Solving the Tech Worker Shortage With Tech

Xero CEO Steve Vamos took some heat from the market late last year when the company missed expectations on profit. The miss was because the company added 1,000 employees in the year to end-September, 30 per cent more than in the previous year. Vamos called it investing for growth.

This is the investment theme for 2022 and beyond: a need for tech workers. The massive demand comes from the search for new opportunities and to create new efficiencies but also from immigration restrictions that has stymied an inflow of educated workers.

But the emergence of low-code tools, machine learning and AI assistance can improve efficiencies — possibly helping companies like Xero in the future do more with less.

Developer time is often spent in the mire of unproductive work: 4.9 billion hours a year are wasted dealing with sub-standard code. Training, tooling, and new ideas are emerging to reduce the need for developers, or to reduce the cost of developer time.

Associated staff shouldn’t be overlooked: data scientists, product managers, product owners, designers, QA engineers and so on — all are needed to add features, fix bugs and make easy to use, helpful software. Developers and ancillary staff all attract significant total compensation in this hot market.

There’s an adage: in a gold rush, sell the shovels and pitchforks. In the tech version of the gold rush, look to the following ideas that may come about from this shortage of tech workers.

1. Bots, AI assistants, and tools to supplement developer skills to generate faster solutions;
2. No-code and low-code tools that offer point-and-click solutions for use cases for casuals and non-developers.
3. More traditional companies that offer developer training, support, and coaching;

AI Speeds Things Up

To augment developers comes machine learning (ML). In its current form, ML algorithms trained on billions of lines of code can help speed through basic programming necessities like syntax and common patterns.

Github, the world’s biggest software repository and tool, has an AI tool called Copilot. Trained on billions of lines of code, it speeds up basic processes -- it can turn English language into code by generating correct code from text inputs. Or it can watch logic being coded and quickly suggest autocorrects, autocompletes and entire middles and endings to functions based on boilerplate or patterns. Copilot was said to be helping with 40 per cent of new code added to Github.

OpenAI's GPT-3 can also generate complex code from regular text inputs (see this fairly mind-blowing example from a year ago), and GPT-4, expected this year or next, will see a further step-change, and there's already talk around GPT-5.

Tools like Copilot aid developers, though there are limitations. But the idea is a developer can get help from an AI to tackle common problems and avoid basic mistakes.

“There’s massive demand for digital solutions but not enough coders out there. There’s a million-developer shortfall in the US alone. So instead of making the world learn how to code, why don’t we make development environments speak the language of a normal human?” said Charles Lamanna, CVP of Microsoft’s Low Code Application Platform, to The Verge.

Github is owned by Microsoft, and it also invested in OpenAI to offer its services via the cloud. Others are working on artificial intelligence but Microsoft’s early investments appear to be well placed. It bought GitHub back in 2018 for $US7.5 billion.

“Prediction,” said Sam Altman, co-founder at OpenAI , back in June. “AI will cause the price of work that can happen in front of a computer to decrease much faster than the price of work that happens in the physical world. This is the opposite of what most people (including me) expected and will have strange effects.”

No-code and Low-code Tools

There are two or three approaches coming from what are called “no-code” tools: mostly online tools with simple drag-and-drop options that let almost anyone build competent web applications using a GUI interface. They also handle piping: the ability to connect different data sources and software tools together and do anything from building a website to ecommerce functions, to handling data flows from one service to another, without needing hands-on coding.

Low-code tools, as the name suggests, are similar to no-code but with more customisation for those who are able to tinker with fully-written code.

The opportunity here is obvious: get ideas live, fast, at very low cost, and without needing to learn much in the way of programming. Pieces of this have long existed such as WordPress but still had significant barriers to entry for non-tech users.

The next generation of tools from Wix, Squarespace, and Shopify are easier to use and solve more problems upfront. They have open app marketplaces on their platform for third-party apps. This has led to a supply of apps and tools that aim to solve most problems someone running an online shop might have, with support from major companies.

As an example, Australia Post has a Shopify app to help create paid shipping labels. For those looking to automate email marketing, or manage stock levels, and so on, there’s an app that can help — often paid, and the platforms clip the ticket as well.

Shopify, up 26 per cent in 2021, has been one of the major success stories of online shopping as the backbone of independent businesses, and is dual-listed in its home country of Canada on the TSX, and on the Nasdaq.

No-code/low-code tools also offer internal solutions for bigger companies. Tools like Airtable, Bubble, and Zapier are powerful add-ons for efficiency boosts and automation.

While these tools may see a reduction in the need for developers, they usually augment the capabilities of non-developers. This is an emerging stage: few of these tools have IPO-ed, and probably the closest to this stage would be Airtable and Zapier, both valued as Unicorns.

Making More Developers

Companies that have emerged to teach developers new skills or provide people with a path into software development have become mega corps, and the field remains evolving. It’s somewhat of a cottage industry, with a mix of on-premise and online course providers.

Companies like Codecademy teach budding developers via online courses. It has grown to a unicorn-size business from a startup. Udemy, an online learning and teaching marketplace, listed this year with a market cap at the time of writing at $US2.9 billion.

Bootcamps or coding schools like General Assembly (now owned by Swiss-firm Adecco) have risen through the need to teach more developers, and online learning company PluralSight was a listed company before being plucked off the Nasdaq by Vista Equity Partners.

The need to get more tech workers isn’t going away any time soon. Immigration restrictions may need to loosen to help solve Australia’s technical shortfall over the coming years. But the emergence of low-code tools, machine learning and AI assistance, faster ways to train new developers, combined with easier-to-use tools and collaborative support, can make development times faster and improve efficiencies and help companies do more with less.