Programming

Summary: Our industry is going through a shift where more business solutions are built using serverless models and higher-abstraction services causing a change in developer skill sets. This will lead to low/no-code platforms further accelerating this shift.

Whenever I ponder the future of serverless I end up in an existential rabbit hole about the future of programming. I can likely attribute this to my Twitter/conference echo chamber, but I feel like our industry is in this constant vortex of technical debate — whether it’s the Kubernetes sub-project du jour, or arguing about the definition of serverless, or hearing every vendor talk about their latest contributions to the world of overly-complex open source stuff that only they will ever understand how to operate.

This technology echo chamber misses the forest for the trees.

Escape this vortex for a bit and you realize that cloud services are getting better, faster, higher (up the stack) and proliferating. And it’s not all just for “developers” where a knowledge of JSON parsing and dependency management still makes even experienced programmers feel sad.

End-user “builder” services are proliferating — like Airtable for building “database-backed” applications without writing code, to Twilio’s Flex platform for building contact centers, to marketplaces of skills for home devices, to “drag and drop” platforms for life and code, to a plethora of customizable commerce platforms. The examples are endless.

All of these things are consumed as services, without thinking about, wait for it, servers. So is this serverless? Whatever. Call it Jeff. Wait someone already did that. Just don’t miss the forest.

And yes, we’ve been talking about no-code/low-code for decades, and every attempt has felt constrained and restricted. But then I remember Google App Engine in 2008 feeling too restricted, and guess what’s happening?

The following 10 years showed explosive growth for IaaS, and quietly the evolution of services higher up the stack is bringing us right back to the app engine vision.

Sometimes innovation skips big steps, and sometimes it has to follow a linear path upwards.

Another indicator no/low-code is coming is the Internet has become an incredible distribution vehicle for services. This wasn’t as true in 2008, where even SMTP and SMS were more niche ideas. Sendgrid and Twilio hadn’t entered the scene and defined API-centric developer evangelism for us yet. But now everything is a service. Try searching for some noun + “API”, like “Cat API” (The Cat API), or “Lamp API” (Philips Hue API). You get the picture. Of a cat.

This was the rationale for my Three Ghosts of Serverless talk. Ebenezer realizes his organization is not prepared for the shift in the way stuff is being built, and that his organization will die. Queue grave scene with bad acting on my part.

Then I read the “The creeping IT apocalypse” post from Forrest Brazeal which from a different viewpoint is telling the same Ebenezer story. That “[the cloud providers] are straight-up reducing the number of people required to deliver technical solutions.” Bah, humbug. And yes, I agree.

Eventually, an entire demographic of programmers who got into programming for the wrong reasons (money?) fade out and find new careers or become “business ops” using said technologies above.

That and a widening salary gap is already creating fair market value for skilled programmers who can solve more complex problems vs newer or less-skilled developers. But I digress. I’ll come back to this topic.

I’ve made bad predictions before. Like when I was 10, running The Sonic Realm BBS, proclaiming the Internet wasn’t going anywhere. So I’d like to hear from others. What do you think?

Chad

Find me on Twitter.