Suggestions
Alex Ewerlöf
Web architecture, reliability, security, privacy
Alex Ewerlöf is a seasoned engineer with a BSc in Hardware Engineering and an MSc in Systems Engineering from Sweden's top technical university, emphasizing the importance of continuous learning in the ever-evolving software landscape.
With over 20 years of experience, Alex has honed his skills in various domains such as robotics, mobile apps, banking systems, insurance, telecom, retail, fintech, medtech, news, and streaming.
Specializing in community building, fostering collaboration, dismantling silos, implementing lean processes, and promoting transparency within teams, Alex is adept at solving complex problems and driving innovation.
Throughout his career, Alex has held key roles in reputable organizations like Volvo Cars, Discovery Inc, Schibsted, RaySearch Laboratories, Bilitti, Klarna, Cobham, Rebtel, and more, showcasing a diverse and extensive professional background.
As a prolific writer, Alex shares his insights and guiding principles in a post reflecting on his 20 years of programming experience, underlining his commitment to growth and adaptability in the tech industry.
Highlights
There's always a good reason for adding more process and it's never about enabling human creativity but about reducing human error, bias, diversity, etc.
#Process adds #friction. You can use that fact deliberately where the cost of friction is less than the risk of the flexibility it blocks.
If you have to use process:
➡️ Keep it lightweight and efficient (cost/value) ➡️ Make sure to have functioning feedback channels to adjust the process as the system, service and consumers evolve ➡️ Regularly review it
Nothing spells "fake work" louder than outdated rituals.
- "We've always done it that way."
- "No thanks! I'm fine using my brain"
😢 Matt Garman (AWS CTO): AI will code instead of developers 😢 Sam Altman (OpenAI): No start-up can compete with us 😢 Jensen Huang (Nvidia): Don't bother learning to code 😢 Various "thought leaders": Programming is dead
🛑 If you think AI can take your job, maybe you're not good at your job anyway!
There is ALWSYS a need for good problem solvers who understand people and can define and validate human solutions.
The current AI tech has reached its limits with the available data and isn't reliable enough. But let's assume some breakthrough will make it exponentially smarter with the ability to "think" in abstract terms.
By definition, there's a gap between silicon brain and carbon brain. Western philosophy often detaches the brain from the rest of the body. But the body and the brain are a whole. AI cannot completely understand the human experience. It can mimic it with good approximation but it is by definition limited in how it interprets and interacts with the world. I'm not saying it's impossible to create super humans but when that happens we have more important problems than "who gets to code?" or "will I have a job?".
What can you do with this information? Here's my personal strategy:
🟢 Allocate a part of your time to experiment with AI. I've invested in a beefy machine to run LLMs at home and build with it. AI may not take your job, but a human equipped with AI may do that. You don't have to be a 10x engineer to keep your job. All you have to do is to maintain an edge over those who are stuck with their first impression and think they know it all. Gain first hand experience.
🟢 Always bet on the human side. I'm confident that we still have some time left before AI reaches or surpasses human intelligence. In the mean time, it's important to elevate your skills beyond coding. You can use AI to code in programming languages you've never worked with, you can use it as a learning tool, but you can hardly use it to think for you or build relationships and understand people. If you're one of those "coders" who made easy money by punching the keyboard, it's time to dust off your copy of Dale Carnegie (just an example) and build relationships and leadership skills.
🟢 If you're unemployed, you have a full-time job to upskill. deeplearning.ai has lots of free high quality courses. If you're employed, don't let your day job to steal your opportunity to stay relevant. I've seen it too many times that "clever" engineers try to add some AI to their current projects and force their employers to pay for their learning/experimentation. Don't! My personal approach is to go down in salary and working hours to experiment with these tools and build something from the ground up (I'll announce it here when it's ready for public. Hint: it's going to be about Service Reliability Engineering.)
Hope this helps. There are a lot of counter-arguments to be had because there's hype, uncertainty, fear, and confusion. Ask your questions in the comments.
