Technologically Skwair ▢ blog
|
Print by Marian Hardey. We Are All Squares. March 2020.
Print, Resistance! Designed by Mariann Hardey, 2025
The most terrifying sound in the technology industry today is not the roar of a hostile algorithm or the crash of a market correction; it is the silence of the woman who has just decided that speaking up is no longer worth the risk. She has disappeared herself, the brilliant mind who has quietly calculated the cost of her visibility and found the price too high. She has realised that while she was busy doing the heavy lifting of diversity work, the water around her had filled with sharks.
When I presented my latest evidence before the Coalition for Academic Scientific Computation (CASC) recently, I opened with an image that often lurks in the subconscious of every underrepresented person in our field. The shark seemed a fitting image. It could represent a specific person. It could also be interpreted as a caricature of a bad boss or a hostile colleague or a politician. To me, the shark represented the water we are now swimming in. It represented a danger that does not need to bite to be effective because it just needs to be visible enough to make us afraid to move. My talks to both CASC in the US and a version of the same talk to the Exobiosim/HPC group in the UK were driven by urgent data gathered this year from interviews with women working in computing, predominantly HPC, who are watching the tide turn against them. In these sessions, we mapped the anatomy of this new hostility. We discussed how diversity work has historically relied on a model of "good citizenship" as a volunteer-based "vibe" without actual resources or institutional protection. This precarious model is now collapsing under the weight of leadership hostility and resource cuts. My slides, which I share below, document the direct quotes from participants who feel they have "gone back decades," who describe the air as "thick with unspoken threats," and who see former allies retreating into silence to protect their own careers. We categorised the external threats, which ranged from legal challenges to "anti-woke" political pressure, but the most chilling finding was the internal retreat: the self-censorship of women who no longer feel safe to advocate for themselves or others. We are witnessing the erosion of allyship in real-time, leaving the most vulnerable to navigate these shark-infested waters alone.
For decades, the work of diversity in technology has been a slow and arduous swim upstream. We told ourselves that if we just worked harder, if we just leaned in, if we just mentored enough girls, the current would eventually change. But recently the current has not just stalled. It has reversed. We are no longer just fighting against inertia. We are fighting against a stark cultural shift fuelled by fear, legal threats, and a political climate that has turned equity into a dirty word. The result is a phenomenon that is perhaps more dangerous than the external attacks themselves. It is the silence of self-censorship.
This silence is the sound of survival. In my research, I spoke to women and marginalised groups across working in international teams. I heard that it feels like we have gone back decades. I heard that people are afraid to speak up because they fear repercussions. This is not a knee-jerk reaction. It is a calculated act of self-preservation in an ecosystem that has suddenly become hostile to our existence. We are seeing a retreat from DEI and EDIA initiatives not just in the White House but in the boardrooms of major corporations and in the quiet hallways of our own universities. Allies who were vocal two years ago are now waiting to see which way the wind blows, engaging in what my participants described as a calculated silence. They are testing the water while we are drowning in it. This retreat is documenting itself through a digital disappearing act. We are witnessing a systematic "going dark" of EDIA resources, a phenomenon confirmed by recent reports from both the tech and academic sectors. Major technology giants like Google and Meta have quietly cut staffing for their DEI programs or ceased releasing the detailed diversity reports that once served as industry benchmarks for transparency. In the academic and scientific computing sphere, the erasure is even more literal. Universities and research institutions, bowing to mounting political pressure and the threat of funding freezes, have begun scrubbing their public-facing websites. The diversity statements are being deleted from hiring pages at major institutions like MIT and the University of Utah, and entire directories of LGBTQ+ faculty and support resources are vanishing behind firewalls or 404 error codes, as seen recently at Northwestern University and the University of Chicago. My research highlights that this administrative action has moved beyond simple funding cuts to the explicit censorship of language. We are seeing a sanitisation of vocabulary where terms like "equity," "privilege," and "systemic" are being surgically removed from mission statements to avoid triggering political targeting or losing federal grants. This is a survival strategy for the institutions, as a way to fly under the radar of "anti-woke" legislation, but for the individuals relying on those support structures, it is an act of erasure. It signals that our identity is now a liability too dangerous to even name in public.
Table summary:
The Digital Disappearing Act: A Record of Erasure
This retreat forces us to confront the uncomfortable truth I wrote about in my book, An Unsuitable Job for a Woman. We have built our house on sand. For too long, diversity work in tech has relied on the volunteer time of the very people it is supposed to help. We have treated equity as a form of good citizenship, a vibe we create without actual resources. We have relied on the passion of the marginalised to fix the systems that marginalise them. We have asked women to do the heavy lifting of repairing a culture that was built to exclude them.
