Why ChatGPT Won’t Kill Coders

In Will ChatGPT Kill Coders?, we examined many demand suppressors that threatened to kill coders’ jobs. We also alluded to DEMAND STIMULANTS, which work as counterweights by boosting demand for coders.

In this post, we will do a deep dive into Demand Stimulants.

Comprising new computing paradigms and markets, demand stimulants not only arrest the decline of coding jobs but increase the headcount of coders.

NEW COMPUTING PARADIGMS

Anyone who has followed the IT industry for the last two decades or longer would have observed many waves of computing paradigms like Web, Mobile, Social, Cloud, and Blockchain alongside the growth of ERP, RAD, AI, and other demand suppressors described in our previous post.

Every new wave of computing paradigm created new demand for coders.

  • Web: A substantial portion of client-server architecture-based ERP had to be extended to support internet-based transactions with suppliers, customers and other stakeholders.
  • Mobile: Many existing applications like CRM had to be “mobilized” (or “mobified” if you prefer).
  • Cloud: Onprem applications had to be migrated to hyperscaler clouds like AWS, Azure, and Oracle Cloud Infrastructure.

Gen AI is the latest wave. By automating coding to some extent, it will suppress demand for coders. However, by requiring training of AI on industry- and enterprise-specific data – “last mile training” as Oracle calls it – Gen AI will also stimulate demand for coders.

NEW MARKETS

Over the decades, we have seen a mushrooming of software product companies and SAAS companies, and waves of digital transformation, and consumer tech.

They created new markets for coding jobs such as:

  • Engineering orgs of COTS (Commercial Off The Shelf) and SAAS companies
  • Tools (RAD / Low Code) vendors including open source software and WordPress Plugins.
  • Gen AI platform developers
  • DX practice of Big 4 and other consulting companies
  • Consumer Internet startups in fintech, food delivery, rideshare, roomshare, travel, and other industries.
  • Software Is The Brand companies. Coined by Forrester, the term SITB refers to the practice among banking, engineering and other non-software industries to differentiate themselves via software. Examples of SITB include Trade Finance and High Value Fund Transfer software in banks, and Internet of Things (IoT) in manufacturing automation companies.
@mattturck: Whatever happened to the Internet of Things? 10 years ago, IoT was the next big thing. Tons of new startups, VC money and hype. Ended up producing just *one* currently independent public company today, Samsara.
@s_ketharaman: It could be argued that IoT has been in use in chemical process plants etc. for 40+ years in the form of sensors, DCS, and PLCs. At most, these products will need to be upgraded to ZigBee and other open Internet protocols to make them true IoT. But those opportunities will likely be tapped by Honeywell, Schneider and other existing process control equipment suppliers and may not create opportunities for new startups.

  • Infusion of AI into the products of non-software companies. e.g. Siemens Industrial Copilot.
Microsoft and Siemens have come up with a manufacturing assistant called the Siemens Industrial Copilot, which will… assist maintenance staff with repairs. – @superglaze.


If recent history is any guide, I predict that Gen AI will create more coding jobs. At most, it might lead to a displacement of jobs from engineering to professional services organizations (PSO). (For the uninitiated, coders who develop the core product on the roadmap of a COTS or SAAS company are in the engineering org, and coders who implement or support the product for a specific customer of such a company or its implementation partners are in the professional services org.)

That said, we must recognize one major difference between Generative AI and the demand suppressors that came before it: ERP, RAD / Low Code etc. obviated the need for code. They did not produce what coders produced. Whereas Gen AI does not obviate the need for code. It produces what coders would otherwise have produced.

@vkhosla: Fun story – at a recent coding hackathon, the winner was a marketer. Not a coder, but a marketer. Previously it was the best coder who won, less dependent on idea and more on execution and ability to jerry rig a solution. Now it’s the best idea that wins since we’ve democratized what it means to program and barriers to entry are much lower (i.e. natural language!)

To that extent, Gen AI does have a direct competitive posturing against coders, designers, and architects, unlike demand suppressors of the past.

Accordingly, it could be argued that the new demand for coders driven by new computing paradigms and markets discussed above can be fulfilled by ChatGPT as well.

Going by my personal experience of using ChatGPT to get readymade code, Gen AI is bound to diminish demand for plain vanilla HTML CSS coding skills.

Moving on to complex websites and applications. Let’s take the following example: 

Me: Code for Uber clone?
ChatGPT: Creating a full Uber clone is a complex and extensive project that involves multiple components, including mobile apps for passengers and drivers, a backend server, payment processing, mapping and navigation integration, and more. It’s a project that would typically require a team of developers and significant resources. I can provide you with a high-level overview of the components you would need and some sample code snippets to get you started, but keep in mind that building a full Uber clone is beyond the scope of a simple code snippet…

You can see the full response here. While ChatGPT did give code snippets for many building blocks of a Uber Clone app, it’s clear from the above extract that coders will still be required to build advanced websites and applications.

What coders can do is to use Gen AI personal assistants to write better quality code faster. In this capacity, Gen AI will supplement – rather than replace – midrange and senior level coders. Gartner has a couple of pro tips for them in this context:

  • AI-assisted software engineering improves developer productivity and enables development teams to address this increasing demand for software to run the business.
  • AI-infused development tools allow software engineers to spend less time writing code, facilitating an increased focus on higher level activities, such as the design and composition of compelling business applications.

While I’ve not had a chance to use it, I hear good things about the Microsoft Github Copilot coding assistant.

If all else fails, in a lighter vein, remember the old Silicon Valley adage:

A well run Tech Company is 2X overstaffed; a badly run Tech Company is 4X overstaffed.

Coders don’t need any more assurance of job security than that!


I can hear coders grumbling that crafting demand stimulants is beyond their skillsets.

They’re right. Like before, product managers and marketers aka normies will create the new computing paradigms and markets sparking greater demand for coders in the age of Gen AI. However, they will need to be supported by coders, designers and architects aka geeks in this endeavor.

Ergo Generative AI presents a unique opportunity for geeks to collaborate with normies.

Disclosure: Oracle is ex-employer.

Comments

No comments yet. Why don’t you start the discussion?

Leave a Reply