant wait when i tell some AI-IDE what i want to build to solve a problem and i dont manage implemetations details

Collapse
 
daniel_brenzel profile image
Daniel Brenzel

There shouldn't be a discussion to revert the progression. (Environmental concerns by side)
The problem in the automation of simple tasks is the increasing complexity of the remaining ones. At some point cognitive load may be getting to big for some people to handle.
There should be a social debate about the implications of this problem.

Collapse
 
nakawawa profile image
x2

"LLMs aren't going to replace the latter" They will, just not yet

Thread Thread
 
emi_black_ace profile image
Jacob Van Wagoner

Everybody said that about "no code" but ironically all it did was increase the demand for developers.

Collapse
 
spookyaction profile image
Louis-Alexandre Simard

I rather keep my job.
By the way like 60% of everything you consume isn't automated. All your clothes are still stitched by hand as there is no robot on earth capable of stitching by itself.

Your electronics are assembled by hand.

The future you describe is one where the only jobs available are low to no skill and liw pay.

We will move boxes and make lattes and own nothing

Sounds amazing

Collapse
 
darkwiiplayer profile image
𒎏Wii 🏳️‍⚧️

Given enough time, AI will probably enable robots to do more and more of these tasks as well. At some point humanity will just have to accept that the model of full-time employment for the entire population just isn't sustainable, and we'll have to start leaning back more. Well, that or we continue this weird "capitalism" nonsense, whereby "not enough work" is somehow a bad thing.

Collapse
 
disukharev profile image
Dima Sukharev

honestly i dont like coding, i love building. im happy AI will automate manual labour typing and could not manage zoo of libraries with 10000 different implementations of the same thing

Collapse
 
teddylumidi profile image
Teddy Lumidi

Most people don't like the idea that most of humanity will be super empowered by AI. And some competitors want to have advantage over their competitors thats why the thought of AI makes some people worried

Collapse
 
tensorprogramming profile image
Tensor-Programming • Edited

I firmly believe that, indeed, some programming teams may be smaller due to a single programmer's increased productivity with the assistance of a large language model (LLM) as a valuable tool. However, any company that replaces an entire department or team with an LLM fails to grasp the true nature of how LLMs function. A noteworthy yet often overlooked example is an amateur Go player who consistently defeated the world's most advanced Go AI by utilizing a basic concept of the game.

Essentially, the researcher managed to triumph over the 5000+ MMR Go AI approximately 96% of the time by employing what they termed the "double capture technique." They created multiple loops on the board to capture stones, but the AI neglected the inner loop. This is significant because it demonstrates that a Go program capable of defeating top Go players worldwide does not genuinely comprehend the game of Go. Even a novice Go player would know to evade such captures as they award numerous points to the opponent, yet this AI appeared indifferent.

Most importantly, these Go AIs employ the same methods used in all LLMs. Although these AIs may be able to pass the bar exam or compose a top-tier essay, they lack a true understanding of their actions. In the context of coding, this implies that these systems are approximating code writing rather than genuinely writing code like a programmer. They can mimic and identify patterns in existing code, but their limited understanding can also lead to hallucinations.

This is also the reason why these tools will always require the presence of an individual with domain expertise to fully leverage their capabilities. Without comprehension, these LLMs cannot operate independently. Even something like AutoGPT needs a chaperone to work properly.

For context, I am an engineer who has been working with GPT-3, GPT-4, and GitHub Copilot X for several months now. While these tools have saved me time in various instances, the experience is not akin to pair programming with a junior developer; it feels more like guiding a search engine to locate the information I need. Although it can generate code and retrieve information more rapidly than traditional engines, the resulting code is just as prone to flaws as any content found on GitHub or Stack Overflow.

This IMO is the biggest flaw with modern AI systems and until someone figures out a different way to build an AI or a way to cover these weaknesses, the current generation of AI won't be able to fully automate these more complex jobs without the help of a human.

Collapse
 
disukharev profile image
Dima Sukharev

absolutely agree here, great text, thanks for sharing

Collapse
 
josetorresks profile image
josetorresKS
Collapse
 
tensorprogramming profile image
Tensor-Programming

Exactly. Hadn't watched his video but I had read the research paper when it came out and it raised an eyebrow...

Collapse
 
josetorresks profile image
josetorresKS

Literally this video

Collapse
 
timillionaire1 profile image
Timileyin Akinyemi

I love this documentation. 👍

Collapse
 
liveuk profile image
Mike Hebblethwaite

Edit suggestion on your twitter profile.

Current "Helping World with software while having fun"

Suggestion ""Helping the world with software while having fun."

Or this as we expand

Collapse
 
disukharev profile image
Dima Sukharev

thanks for that <3

Collapse
 
sabbir2609 profile image
sabbir2609

well said

Collapse
 
njengah profile image
Joe Njenga

This is spot on. I saw this two years back 😊

Collapse
 
christian_go3 profile image
Christian GO

Fear mongering to sell a scam product.

Collapse
 
disukharev profile image
Dima Sukharev

ohh the internet, i love it

i've found a Stanford CS 101 for you for FREE ❤️

Collapse
 
nakawawa profile image
x2

nonsense, coders will use GPT and engineer's tools to create products for user demand, this is about money, and survival, stop gatekeeping what we should learn son.

Collapse
 
disukharev profile image
Dima Sukharev

dad, i asked you not to comment on my posts, people are watching! i told you and mom that coding is manual labour and it will be less and less demand for clicking keyboard when Copilot gets more accurate and engineers solve problems 666 times faster with less need to hire more coders <3