I have called this set of observations the "intimacies of labour" (again, see my book). It is the identity work that women must perform just to exist in these professional spaces. It is the mental calculus of deciding whether to be one of the boys or to embrace the label of "Woman in Tech." It is the exhausting effort of bridging the gap between our gender and our professional legitimacy. We are expected to be soft enough to be likeable but hard enough to be competent. We are expected to fix the pipeline while navigating a workplace designed for a man who has no caregiving responsibilities and a wife at home to manage his life. As I see it, the label "Woman in Tech" itself has become a straitjacket. It is a status characteristic that signals difference rather than competence. It implies that our gender is the problem to be solved. It suggests that if we just had more training, or more confidence, or better negotiation skills, the inequality would vanish. This deficit model absolves the industry of its responsibility. It allows tech companies to paste pictures of diverse faces on their websites while their internal cultures remain toxic and exclusionary. Now we are doing this heavy lifting while swimming with sharks. The emotional labor required to sustain our careers is compounded by the fear of political and professional backlash. The hostility from leadership is palpable, with diversity initiatives facing increasing backlash under the guise of preventing reverse discrimination or protecting free speech. My research uncovered reports of senior leaders celebrating the savings from cutting diversity programs and framing equity work as a distraction from excellence. The message from the top is clear: diversity is a waste of resources. AI-bro culture actively resists and prejudices representations of and the inclusion of women. We are, shark bait. We are also F***-ing exhausted. The women I interviewed told me they are considering leaving the field altogether because it is not worth the constant battle. They described the current environment as a full-blown assault on their right to exist. If we lose this generation of women in HPC and technology, we do not just lose diversity numbers. We lose innovation. We lose the future. This is why we must stop worshipping at the altar of metrics. In the data-driven world of computing, we love to count things. We count heads. We count retention rates. We count the percentage of women in the room. But Goodhart’s Law reminds us that when a metric becomes a target, it ceases to be a good metric. We have focused on the appearance of diversity rather than the reality of inclusion. We have allowed organisations to game the system and test the water without ever jumping in. The result is a surface-level diversity that collapses the moment the political weather changes. We have a lack of tracking on retention rates for underrepresented groups because we have been too busy counting who walks in the door to notice who is walking out. We need to stop asking how many women are here and start asking who feels safe enough to speak here. The path forward requires us to name the fear. We must stop pretending this is business as usual. We need to acknowledge the hostility from leadership and the fear of repercussions. We cannot fight a shark we refuse to see. We must reclaim the language of equity and refuse to let the anti-woke agenda define our terms. Diversity is not reverse discrimination. It is essential for robust science and innovation. We must also reject the trap of the volunteer revolution. The isolated volunteer is vulnerable to the shark. The coalition is a fortress. We need to look at the broader landscape of resistance, such as the lawsuits organised by civil rights organisations and the collective actions taken by NGOs and educational bodies. We need to connect our internal struggles with these external movements. We need to build structures that do not rely on the free labor of women to sustain them. We need to professionalise this work and resource it properly. Most importantly, we must refuse to accept the premise that we are the problem. The problem is not women. The problem is a dominant bro-tech culture that protects its own power at the expense of everyone else. The problem is an industry that demands we do the heavy lifting of inclusion while it actively dismantles the supports we built. SO WHAT CAN WE DO ABOUT THIS? I've been reading with avid curiosity and a certain gritting of teeth, the popular science writings of Yuval Noah Harari recently (good to go way outside your comfort zone). As I understand it, Harari's main conceit is Homo sapiens rules the world because we are the only animal that can cooperate flexibly in large numbers. We do this by creating shared stories and experiences. BUT, all too often, these stories are to favour money, nations, and corporations (uh-oh). For the last thirty years, the technology sector has operated on a specific, damaging fiction: the myth that supporting diversity and accessibility is a moral luxury, a charitable add-on to the real machinery of innovation. We have told ourselves that the inclusion of women and minorities is a matter of politeness, rather than a matter of default support. Over the past six-nine months actively excluding groups of people is something AI-bro and tech-bro culture is out in the open actively celebrating. A homogeneous team building a global system is not just unfair; it is computationally incompetent. It creates blind spots that are no longer just social inconveniences but systemic vulnerabilities. If we wish to own the ocean rather than merely survive the swim, we must stop treating equity as a social crusade and start treating it as an engineering specification. We need solutions that do not rely on the benevolence of the powerful or the exhaustion of the marginalised. We need structural hacks that rewrite the code of the institution itself. (I write about this in my book btw, arguing we do not ask more of minority groups to advocate solely by themselves). So here's my shopping list of stuff to action (in response to the question of "what can/should we do to support DEI