Is <a href='https://ainewsera.com/how-to-use-new-google-gemini/artificial-intelligence-news/' title='Discover the Ultimate Guide for Mastering Google Gemini AI in 2024' >AI</a> going to completely replace coding?

Is AI going to completely replace coding?

Over the course of the last 10 years 15 years, almost everybody who sits on a stage like this would tell you it is vital that your children learn computer science, everybody should learn how to program. In fact, it’s almost exactly the opposite. It is our job to create Computing technology such that nobody has to program and that the programming language is human. Everybody in the world is now a programmer. This is the miracle of artificial intelligence. The countries, the people that understand how to solve a domain problem in digital biology or in education of young people or in manufacturing or in farming, those people who understand domain expertise now can utilize technology that is readily available. You now have a computer that will do what you tell it to do. It is vital that we upskill everyone and the upskilling process I believe will be delightful surprising. The CEO of stability AI also has similar thoughts on this, there are no programmers in 5 years. So those of you with kids who you having, you know with python lessons and so forth maybe it’s instead helping them to understand how to ask great questions or give great directions or prompts. I think we always have to look at the unchanging versus the inevitable.

The Evolution of Programming Languages

If we look back at the history of programming languages, we can see a trend of making coding easier and more accessible to a larger audience. From the complex languages of the 1950s like Cobal and Fortran to the more user-friendly languages of today like Python and JavaScript, every new language was designed to simplify the coding process. This trend of democratizing coding has led us to the current era of AI coding assistants, where artificial intelligence is poised to take over the bulk of coding tasks.

Layers of Abstraction and the Future of Coding

With the introduction of AI coding assistants, we are witnessing another layer of abstraction in the coding process. Just as each new programming language made it easier for humans to interact with computers, AI coding assistants are streamlining the coding process even further by allowing users to communicate their coding requirements in natural language. While AI coding assistants are not without their challenges, they are continually improving and becoming more proficient at generating and debugging code.

While AI may eventually handle the majority of coding tasks, there will still be a need for human intervention in areas such as problem-solving, user interface design, and overall software development. Coding will become more of a collaborative effort between humans and AI, with each bringing their unique strengths to the table. The future of coding may involve humans providing the creative direction and problem-solving skills, while AI handles the technical implementation.

Ultimately, coding was never the source of value itself, but rather a means to an end. The core skill of problem-solving will remain essential, but the traditional precision demanded by coding may no longer be a barrier to entry. The future of coding may involve humans working alongside AI coding assistants to create innovative and efficient solutions to complex problems.

In conclusion, while AI may revolutionize the coding process and automate many tasks, there will still be a place for human creativity and problem-solving in the field of software development. As AI continues to evolve and improve, it may become a powerful tool for programmers to enhance their skills and streamline the coding process. The future of coding is likely to be a blend of human ingenuity and artificial intelligence, working together to create innovative and efficient solutions.


27 COMMENTS

  1. When we reach 1 billion-ish tokens context, AI will be able to create their own programming language, based on lowest-level languages (either ASM either machine), and we'll just put ai in competition with other instances to find the best solutions, then train AI with this set of best solutions… and so on.

    BTW you always forget to mention mixtral/mistral, but this AI is crazy, i'm using it daily since i discovered it, it runs locally, and i went from struggling to do short python scripts with chatGPT to directly asking User-interface included codes to mistral vocally…
    I mean, as an example, 2 days ago i needed a program to compare .webp pictures for similarities, i KNEW i'd likely need at least an hour searching, comparing, trying softwares, so instead i just described what i needed to mixtral and half an hour later i had exactly what i needed, and even more. However it's true that some knowledge was needed (which image comparison method to use as an example) to formulate the demand correctly. (histogram and average hash on HSV, comparison).

    Can't wait to see the first AI-OS and get rid of windows/apple bloatwares/spywares, but at the moment the AI is as toxic as most other fields since people don't encourage the open source…

  2. As a young graduate, i am sure that these AI assistants are very good at almost every task.we shouldn't be afraid to use coding tools,someday these code editor will become a complete context window and the same programmers will be using that.

  3. Is it dead? Hell no, it isnt. Those code assistants currently can barely do more simple tasks or better said assist, let alone going much more deeper and complex. Everything else about how soon AGI will replace manual coding etc. is a speculation…usually by uneducated people who never coded in their life.

