Programmers may disappear, but everyone has become a “programmer.”
Author | Wildcard
All day today, people on X have been discussing, “Happy 1024 Programmer’s Day—will this be the last Programmer’s Day ever?”
This day happened to fall on a Saturday. On Weibo, the topic was titled “This Year’s Programmers’ Day Comes with a Day Off,” accompanied by a caption stating that AI tools have taken over repetitive coding tasks, allowing programmers to finally enjoy a holiday without being interrupted by urgent requests.
For the fifth year in a row, Insta360 has been awarding golden keycaps to outstanding programmers. This year, they’ve added an extra “F5” key, symbolizing continuous updates and iterations. At the awards ceremony, a young colleague joked, “This year’s golden keycaps should go to AI.”
One organization’s holiday tweet was more straightforward: “This Programmer’s Day, we’re hiring more—anyone who can instruct AI to build things counts as a programmer, too.”
Meanwhile, on Xiaohongshu, a programmer posted under the “Ancient-Method Programming Era” tag to clear out old books: “Illustrated HTTP,” three copies for one yuan each, and algorithm books for ten yuan apiece. Another post was titled “Programmers’ Books—Even Duoguyu Won’t Take Them,” and among the few that eventually sold, the most expensive turned out to be “The Mythical Man-Month.”
A festival has been given a holiday, expanded its recruitment, and even its “bible” is on sale at a discount in the secondhand market.
Festivals don’t usually die because they’re canceled. They die when no one needs them anymore to define who they are. To understand this, we first have to go back to how this festival came into being.
From coders’ inside jokes to a holiday for everyone
1024 has never been an official holiday. Its distant cousin is Russia’s “Day of the Programmer,” established in 2009 as the 256th day of the year; when it spread to China, the date became erratic due to leap years, so it was replaced with the more memorable power of 2 raised to the 10th. Starting around 2015, October 24 has become the de facto “Programmer’s Day.”
This is a secret code known only to those in the industry. Growing up alongside it came an entire set of self‑deprecating jokes. Around 2006, “IT migrant workers” began circulating online; from 2009 to 2012, the term “coder” spread rapidly, followed by “code slave” and “code livestock.” The premise of this self‑mockery was a threshold: only with a salary significantly higher than one’s peers and work that outsiders couldn’t understand did people feel confident enough to call themselves migrant workers. In 2021, the Ministry of Human Resources and Social Security included software professionals in its statistical category of “new‑generation migrant workers,” prompting programmers to flood social media with posts like “Coders really have become farmers,” their tone more lighthearted than angry.
Then, this secret code was invited out of the circle. In 2020, CSDN held the first 1024 Programmers’ Day in Changsha, with an academician delivering the opening address; in 2023, it appeared on CCTV, and by 2024, the Weibo topic had amassed 4.07 billion views.
Interestingly, the festival’s slogan has been evolving over time. In 2023, the Changsha Conference shifted its focus from open source to artificial intelligence for the first time. By 2024, the Programmer’s Day event hosted by the China Software Industry Association had adopted the theme “AI surges toward the future—our shared destiny is inseparable from AI”; meanwhile, at the opening ceremony of the Changsha Conference, CSDN founder Jiang Tao demonstrated live how to use natural language to have AI write code.
This year, the conference was renamed “1024 Programmers’ Day and AI Builders’ Conference,” with the theme “Everyone Can Build.” At the opening ceremony, founders of several robotics companies were in attendance. The term “builders” is far broader than “programmers,” and among the new programmer projects listed in that “recruitment‑boosting” tweet were a consumer‑cooling‑off‑period reminder, a wardrobe‑styling generator, and a chatbot capable of engaging Li Bai in conversation—projects created by people who hadn’t written a single line of code by hand. AI programming tools also treated this day as a user‑acquisition milestone: ZCode gave away 500 million tokens, TRAE offered check‑in raffles, and the official announcement revealed that 92.5% of participants received a 10‑yuan coupon.
The inclusion of more participants has expanded the festival’s community, and AI has made that community virtually borderless. A festival open to everyone, in essence, no longer belongs to anyone in particular. But as the community grows, what are those who once stood at its center experiencing?
The unluckiest cohort of computer science students
Computer science students who enrolled in 2021 and 2022 jokingly refer to themselves as “the unluckiest class in history” upon graduation this year. When they first entered school, they were surrounded by promises of fruit days and offers that seemed within reach after just a few rounds of interviews—yet by the time they graduated, almost no one was willing to hire entry-level candidates.