in the current climate?" asked both at CASC and ExoBioSim events): First, we must reclassify homogeneity as a security risk. In cybersecurity, we do not ask the virus to be nicer; we build firewalls. Similarly, we should stop asking male-dominated teams to "be more inclusive" and start treating extreme gender imbalances as a critical failure in project auditing. Funding bodies and shareholders should view a team of ten men not as a culture fit, but as a high-risk asset prone to groupthink and data bias. We must demand that Red Teaming (as in the practice of rigorously challenging plans and code) be applied to human capital. If a team lacks diverse cognitive inputs, it should be flagged as unstable, and its funding paused until the security flaw is patched. This shifts the burden from the woman raising her hand to groups of people auditing together. Second, we must shatter the illusion of the meritocracy by introducing radical financial accountability. For decades, we have allowed leaders to outsource their conscience to volunteer committees. We must now attach their survival to the survival of their staff. Executive bonuses and grant renewals should be mathematically tethered not to recruitment—which is easy—but to retention, which is hard. So, in this scenario, if the women leave, the money leaves. If the shark drives talent away, the shark starves. This aligns the selfish interest of the leader with the collective health of the group. Finally, we must harness the power of the Strategic Glitch. The current system functions because women and minorities act as the shock absorbers, smoothing out the friction of a toxic culture with their unpaid emotional labor. We organise the events, we mentor the juniors, we soften the blows. It is time to stop. We must allow the friction to be felt. If the "good citizenship" work is not paid, it should not be done. And louder at the back, IF THE GOOD CITIZENSHIP WORK IS NOT PAID, IT SHOULD NOT BE DONE. Let the panel be all male. Let the report go unwritten. Let the vibe of inclusivity collapse so that the raw, jagged reality of the exclusion is visible to everyone. I am advocating for a dysfunctioning system to be allowed to crash. Martyr's Trap I can feel the pushback on my 'finally' point. Is this the Martyr's Trap? So let me join the dots more comprehensively. The current system functions only because women mask the liabilities. We fix the bad PR before it happens. We smooth over the HR disasters. We make the dysfunction look functional. By stopping, we are not quitting; we are simply returning the risk to its owners. Here's my take, in withdrawal from unpaid effort. In doing so, we are handing leaders a liability. We push them into a market that will punish their blindness. An all-male AI development team is not a club; it is a lawsuit waiting to happen. It is a product recall in the making. It is an evolutionary dead end. Case in point, an AI-enabled teddy-bear caught talking sex and knives. We are not fighting a battle for "kindness." We are fighting for the cognitive capacity of our species to navigate the future. The shark is in the water only because we keep feeding it. It is time to change the diet. How to Avoid the Trap: The Art of the "Bureaucratic No" However, the danger to the individual is real. I've spoken to women and minority groups who have lost professional roles, lost work, lost contracts, lost their jobs. If you simply stop doing the work, you risk working in a culture that cuts across your morale code and violates your sense of right and wrong. For many the anxiety here is too much and they self-censure, or disappear altogether. To avoid this, the glitch, in the way I see it, must be engineered with the same precision as the system itself. We do not just stop; we reclassify. How?
A Note on the Architecture of my Argument Hold the line! I must, however, pause to acknowledge the specific architecture of my own position. I write this as a researcher based in the UK, where despite the turbulence of the sector, I possess a degree of contractual security that many of my colleagues in the US tech industry or precarious academic roles do not. It is undeniably easier to advocate for a strategic crash when you are standing on relatively firm ground. Yet, it is crucial to state that the strategies I outline here, the reclassification of labour, the risk assessments, the collective refusal, are not abstract theories born in the safety of a university office. They are the direct, distilled output of the research I have conducted this year. They echo the exact frustrations and desires expressed to me by the computational professionals, the HPC engineers, and the data scientists I interviewed. They told me they felt as though they "were holding up the sky". They told me they "wanted to let go". These recommendations reflect exactly what people I have spoken to told me they needed to do to survive. The Ultimate Reframing I often read articles concerning the "They" in reference to toxic leadership and individuals in positions of power who directly impact tech culture. Let's be specific here. "They" are the beneficiaries of a system designed to extract value from our silence. They are the leaders who view equity as overhead and diversity as a cosmetic feature rather than a structural necessity. They want us to act as the invisible load-bearing walls of an institution they treat as a mere façade, absorbing the stress so they can occupy the penthouse without feeling the tremors. By refusing to do the unpaid work, we are not abandoning the structure. We are unionising the resistance. We are collectively handing back the weight of their own negligence. This is the power of the coalition and the trade union. It transforms a personal refusal into a structural renegotiation. When we stand together to enact this Strategic Glitch, we force the leadership to confront the cost of their own apathy. If they choose to let the infrastructure collapse rather than resource it properly, then let it fall. We are not the help. We are the engineers. And when the dust settles, it will be our collective blueprint that determines what rises next. Comments are closed.
|