Featured image of post Thoughts|Do Machine Learning Engineers Dream of Sexing Chicks?

Thoughts|Do Machine Learning Engineers Dream of Sexing Chicks?

The tragedy of Japanese Americans may not be far from today’s Chinese tech elites in Silicon Valley.

This article’s first section draws extensively on the research and data presented in Chick Sexing Is a Job Full of Psychological Suggestion

A Glimpse of History Through a Chick’s Cloaca

In the late 1930s, American farmers faced a long-standing dilemma: how to determine the sex of chicks immediately after hatching? In the poultry industry, male chicks hold little economic value—they don’t lay eggs, and their meat is inferior—so early identification and culling are essential to reduce feed costs. However, male and female chicks are virtually indistinguishable to the untrained eye. This gave rise to a niche but highly specialized skill: visually identifying chick sex by inspecting their cloaca.

The technique is known as “chick sexing.” At the time, experts capable of performing it with speed and accuracy were extremely rare. In 1925, Dr. Kiyoshi Masui of the University of Tokyo discovered a reliable method to determine the sex of common chick breeds within 24 hours of hatching. It involved inspecting the chick’s cloaca: male chicks typically exhibit a small bump on the lower edge—a vestigial sexual organ—that females lack. By the 1930s, Japanese poultry workers had developed Masui’s method into a practical, rapid technique and founded the All-Japan Chick Sexing School to train new experts. Apprentices underwent three years of theoretical and hands-on training and had to prove themselves by sexing thousands of chicks daily. With required accuracy above 98%, it was a high-pressure profession. Some sexers could identify up to 10,000 chicks in an eight-hour shift, and a single worker could process over 300,000 chicks in a typical 100-day hatching season.

As industrial poultry farming scaled up in the U.S., American hatcheries grew desperate for this expertise. Yet few outside Japan could match the rigorous standards of graduates from Japan’s chick sexing schools. Beginning in the 1930s, many trained Japanese sexers were recruited by American companies and traveled across the Pacific. On arrival, their position was unique: respected by farmhands, relied upon by technicians, and well-compensated by employers. They weren’t scientists, yet their skill was indispensable; not white-collar, yet they carried a dignified command of technique. Still, they lived under the shadow of racial prejudice and discriminatory laws—such as the Alien Land Laws, which barred non-citizen foreigners from owning or leasing land in states like California, Oregon, and Washington.

Then came Pearl Harbor. On December 7, 1941, Japan attacked the U.S., and war was declared. The geopolitical rupture had swift and brutal consequences for Japanese Americans. Many had lived in the U.S. for years, some were citizens, but now all were suspected as potential enemies. On February 19, 1942, President Franklin D. Roosevelt signed Executive Order 9066, authorizing the forced relocation and internment of all Japanese Americans from the West Coast. Over 120,000 were incarcerated in remote camps; thousands more fled east. Officially, it was a matter of “national security,” but in practice, it meant the stripping away of freedom, property, and dignity.

Ironically, while the U.S. government confined their bodies, it couldn’t afford to discard their skills. In 1943, the War Relocation Authority began distributing “loyalty questionnaires” and permitted those deemed loyal to work outside camps. Some chick sexers, despite being interned, were specially cleared by the Department of Agriculture to resume work—processing tens of thousands of chicks per day for the war effort. They became “imprisoned experts,” operating quietly yet indispensably in the machinery of American agriculture.

Only in 1988 did the U.S. Congress pass the Civil Liberties Act, signed by President Reagan, which formally apologized for the internment policy and offered surviving victims $20,000 in compensation. But the number of people covered was limited, and no financial sum could undo the psychic and social wounds inflicted.

Silicon Valley’s Technocrats and the Indifference of the State

Eighty years later, the cloacal sexing method is still used in a few hatcheries, but has largely been supplanted or supplemented by alternative techniques. Today, early sex determination in poultry often involves breed-specific traits—such as feather color or growth rate—and non-invasive methods are increasingly favored. In some pilot farms, AI vision models are now applied to chicks just days after birth, using high-resolution images, behavioral patterns, or even vocal signatures for prediction. These models are powered by convolutional neural networks, data augmentation, transfer learning, and edge deployment—all built upon the expertise of AI engineers, machine learning scientists, and data specialists.

Beyond agriculture, in the broader AI industry—especially in Silicon Valley and elite research institutions—Asian, and particularly Chinese, professionals form the backbone of the field. According to the National Science Foundation (NSF) 2023 report, 31% of PhDs in computer science and mathematics in the U.S. were earned by students from mainland China. Company reports further reveal:

  • Asians comprise 47% of Google’s engineering workforce, with Chinese professionals forming the largest subgroup.

  • At Meta, NVIDIA, and Amazon, Chinese engineers and researchers are deeply involved in AI architecture, chip optimization, and cluster scheduling.

  • In top-tier AI conferences like NeurIPS and ICML in 2023, over 25% of first-author papers were written by researchers with Chinese backgrounds (including international students).

Today’s “chick sexers” are those who fine-tune models, label data, write papers, and burn the midnight oil deploying inference engines. They are the quiet workforce sustaining America’s AI dominance—vital yet unseen, from OpenAI to Anthropic to Tesla.

Yet, as in the 1940s, rarity of skill does not protect against political hostility. In recent years, the U.S. government’s suspicion toward Chinese talent has escalated into institutional discrimination and a veritable witch hunt.

In 2018, the U.S. Department of Justice launched the “China Initiative” to combat espionage and intellectual property theft. However, the initiative quickly drew fire for racially profiling scientists. According to MIT Technology Review, over 90% of researchers prosecuted under this program by 2020 were of Chinese descent.

One of the most high-profile cases was that of MIT mechanical engineering professor Gang Chen. A U.S. citizen and top expert in thermal science, he had received grants from the Department of Defense and Department of Energy. Yet in 2021, he was arrested for allegedly failing to disclose a Chinese academic collaboration worth under $20,000. The case was later dismissed; no evidence of espionage was found.

Another case involved University of Tennessee professor Anming Hu, who was accused of links to Chinese institutions. The FBI surveilled him for two years with no solid evidence. When the case went to court, the jury refused to convict, and the judge declared him innocent. Still, Hu lost his job and suffered profound reputational harm.

These stories illustrate a chilling logic: when your skills are useful, you are a “model minority” in the lab or deployment line; when geopolitics intervene, you become a suspect based on heritage alone.

Today’s Chinese community—especially H-1B visa holders, STEM PhDs, and early-career faculty—are following a disturbingly familiar trajectory:

  • Hired for their rare expertise, diligence, and output;

  • Suspected for their passports, skin color, and national origin;

  • Targeted, in part, due to the silence and caution that pervades the Chinese diaspora.

This isn’t alarmism—it is the echo of history. As Italian historian Benedetto Croce wrote, “All history is contemporary history.”

Just like the wartime chick sexers, today’s engineers and professors continue to generate immense value while trapped in an invisible cage of distrust. You might spin up a U.S.-based inference model on a cloud server—but your government doesn’t trust you not to upload your code to a former advisor in China.

So whether you are an H-1B engineer, a green card applicant in backlog, or a Chinese American voter, it is time for a collective reckoning: technology won’t shield us from prejudice, and silence won’t guarantee safety.

We must stand up—against racial profiling, political scapegoating, and the misuse of “national security” to justify discrimination.

Technology can be shared—but dignity must be defended. The Chinese diaspora must not quietly walk into history’s shadow again.

Answer to the Ultimate Question of Life, the Universe, and Everything
Built with Hugo
Theme Stack designed by Jimmy