What is the Future of Software Development?

What is the future of software development? Will it be VR? AR? Web only? Will everything be delivered over the web? Is desktop development dead? What should I be learning now? These are the questions we will be addressing in this episode of Dev Questions.
Website: www.iamtimcorey.com/
Ask Your Question: suggestions.iamtimcorey.com/
Sign Up to Get More Great Developer Content in Your Inbox: signup.iamtimcorey.com/

[@gurmin1]

I've been a developer for over 30 years. I advise new developers, _"The only constant in this business is change. Unless, you're standing before a vending machine."_

[@IAmTimCorey]

Thanks for sharing!

[@AlexxXRecorD]

Cool)

[@arashsafi525]

Thanks Tim, you explain every things very well.

[@IAmTimCorey]

You are welcome.

[@jdr8364]

This was well explained and got tons of information that gave me an idea on how to approach my career. Thanks a lot Tim, your a LEGEND!

[@IAmTimCorey]

I am glad it was helpful.

[@andergarcia1115]

Wonderful question and wonderful answer. Thank you Master

[@IAmTimCorey]

You are welcome.

[@faisalalhoqani6151]

Great episode, dear Tim, Thank you.

[@IAmTimCorey]

You are welcome.

[@carljung2850]

This guy keep posting good stuff for free ,amazing 👏

[@IAmTimCorey]

I am glad it was helpful.

As always, good one! 👌

[@IAmTimCorey]

Thanks!

hi Tim Thank you very much for always sharing valuable material for everyone. The approach you present is excellent, certainly the future is uncertain and depends on many factors that are impossible to control. I think that by taking your advice we can adapt to the comings and goings of technology. Today AI is in fashion, tomorrow who knows, we can only try to be part of the best developers who, due to their experience, look forward to the coming changes, but without ceasing to be cautious to face them according to each case, punctually.

[@IAmTimCorey]

You are welcome.

[@DeltaXML_Ltd]

Really interesting video, keep it up!

[@IAmTimCorey]

Thanks!

[@mikey803]

I started with Basica and Cobol. Both done on a 5-1/4" floppy disk. On a side note, when I was guessing you were going to say, "it depends", I guessed wrong. lol

[@IAmTimCorey]

Next week. 😉

[@techwithdipufrom0ton621]

❤it.n Thanks for sharing

[@IAmTimCorey]

You are welcome.

[@michaelnurse9089]

I see two trends I don't like. 1) That senior developers will be able to do a lot of stuff previously done by junior developers using AI, especially as it continues to progress. 2) That AI will infiltrate low/no code and be able to take it far, far further before a developer is needed. Both these trends will help those already well established but make it difficult for those trying to break in. Of course, the future is a strange place and things normally turn out differently to what you expect.

[@IAmTimCorey]

Why couldn't junior developers utilize AI to be better than typical junior developers, thus making themselves more marketable earlier?

[@michaelnurse9089]

Good answer. I guess you have to first answer the question is the demand for programmers effectively unlimited or limited? I could make a case either way. If it is the former then AI will only be a good thing and usher in a doubling of productivity. If it is the latter, then AI could simply 'steal' all junior positions. @@IAmTimCorey

[@IAmTimCorey]

My opinion is that development is effectively unlimited. Here is my reasoning: First, have you ever worked at a company on a software project that was fully completed? Not launched or declared production ready. Fully completed. No bugs, no desired features, no edge cases, no missing platforms. It doesn't happen (unless the project is so small as to be insignificant). Even Notepad in Windows has been getting updates. Second, are the number of things using computers/processors/etc. growing or shrinking? Toasters have a CPU now. Refrigerators connect to the WiFi. In my backyard, I have a personal weather station that evaluates my local conditions, networks with other stations in the area, and generates a customized weather report for my house. We keep adding new devices and new features to existing devices that take software development. Third, as software development gets easier, more companies that have had to rely on off-the-shelf solutions can start building their own custom solutions. For instance, because I've been able to do that already, I've built solutions that have helped me grow my business well beyond what it would have grown if I had not been able to build my own solution.

[@rikudouensof]

Unless it is windows forms got me. Solid advice. I jumped into UWP, now I got only Xaml from that. Thanks for the video

[@IAmTimCorey]

You are welcome.

[@ai_coding]

Great tutorials keep up the great work! Not sure if anyone has asked this question but is it possible to run c# gui on a pi 4 & on boot? Thanks

[@IAmTimCorey]

Thanks! And a GUI? Not with a first-party Microsoft project. I think Uno and Avalonia might be able to do it, though. A Pi runs Linux, so that's all you would need to do is create a Linux desktop app.

