“The world is changing: I feel it in the water, I feel it in the earth, and I smell it in the air.”
These are our times….
We live in both exciting and terrifying times. We have a covid crisis, war in Ukraine, a resource crisis, an energy crisis, an environmental crisis, deglobalization and why not, another banking crisis. People are worried if they can pay all their bills and feel left alone by their government which leads to civil unrest. Taking that all into account doesn’t this sound like a great moment for Artificial Intelligence to wreak havoc and start destroying the (IT) job market?
Are you chatGPT-tired already?
The ChatGPT hype is reaching new heights, even for a hype-fueled industry like our IT industry which has seen its fair share of big hypes the last 10 years. The chatGPT hype is different from these earlier ones because it has put more emphasis on the threats instead of just the opportunities. Ethics, bias and job security have become important topics.
IT has been infamous for killing jobs in the last 20+ years but did anyone pay attention? We now stream movies and series instead of renting a videotape or DVD (blu ray?) from a video rental. Do you still buy books in a bookstore? Is there still someone in your office who brings you the mail (not e-mail.)
In the top 20 lists of jobs which will be replaced by AI, no IT engineering roles have been mentioned (yet). This makes no sense, current AI is getting better and better in natural language processing. It should have no problems at all to be a lot better than any human in handling structured languages like programming languages and scripting languages.
I believe that the IT Industry will see significant changes in the coming 12 months already.
So, how will AI affect software development?
I see both threats as opportunities in our field. Obviously I asked chatGPT to tell me more about it and I got some boring and too obvious answers without saying anything. Hence; there is still hope for mankind!
sidenote: as a Tesla-owner-victim I have been fooled for 9 years about their “self driving” capabilities. In the end, co-pilot was at best an assistant.
Have a look at the following threats and opportunities:
- No more coding assessments
- The 1000x developer
- Fostering new engineers
- Buying or building software, revisited
- Goodbye Scrum! 🙂
- Venture capital
- are we doomed or is it a brave new world?
No more coding assessments!
Coding assessments suck. Yup, I don’t believe in them because you can easily fake them and most assignments can be found online. These “small tasks” always contain some “nice to haves” which actually mean more work.
With the new AI technologies, you no longer have to spend much time on it. Follow these steps and you’re done in an hour:
- Ask the AI to write the code
- (add some nice to haves like comments and style preferences (e.g. you don’t like if-else structures.)
- Ask for some funky things to show off (chose wisely, padawan)
- Put all in your IDE and run a code-beautifier over all code: yeah, we need proper code-styling.
- Compile, check the tests and submit
If you would be asked on the spot in a coding interview you can use AI as well. They probably won’t have a compiler with them to check it anyway.. Wait… They might use AI as well to verify!
The 1X developer is irrelevant within 12 months
Hello brand new day of productivity! Gone are the days of working long days and nights to deliver the next release of your product. The time is here for everyone to become a 1000X developer.
With the aid of AI every software engineer is suddenly coding on steroids. Within 12 months you can expect 40% of your code (mostly boilerplate) to be generated by AI. This will quickly increase over all parts of the life cycle. Will it generate 100% of the code? Probably not but don’t be surprised.
Which will cause some interesting side effects:
- Working less hours, but delivering more value: better work/life balance and less stress.
- Higher quality with less defects (let the AI verify for you while you take a plunge in the pool)
- Scarcity in the IT workforce is decreasing.
- Be on the lookout for new tooling which helps
- What coding solution you use for productivity is no longer required.
- Legacy code bases can quickly be modernized into newer tech stacks
- (this sounds painful) just keep the old/legacy codebase but ask the AI to restructure and improve it.
- You outsource software development to AI, instead of people.
- Stop writing maintainable code. People will not maintain it anymore; AI will take care of it.
If you haven’t started to integrate AI into your daily engineering activities, do it asap. Don’t become irrelevant but instead become a super-version of yourself: the Tony Stark of your company.
How to grow new engineers?
