The software that ate itself
It was reported this week project management software developer Atlassian, Australia’s largest software company, is cutting 10% of its workforce. In addition, database giant Oracle is also planning a fresh round of layoffs, hailing the efficiency of artificial intelligence (AI) ‘coding’ and, apparently, to free up cash to pay for a data centre.
They are not alone. One Website, jobloss.ai is attempting to track AI-related layoffs, breaking out both cases where the connection has been explicitly made by management and ones where vague gestures have been made toward the technology.
For his part, Atlassian’s chief executive Mike Cannon-Brookes appeared unable to truly stand over his decision or else has been, mistakenly, informed that deploying the language of therapy is appropriate when firing people. In a video message, Cannon-Brookes told remaining workers to “be kind to yourselves and others [and] check in on your teammates,” rubbing salt into the wounds of new P45-holders by saying “you are an important part of the Atlassian story”.
Atlassian’s internal communications channel on the application Slack was kept open “on mobile devices for 6-12 hours (depending on timezones) for impacted employees who want to say goodbye to teammates across the world”.
By this point we are all well aware of the contours of the arguments about AI and job losses, but a complicated – and, frankly, incoherent – tone is emerging as businesses scramble to either defend their AI spending or else fend off the challenge posed by the technology. What is perhaps less well-understood is that AI deployment is revealing some serious contradictions in labour relations in general.
Despite a tight labour market (in plain English: worker shortages) tech companies in particular seem hellbent on making do with less. Or at least making do with more money by making do with fewer staff. The tech sector has been racking up layoffs at precisely the same time as complaining about a labour shortage.
The myth of productivity
At the very least, we are seeing an interesting reversal in messaging: out go top chefs, ping pong tables and the general presentation of work as some kind of ludic practice. In its stead comes a grim lockstep march driven by the abstraction that is data (masquerading as knowledge) and endless labour with no time allotted to attend to its meaning.
Beneath that, though, is an even starker irony: AI isn’t just taking jobs, it is collapsing the specific business model that created those jobs. Atlassian’s products, Jira and Confluence, and project management software broadly, are products whose premise lies in large numbers of human workers who need coordination tools. Fewer workers, or AI that can handle coordination tasks, means fewer users. Seen this way, Atlassian is racing to get ahead of a disruption to its own revenue model.
Indeed, the company has lost more than half its market value since the start of 2026 as investors fear AI will make the company’s software obsolete. Whether or not this is actually the likely outcome is unclear, but widespread job cuts will only accelerate the arrival of that particular future.
As I have argued before, it is true that software companies share prices are on the slide because AI can replicate their functionality, but whether or not enterprises trust machine-developed homebrew software is another matter. In either case, though, the more fundamental question, and one that is in dire need of attention, is what it means to organise human labour around tools that presuppose human labour is itself dispensable.
Subscribers 0
Fans 0
Followers 0
Followers