[@codefoxtrot]

20 years ago, many of my college professors touted all of development would standardize around Java. It was to be in every device, every appliance, every vehicle-- it would be all that there ever was. They also proclaimed Software Development would be a dead-end in the United States, that this work would be outsourced to other countries, and that we should instead learn to manage software requirements, design, and acceptance testing. I've, thus far, really enjoyed making a career out of how wrong they all were. Change is the only thing you can bet on. Don Draper: "Good. And let's also say that change is neither good nor bad, it simply is. It can be greeted with terror or joy, a tantrum that says, I want it the way it was, or a dance that says, look, something new!"

[@codefoxtrot]

Also: Visual Basic Foreverrrrrr!!!!

[@IAmTimCorey]

Yep. There have been a lot of bold declarations that have not come true. The most important way to handle an uncertain future is to be wise about what you do in the present.

[@g2bam]

There’s a need for about 3 computers in the world and 640K is more than enough memory any computer should ever need, so once said a certain Bill Gates. Said to me over 40 years ago… “I would give up this programming if I were you because in 5 or 10 years the receptionist will be doing it in between phone calls”… said to me last week, 40 years after the first statement… “I would retire or give up this programming if I were you because next year the computer will write its own code in between running the whole business on its own” Object lesson, don’t listen to anyone who themselves are not professional programmers, even if they’re highly technical and doing some serious stuff in another IT discipline. Fact, you cannot design nor produce software for human needs/use without involving a human being. Now, it’s my turn to make a prediction! In 40 years time there will still be human programmers, top level highly skilled ones, not company receptionists and while there may be some element of self-writing by AI, but it will not be entirely without the human touch and ‘real’ intelligence. It may very well be the case that all or a majority of the programming work would be to write systems that produce systems! Enjoy your life as a programmer for the next 4 decades or more!

[@raynoldj]

Hi Tim. I'm a 41 year old guy without any coding skill except html. Do I need to learn css and javascript as a pre requisite before taking your C# Masterclass course?

[@IAmTimCorey]

No, you do not. You can start the C# Mastercourse without knowing anything as a prerequisite. Knowing HTML will be helpful when I introduce the C# web project types, but that won't be an emphasis.

[@raynoldj]

@@IAmTimCorey can I build website using C# or is it only for backend?

[@gahshunker]

Tim, how do you feel about the low code/no code (drag and drop editors) hype train? do you think there is place for it to be a thing, or is it just a gimmick? Great video, thanks!

[@sivaprasad905]

It's obvious they will get better as time pass by and companies will always choose what's more beneficial to them

[@IAmTimCorey]

I think it definitely has its place. The funny thing is, low-code/no code has been around for a long time. We used to create web pages in FrontPage using drag and drop back in the 90's. There will always be a place for helping people create something that they need without involving a developer. With modern tools like the Microsoft Power Platform, more powerful work can be done without building a full system. However, what you do need is a person who can think like a developer because that's what they are doing. They need to make good decisions about the logic of the design they come up with in order to make something that is useful and accurate. These solutions will never remove the need for custom software entirely, but they do fill an important gap. There is more development needs in the world than developers can fill. This allows for more of those gaps to be filled by non-coding developers.

[@IvanRandomDude]

It's not new. Dreamweaver was a thing in 90s.

[@calmhorizons]

@@Luke5100 Also, the technology is, by definition, always going to outpace the tools, because the tools are built with the technology. It's a bit like worrying that the existence of car factories will put mechanics out of work.

[@konstantinrebrov675]

@@IAmTimCorey Ok, what do you think about AI autocoders, such as ChatGPT and Devin AI?

[@robertmazurowski5974]

What if we live in final times where technologies will not change as fast anymore.

[@IAmTimCorey]

History tells us that this isn't the case. People have frequently thought they were at the height of civilization and technology. The Roman empire thought they had it all figured out. They had great roads and aqueducts that moved water to places where it needed to be. People fought against the invention of the car because they thought the horse and buggy was the peak. When the electric light was invented and then lights were introduced into houses, people thought technology was moving incredibly fast. Yet my great grandmother was alive to watch the Wright brothers' second flight (in person), to ride on a jet airplane, and to watch men walk on the moon. To her, technology was moving incredibly fast. Yet the technology of the 1960's seems primitive compared to what we have now. When I was in college, I had an engineering professor explain how the logic gates on a processor were so small that we would never be able to achieve more than about 400 MHz from a processor. We have chips that go over 10 times that speed installed in regular desktops now. Technology innovation is speeding up, not slowing down. The reason it doesn't plateau is because we are building on the shoulders of the last generation. I'm not sure where it goes from here, but machine intelligence is on the rise. So is VR/AR. Robotics are also improving rapidly. The bottom line is that we haven't come close to our potential. We haven't reached the limit of human ingenuity. We have a LONG way to go.