When Junior engineers start their career, they start with “simpler” or smaller tasks. These tasks are replaced by AI driven development tooling. Our task is to fight this exponential growing learning curve. Becoming a senior engineer takes time (and sweat and tears and a lot of patience.) No longer can we gradually increase complexity or abstractions to help grow our greatest talents into the new heroes. You could say; let the AI train them and tell them how it works and how it should be done. Theory will get you half way, practice is another ball game. I have not met a senior engineer who didn’t experience the frustration of “it should work, but it doesn’t and I’m completely lost.” The beauty of being stuck and getting yourself unstuck can not be taught, it must be given to you at the exactly wrong moment; just before go-live!
The buy over build principle; no longer a dilemma
In software engineering the BoB principle helps organizations to choose between building software themselves or buy it instead. There are pros and cons for both options but in reality it is a dilemma; choosing for tailor-made software results in higher costs and longer time to market and the need for a team to maintain the software. Choosing for off-the-shelf software results in tedious procurement processes and changing your own processes to make it fit with the product philosophy. Embedding it in your organization (organizational, processes and technological) can be time consuming.
When you have a team of 1000x developers, you no longer have to worry about the cons of tailor made software. As long as you have a clear understanding of the requirements, you are good to go. Once done, let an AI-powered DevOps service maintain your solution (including fixing/optimizing and releasing) and you are in full control, without the recurring discussions on increased license costs.
Goodbye SCRUM, hello AI Driven Development
SCRUM is both loved and hated by many engineers these days. Yes it has it good things but it it is meeting-heavy and focused too much on the rituals instead of the real goals (self learning and self managing teams)
With AI now powering more and more of your tools and services, SCRUM is no longer good enough. Why wait for 2 weeks to release a new version when you’re done today.
Yes today; that means, you discuss the requirements in the morning and that results in a first version being generated. You fine-tune before lunch. In the afternoon it’s being tested and a new version goes live when done.
New issues arise though:
- Not enough requirements for the team to keep them busy
- Requirements gathering remains challenging.
- Smaller teams and cross-functional roles
- (Biz)Dev(Sec)Ops teams need to rethink how to maintain fast-growing and quickly changing software.
1 month of AI Driven Development > 10 years of old fashioned Software development (SCRUM)
Venture capital: “Houston, we have a problem”
Software product companies have been hyped for the last few years. With software being so easily developed by AI with a little bit of help from the new Tony Stark of IT (you), what will it do to valuations of these companies? Software development runs the risk of becoming less valuable. For great software products, there will always be room in the market but you can expect demand for it to decrease in the coming years. The new type of software products companies might very well be in the models which generate the product, unique for the customer. Those who can combine a great product with client-specific needs will be rocking the show!
Brave new world
We are on the verge of entering a new era in IT. This will create a lot of buzz and noise. It will be our task to find the new standards of the future. There is not a day without new AI tools appearing and dissappearing. A lot of things will be shaken up and many certainties will expire.
For me it feels a bit the same as the rise of the Internet, many years ago. There were so many opportunities and so many wrong bets. Many of the big names have ceased to exist; can you still remember AltaVista? The AI bubble is growing quickly and is fueled with a lot of hype. What megacorps will lose relevancy when the AI bubble will collapse?
What are your thoughts about AI on our industry?
Who am I?
I’m Freek van Gool, and I proudly serve as the CTO at techspire. Our team is on a mission to elevate the standards of the IT industry and create a more sustainable future for the generations that follow us. With over 20 years of experience as a passionate technologist, I have faced countless challenges and tackled impossible problems head-on. Whether it’s building complex software systems or finding innovative solutions to thorny technical issues, I thrive on the excitement and satisfaction of solving difficult problems. As a persevering optimist, I always look on the bright side of things, whether I love running, grilling steaks and spending time with my feline friends. And let me tell you, I am convinced that the IT industry is hands down the best industry to work in. Every day, I feel a spark of passion and excitement in the very tips of my toes, and I can’t wait to see what the future holds for us at Techspire.