Mark recalls the moment he began to question whether Block’s AI tools might take over his job—or even replace him entirely. This occurred during an elaborate anniversary celebration hosted by his fintech company last September. As executives showcased the productivity advantages of a new internal AI tool, Mark, then part of the product department, shared his apprehensions with coworkers. He expressed skepticism about the technology’s capability to independently advance the business without the involvement of employees like himself in steering vision and strategy.
These AI systems were not self-sufficient; he had to guide them. Mark believed that his role was still essential.
“You can’t really AI that,” he told the Guardian, adding, “an employee is more than a series of tasks.”
Unfortunately, Mark was one of approximately 4,000 Block employees who were laid off last week. CEO Jack Dorsey stated that the substantial staff cuts were a result of enhancements in AI productivity. “A significantly smaller team, using the tools we’re building, can do more and do it better,” Dorsey explained in a letter to shareholders.
However, in conversations with the Guardian, seven current and recently laid-off employees challenged Dorsey’s claims that existing AI technology can replace workers on such a large scale. Most of these workers, representing various departments including engineering and product, spoke anonymously to protect their jobs and severance. While they acknowledged that Block’s AI tools could help in their roles, many felt the layoffs were a strategic move by Dorsey to restore investor confidence following a decline in Block’s stock due to significant investments in a volatile cryptocurrency market.
George, a current Block employee, stated that the layoffs were “posturing for the market” and emphasized that investors perceived Dorsey as a weak leader: “This was a bold move to shift the company’s focus from crypto to AI and alter the public market’s narrative surrounding the organization.” Following the layoffs announcement, Block’s stock experienced an increase.
In a recent Wired interview, Dorsey explained that he drastically reduced the workforce because “something significantly shifted in December in the sophistication of [AI] tools, including Anthropic’s Opus 4.6 and OpenAI’s Codex 5.3.” He rebuffed accusations of over-hiring during the COVID-19 pandemic, asserting that the company was performing on par with or ahead of its peers in terms of gross profit per employee. Dorsey remarked that the structure and hierarchy of companies often hinder their effectiveness. He conveyed his goal for Block to someday function like a “mini AGI.” Block did not issue a statement to the Guardian.
The layoffs at Block coincide with growing concerns regarding the effect of AI adoption on employment in the United States. A Goldman Sachs report from February highlighted the potential for AI to drive unemployment upward, estimating that it led to 5,000 to 10,000 net job losses per month last year.
Creating Their Own Replacements
Initially, Block encouraged its employees to leverage AI technology more frequently. However, in the past nine months, that encouragement evolved into a mandate, workers shared with the Guardian. Dorsey emphasized in a January all-hands meeting that “the way we built things in the past is not going to work anymore.”
“We have to shift. There’s no question,” Dorsey affirmed.
Some employees, including Mark, believe they are being asked to create and train the very tools that the company plans to use to replace them.
“The way they are using these tools as a rationale for firing half the workforce is preposterous,” Mark stated. “In retrospect, it resembled a thinly veiled strategy to solicit input from employees on which tasks should be automated. You essentially have staff instructing management on how to automate them out of their positions… yet these tools are not nearly comprehensive enough to encompass someone’s full role.” Another former Block employee echoed this viewpoint in a public interview with Business Insider.
Even those employees whose jobs revolve heavily around AI tools express skepticism about the capacity for current tools to replace workers on this scale. “We’re just not there yet,” asserts John, who aids his colleagues in using AI.
“There’s a difference between what’s technically feasible and, forgive my French, the CEO’s overreaching interpretations of how AI actually works,” John remarked.
While AI tools have undoubtedly accelerated some engineering tasks, human involvement is still necessary. Block executives recently announced during an earnings call that they have observed significant increases in productivity, stating that engineering tasks which once took weeks could now be accomplished by small teams in mere fractions of the time thanks to “agentic coding tools.” They noted a “greater than 40% increase in production code delivered per engineer since September.”
Nevertheless, John clarified that all code changes at Block still require human approval before they can be deployed. He pointed out that as of three months ago, about 95% of AI-driven code revisions necessitate human adjustments to meet company standards. “They’re not up to snuff on the first attempt,” he explained.
Block is also closely monitoring employees’ AI usage, tracking specific tools and tokens, as indicated by several staff members. Evaluations of employee performance, which partially rely on manager assessments, now incorporate inquiries regarding proficiency in AI usage.
Liam, a recently laid-off software engineer, recalls feeling the pressure from his manager as he was questioned about his current AI usage and what actions he could take to improve it. “It was abundantly clear that if you weren’t utilizing AI, your job was at risk,” he said.
The swift shift towards AI has led to widespread frustration, as reported by John, who supports others in using the technology: “People are exhausted by AI.”
Carl, another current employee, shares his ethical opposition to AI due to the detrimental impact of the data centers that power it on local communities. He refuses to engage with these tools, stating, “You’re not compensating me to train your equipment, so I’m not going to do it.”
However, even those who might have been more receptive to AI find themselves frustrated. The enforced integration of the technology has “created more obstacles,” as employees recognize they were being monitored for their usage and ordered to adopt it, even when it was the less effective option, according to Oliver, a recently laid-off worker. He, along with others, indicated that Block’s AI systems are not yet capable of taking the lead on tasks in highly regulated areas, such as banking and money transfers—crucial aspects of a financial technology firm’s operations.
Naoko Takeda, formerly a data scientist at Cash App, which is part of Block, recently posted on LinkedIn about surviving the layoffs while expressing feelings of overwhelming dread and survivor’s guilt. Despite a significant pay raise offered to remaining employees, she chose to leave the company.
“Over the past year, AI was essentially thrust on everyone. Everything revolved around AI. We were instructed to employ AI as much as possible,” she lamented. “It feels incredibly dystopian to be forced to use tools that hasten the disappearance of the jobs we rely on for our livelihoods. Personally, I didn’t see any substantial productivity enhancements from AI, nothing nearly significant enough to justify the termination of half the company’s workforce along with their invaluable institutional knowledge and expertise.”
Are Bots Bad for Business?
Aside from the implications for employees, Block’s expansion of AI could negatively impact its business. George, still a team member, describes how some customers have expressed dissatisfaction with the company outsourcing initial customer support requests to chatbots.
“We’ve seen from internal surveys that [these bots have] made significant errors,” he mentions, noting instances such as advising customers to cancel or close their existing accounts—a suggestion the company would obviously never endorse.
More generally, employees articulated that although AI is beneficial in back-end processes, customers often do not prefer interacting with automated bots, particularly for serious issues. “It’s frustrating; it feels like you can’t convey your point… it’s almost like they’re reading from a manual, and this isn’t the real problem,” Carl explained.
Other staff appreciate the productivity benefits AI brings, but they highlight its lack of discernment and emotional intelligence. “It can construct a building, but does it comprehend architecture?” commented Oliver.
Faced with these challenges, the remaining employees find themselves stretched thin, as entire teams have been decimated. Current staff describe being in “survival mode” with morale reported as “in the gutter.” An internal Slack message from Block’s engineering lead, expressed gratitude to employees following the layoffs, but received a mix of hundreds of thumbs-downs, tomatoes, middle fingers, and clown emojis in response from the staff.
“Everyone I know who remains feels overwhelmed because they now realize their workload has increased fourfold or tenfold, and AI is not going to resolve that,” Oliver lamented.