[@robertmazurowski5974]

@@IAmTimCorey this is a great comment. But regarding programming paradigns there isn't really much innovation. I feel like every language feature is done. I feel like frameworks are not new, they just make certain aspects easier and new frameworks have an advantage of not being bloated. And I feel there is a cycle of new framework, very fast to work with vs an old bloated, and the cycle repeats. C++ to JAVA to JAVA EE. JAVA Spring when it was new to Spring in 2015 to spring boot. Python Django to Fast Api.

[@IAmTimCorey]

Even in the .NET ecosystem, there have been LOTS of innovations recently. For instance, Blazor Server came out with .NET Core 3.1. It is the only framework I know that is a server-side project that acts like a client-side application. That was 3 years ago. Then, they put out Blazor WebAssembly and made C# work fully client-side and be offline-capable as a PWA. Then 18 months ago, they came out with Blazor Hybrid, which allows you to write C# code that can run on the Windows and Mac desktops (natively) and on iOS and Android (natively). Now in a couple of months, they are releasing a unified Blazor that allows you to start out as Blazor Server, then transition to Blazor WebAssembly. It will also allow you to only enable the client-side features on specific pages or parts of pages. That is a HUGE leap forward. But I've seen the internal demos - they go WAY further. And that is just Blazor. Just in the past few years, we have added gRPC, which is a faster web communication protocol (compared to API), that also allows for stream communication (communication that can continue over time rather than point-in-time calls). .NET Core 3.1 was orders of magnitude faster than .NET Framework 4.8 with the same C# code, and there have been significant performance improvements with every .NET release since (.NET 8 has HUGE performance improvements). AI is now being implemented inside the .NET ecosystem as well. As new platforms come out, .NET goes to them as well. The language has expanded and changed as it needed to in order to address the current needs of the developers. If you look at C# 10 compared to C# 5, you will see a TON of changes. When we get to C# 15, we will see a TON of changes compared to C# 10. I think the problem is that we can be too close to the situation to see the incredible changes to the language and how it is used. Just like how we can hang out with our kids and it seems like they are always the same size, but if we compare them to pictures from six months ago they have obviously changed a lot.

[@terjes64]

I develop a desktop app, its what my industry wants. Yet parent company say our software is legacy, because we are not supported on all platforms. 🤔

[@IAmTimCorey]

Unfortunately, companies aren't required to be logical.

[@TheVincent0268]

blazor hybrid is promising

[@IAmTimCorey]

It really is.

[@annagulaev]

The future of development is to throw more and more people with more and more complex tools at problems that programmers used to be able to solve and small business used to be able to afford.

[@IAmTimCorey]

Just the opposite. For example, Microsoft Power Platform allows "regular" users (actually, power users) to be able to build applications that were previously reserved for only custom built applications. The rise of AI has allowed small businesses to take advantage of AI in a way that they never could before. For example, I am a small business. I employ 4 people besides myself. We have been slowly adding English captions to our paid videos over the past two years. We did it slowly because it costs $1.25/minute to add captions to a video. That means that my C# Mastercourse cost over $5,000 to add captions to it. And the reality was that the company we used mostly just used an AI tool to add captions and then just tweaked the end results if needed. Now, I have direct access to the AI and we've been able to add captions to every video in my library. Next up is translations. Translations used to cost between $3 and $8 per minute based upon the language (and you had to have done the English captions first, so add another $1.25/minute to the cost). Now, AI has come along enough that we will be able to add translations to our content very soon in multiple languages for free. Now these captions and translations aren't perfect, but they are very close. And the result is a product that is more accessible to more people for very little cost (just our time and some validation we want to do). The future has always been one of making things easier, not harder. It may seem like things get harder, but what is actually happening is that we are doing more than we used to. For instance, in the early 2000's, deploying a website involved FTP. Scaling a website involved extremely complex and expensive processes and didn't typically include global scaling, just scaling at a particular location. Now, for $9/month, my website ( www.iamtimcorey.com ) is deployed automatically based upon a push to the prod branch in our repo and is globally scaled and replicated automatically. I can bring up a new site like this in under 10 minutes (from code to automatically deployed and replicated with a custom domain name and SSL certificate). That is an incredible reduction in complexity. It is also an incredibly affordable solution for small businesses.