  4. Well, I was taught Python by not the best professor, and I honestly am afraid to ask actual humans for help with Python because I'm pretty sure they'll criticize my naming conventions rather than just fucking helping me. I don't use underscores and lowercase for anything usually because it's ugly. It's still obvious what's what, I think. I mean, if I gave someone a snippet, I'd just explain what is there. I started doing some functions with capitals and underscores, but some just look wrong like that.

    My professor taught us snakeCase for Python, which is Java standard. He taught us a lot of things that are apparently wrong.

  5. It makes sense. We're going in the direction of higher and higher level programming languages, with LLMs almost reaching the endgame:
    Binary -> Assembly -> Low-level language -> High level language -> Natural language

  6. All of the "build this app for me" examples are pretty much the same boilerplate app. Real apps have more delicate logic requiring more than common patterns. It would be more impressive if they ever built anything new.

  7. One thing about code generation is, it works great if you tell it things like "write a python pytorch image classifier". Thats. great if you know what all those words mean and why you'd want AI to do that. And you also know what to do with the code after AI writes it. So you still basically need to be a programmer. Just to pose a worthwhile prompt, and use the generated code.

  8. well I‘m using AI wherever I can every day, but I can tell you LLM models are language models, they aren‘t able to do big engineering things or have you ever seen a complete application like a app made only with AI. there are just to many problems in the Details. the production ready stuff must be secure, reliable and 100% be tested, must cope with human designed hardware and systems , etc etc.

  9. Deep down you know, at some point people, won't steer or manage AI. There will be nothing to do and AI will be the one doing the managing of humans.
    The humans that stand a chance of control are the transhumanists. A hard truth many are still not willing to accept.

  10. interesting. but what a bug touches the whole system of AI in the world one day in the future? and then the system is down? then what? human programmers will be required to code because the AI system is down temporarily. this incident is possible to happen ? finally the AI system is itself a bunch of codes and can be touched by a virus or or any other type of bugs. so the fact that AI CEO says, we will need no programmers in 5 years from now doesn't take into account the fact i mentioned. so coding will be always a in need, cos the rules of coding game will change from coding to make a program to coding to fix AI bugs and make AI programs for small businesses which can't afford paying for installing an AI system. so don't stop learning coding and encourage your kids to do it.

  11. I failed my Python final ten years ago in college. Without being able to reference Open Stack, I wrote my name at the top of the form, submitted the test, and thanked the professor for the course. Now, as a Ph.D. student I use ChatGPT to help me write code for R to analyze data and produce pretty good visualizations. Sometims I sit back and reflect on how far technology has come to be able to include someone like me in this level of computing. We have coders to thank for that. Luckily for them, we will always need coders to code the AI. Still, I'm excited to see where this democratization of computing will take us! Thanks for the inspirational video! You've put some wind in my dissertation's sail 🙂

  12. Agree we should move with the flow. Focus on our jobs as being creating software rather than being coders so we take advantage of the tools that are available and not put our heads in the sand.

  13. If companies push prices on applications due to higher competition, you will have to code faster. And if AI solutions make an appication for a lower cost and faster, companies or users will probably choose this.

  14. Why have the videos of the Nvidia and StabilityAI CEOs so many cuts? Why do the words don't match with the lips movement? So is it realy what they actually said? – Overall a great video Matt, thanks a lot!

  15. AI is a tool. It will multiply the efficiency and productivity of human coders. One human coder will be able to do the work of many pre-AI human coders. Under capitalism this will cause a glut in the market for human coders which will decrease wages and increase unemployment. Also since labor creates all value, AI-generated coding pay will drop by even more than if there was just a glut of human coders in a non-AI world. Just like a midjourney prompt jockey can’t charge as much for AI artwork as a master painter can for a real painting. This is the downward spiral of capitalism that creates poverty out of technology-enabled abundance.

  16. Absolutely not. I use gpt-4 extensively for writing code and it literally sucks. There is now way any chat model will be able to deal with the requirements and technical complications us we need to deal with as devs is gonna get solved any time soon. And i am saying this as a developer not because AI wont be able to write code properly in the future, but there are ao many moving parts to it that humans will always be required to adapt and react to new and conplicated situations alarrive in various different moving parts and coordinating them togethet.

LEAVE A REPLY

Please enter your comment!
Please enter your name here