Vibe-Coding: Will Software Engineers Disappear by 2028?

“Vibe coding is already present in your company,” according to David Hsu, CEO of Retool and governance is the only way to secure it.

Let's put this into context. Think of a Fortune 500 company’s CIO who has recently come across a vibe-coded application spotted on the public internet. This app was developed by an employee of the Fortune 500 company in question. The developer obtained a data dump from Salesforce pertaining to customers, then fed this data to an external AI tool not part of the company’s set of AI tools. The employee prompted the tool to host the resulting application on a public URL. 

Since it was an external AI tool, a security review couldn’t be carried out, nor could the IT department view it. Without any internal visibility into a tool created by the employee, the actual customer data was live on the open web. It’s a scary situation, a breach of trust, so imagine the CIO’s wrath upon seeing it. 

This is why, in the recent episode of the Tech Transformed podcast, host Shubhangi Dua, Podcast Producer and B2B Tech Journalist at EM360Tech, sat down for a conversation with David Hsu, CEO and Founder of Retool. They talk about the repercussions of vibe coding in addition to examining why most enterprises have no conception of how much of it is already taking place within their own enterprise. 

The discussion then turns from the immediate issue of governance concerning uncontrolled AI app-building to Hsu's prediction that software engineering as a profession could vanish within two to three years.

“All the CIOs that I speak to are genuinely frightened about vibe coding, and even if you don't want to admit it, vibe coding is already present in your company,” he tells Dua. This is partly why Hsu founded Retool. 

Retool secures vibe coding for enterprises. Hsu explains that Retool can be connected to a company’s Snowflake database, for instance or their Databricks data lake. “Users should be able to have access to particular tables only,” he says. Access controls are required at the group level. 

“If you're in this particular Okta group, you can only access this table, and it's read-only. Whereas if you're in this other Okta group, you can actually write back to this other Salesforce over here." So one Okta group could be granted read-only access to a particular table, while another group might be given write access to a connected Salesforce instance. All applications that are built on top of that connection automatically inherit the same restrictions.

Once rules of such a nature are put in place, people could then be set free, explaining why he views governance as something that enables freedom rather than something that restricts. 

Why Governance Enables Freedom

The CEO of Retool says that vibe coding cannot be stopped; however, it can only be directed. People are already using their personal email addresses to sign up for AI coding tools. They link them to enterprise data and then harness whatever the tools produce. The critical point for enterprises is that banning this practice only pushes it outside IT's oversight.

Hsu flags that when employees vibe-code today, vibe-coding platforms aren't centralised. “Instead of allowing a Large Language Model (LLM) to explore your company data and use other data or have people download CSVs and upload to them, which is what they're gonna do if you don't give them access in a successful way.”

In such a case, since IT is already aware that the hosting environment, the audit trail, and the data connections are secure, employees will no longer need permission for each individual application and will therefore be able to develop freely within a system which has already been centrally secured.

Why is the Token Spend So High? 

In addition to the problem of having too many applications, Hsu pointed out another gap in governance: cost. Most AI coding assistants such as ChatGPT, ChatGPT Work, Claude, and Claude Code function as single-player tools on a person's own local machine. As a result, when a CIO's bill goes up, there is usually no means of working out which teams, projects, or chat sessions are actually causing the expenditure without having to interview the employees individually.

A centralised platform reverses this situation as all activity goes through a single system. IT can see precisely which automation, which agent, or which individual chat session is using the most tokens and can therefore set model-access policies by group. For example, limiting one Okta group to cheaper open-source models while granting the C-suite access to frontier models.

'Software Engineering May Not Exist as a Profession Anymore'

When Dua asked Hsu to describe the enterprise work environment in 2030, he said that software engineering may not exist as a sole profession anymore. Hsu referred to forecasts such as the widely circulated AI 2027 scenario, which predicts that fully automated coding will take place between late 2027 and early 2028.

“It's not that software engineers are working alongside agents; instead, I think that software engineers themselves could have disappeared by 2028,” he states. In his view, rather than being augmented by AI agents, the job is in fact replaced by them, so that people who are not software engineers gain the ability to create applications which used to require hiring a developer.

When asked about the future of the CIO role, the founder of Retool says that by 2030 90 per cent of the work within a typical company will be carried out by AI employees rather than by humans, whether it be through recurring automated workflows or via fully autonomous agents. The CIO's role changes to that of a "chief people officer for AIs.” 

The chief people officer for AIs role would involve making decisions on which agents are hired, fired and given access to particular data, and covering what can be described as onboarding through context engineering.

Overall, the idea put forward by Retool is that vibe coding is not something that companies can simply allow to run unchecked, nor can they prohibit it. However, it’s already being used on employees' laptops, frequently with access to actual company data, and there is no central oversight available. 

Hsu's solution is to centralise the deployment of applications that are built using AI, so that IT can set the security and data-access rules just once and then employees are free to build on top of them. 

The more significant assertion behind this proposal is that the governance layer will become even more important as software engineering itself becomes automated, possibly within two to three years, which will in turn shift the CIO's role to that of managing a workforce that is mostly made up of non-human entities.

Takeaways

• Vibe coding is already happening inside most enterprises, sanctioned or not
• A Fortune 500 CIO found a vibe-coded app exposed on the public internet
• One customer's AI coding spend hit 30x their original budget estimate
• Two employees independently built duplicate apps that returned different numbers
• Hsu forecasts fully automated coding arriving by late 2027 or early 2028
• By 2030, Hsu estimates 90% of company work will be done by AI employees
• The CIO's future role: a "chief people officer" governing AI agents, not humans