[@annagulaev]

@@IAmTimCorey You are talking about using tools. I am talking about programming. I'm talking about using a framework that is more complex than the problem you are solving with it. I'm talking about anyone can launch a website, but custom programming gets a project thrown to code farms, who CAN make use of complex tools because they do the same thing over and over, but then the business is reliant on that technology, and is SOL when it is no longer the hotness. I am talking about business logic, the parts that are not boiler plate, completely stumping modern programmers. I'm talking about programmers doing resume programming, leaving their employers reliant on tech that will need to be supported after they've landed their next gig...tech that will be obsolete before the company is done using it. I'm talking about turning every little to-do item into a research project and an excuse to add something to your resume, and taking forever to complete it.

[@BrentHollett]

The "future" is AI taking on the role of a programmer to allow less skilled people to produce FFP products, at lower costs.

[@BrentHollett]

@@Luke5100 you're mistaking "Any old person can do open heart surgery". The AI will do the surgery, flawlessly, and humans will be in the Adminstrative roles to ensure the client pays. Robot arms mean any surgeon can do surgery from any distance. AI means you don't need a surgeon.

[@IAmTimCorey]

@annagulaev - I'm not sure what you are talking about. That's not at all what is happening. If you think that is what is happening, you've either been badly misinformed or you have a very incorrect perspective on the matter. @BrentHollett - That almost certainly won't happen in the next 20 years. Maybe not in our lifetime. AI is not some type of all-knowing thing. In fact, it has proven to be a lot dumber than we had hoped, even if it is an incredible tool. For instance, they've got GPT-4 working well. That was an exponential leap in inputs, which led to much better outputs. However, they've found that adding even more inputs actually makes the model worse, not better. Even worse, they aren't sure why or what to do about it. But let's get practical here for a minute. AI-driven heart surgery would be complex. More complex than full self-driving (FSD). So let's look at how that is going. Driving on the highway is fine. Even city driving is mostly ok. However, the thing that AI can't really handle is unique cases. For instance, what happens if the road is covered by snow? What if there is construction and the lines on the road don't match up with where you need to drive? What if there is an accident scene and the police are redirecting traffic into the "wrong" lane to get around it? Our world is full of complexities and edge-cases. Trying to account for all of them is nearly impossible. Trying to get an AI to do it for us is even worse. Now ramp up the complexity of open-heart surgery, where every situation is a bit different, where there can be any number of complications, and where a single mistake can cost a life. Do you think that will be something that AI can do anytime soon? Nope. We can't even rely on AI to correctly answer a simple question and not lie about the answer.

[@niksatan]

I think there would be no significant change for Software Development in the future - ChatGPT will maybe replace Google, and that's it.

[@IAmTimCorey]

There will definitely be significant changes in software development in the future. Over the past 20 years, we've had multiple disruptions. Moving forward won't be any different.

[@SuperDrJeckyl]

There are ALWAYS significant changes.

[@williamFrSFO16]

I’ve been in IT, since 1978. At that time, COBOL was the language to get good at. Programs were written and modified using an IBM keypunch machine, executed on an IBM 360 Mainframe, with data input on cards from Data Entry. (Come a long way, eh?) After the turn of the Century, ALL of it was “dead,” but the logic. It was up to the programmer, to learn new languages to execute on networks. What Tim says is perfectly clear, correct! The variety and and and use the programming has changed so dramatically that it’s everybody’s life and could be picked from that part of peoples lives. I am wondering where AI will have a fit for the future of programming. Will AI be programming itself? If so, then, why are we going to do? Thanks again, Tim!

[@IAmTimCorey]

@williamturnley282 - I believe @Luke5100 was referring to @niksatan with his comments, not you.

[@williamlturnley282]

@@IAmTimCorey Hi Tim! Thanks, it was a real long day at work. My apologies to @Luke5100! I deleted the errant post.

[@pippaloves]

Power Platform is the future.

[@IAmTimCorey]

It is part of the future, but not the whole one (nor a large part of the future). What people find after the initial hype is that "regular" people aren't good developers typically. Not everyone has practiced and honed their logic skills. There is more to development than syntax, and that is basically what Power Platform replaces - the syntax. You still need to know what questions to ask and how to create the desired outcome without unintended consequences.

[@dasfahrer8187]

Well, this was a waste of time.

[@IAmTimCorey]

What were you expecting? Half-baked predictions that are sure to be wrong?