The data is on their side. A Stanford study shows that the number of software developers aged 22 to 25 in the U.S. has declined by nearly 20% from its peak at the end of 2022, while senior developers have largely held steady. In China, according to MyCOS’s Employment Blue Book, computer science majors have, for the first time, dropped off both the list of “green‑card” undergraduate majors and the top ten highest‑paying fields; meanwhile, data from Maimai indicate that demand for backend development positions has fallen by 41%, while demand for AI‑related roles has surged by 7.9 times.
What was pulled out wasn’t the entire ladder, but rather the bottom rung.
The most compelling example comes from the company that has taken AI-driven programming to the extreme. Mid-month, Anthropic released a report detailing changes in its workforce. From early 2025 to the present, the share of junior-level employees in the engineering division has dropped from 21% to 14%; among newly hired engineers, the proportion of entry-level positions has fallen from 27% to 9%, representing a two-thirds reduction.
Of course, there are differing opinions as well. Amazon CTO Werner Vogels said in an interview in Shenzhen a few days ago that the number of junior engineers Amazon is hiring “remains the same as in the past.” New hires need to learn how to collaborate with others, understand a system that has been running for twenty years, and determine which AI-generated code requires careful review.
He also introduced a concept called “validation debt.” AI generates code, but no one has ever verified whether it’s correct; this portion remains unpaid, much like technical debt, and the team may not even know what issues it harbors.
But even in his narrative, the value of newcomers has already shifted away from “writing code” and toward validation, collaboration, and judgment.
For those who stayed, the work itself has also changed.
In September this year, DHH, the creator of Ruby on Rails, announced at the Rails World conference in Austin that his company, 37signals, would no longer rely on hand-written code as the standard approach to software development. He dubbed this move “Pencils down,” meaning putting down the pencil. The same person who once said he’d rather retire than hand over his keyboard to AI asked the audience onstage how many people still write large amounts of code by hand.
Only five people raised their hands.
Domestic programmers are undergoing the same transformation. An engineer working on the backend at a major tech company says that in the past, he spent 60 to 70 percent of his day writing code, with the rest devoted to clarifying requirements, designing systems, and debugging issues. Since AI was integrated into his workflow, that ratio has completely reversed. A PhD student recalls that two years ago, when he opened an AI‑powered coding window, he had to avoid his labmates’ gaze, because “using AI made it seem like I lacked competence—like I was slacking off.” Now, he actively recommends it to his senior colleagues.
So, this year’s 1024 falls on a Saturday, meaning programmers can finally take a day off—good news, right?
Anthropic’s report also includes a set of figures. In March this year, the total working hours of Claude agents within the company surpassed those of human employees for the first time. By August, the working hours of agent tasks overseen by employees were already six times those of humans themselves.
The engineers’ daily routine described in the report involves simultaneously running multiple programming agents on their terminals, switching back and forth among a dozen or so contexts. Backend data shows that interactive AI agents spend roughly 80% of their time idle—not because the models are slow, but because they’re waiting for humans to provide the next instruction.
Writing code is no longer the bottleneck; human attention is.
Weibo claims that AI has saved programmers their day off. That’s not entirely wrong, but it could be rephrased. This year’s 1024 didn’t require overtime because the tasks that would have required overtime have already been taken over by AI.
Programmers are shifting from “code writers” to “code reviewers,” “AI task assigners,” and “signers off on outcomes.” These roles are all crucial—some even more challenging than writing code itself. Yet they share one thing in common: none of them really requires the “1024” password.
A profession’s name becoming disconnected from the work it actually entails has happened throughout history. And the last time this occurred, it happened to be related to “computing” as well.
Before the advent of electronic computers, “computer” was a profession.
This term refers to a person—specifically, someone who performs computations. In the late 19th century, the Harvard Observatory hired a group of women to analyze photographic star charts one by one; they were known as the “Harvard Computers.”
Around 1891, female human computers at the Harvard College Observatory were analyzing photographic star charts. Image source: Harvard College Observatory
During World War II, the National Advisory Committee for Aeronautics at its Langley Laboratory in Virginia employed hundreds of female mathematicians to manually compute wind-tunnel data and flight trajectories. Among them was a segregated division known as the “West Computing Group,” which was composed entirely of African American women.
In the summer of 1949, in the “Computers’ Room” at the NACA High-Speed Flight Station, operators manually processed flight test data. Image source: NACA / NASA
In 1945, the world’s first general-purpose electronic computer, ENIAC, was completed at the University of Pennsylvania. The six women who got this machine up and running were selected from among the human calculators who had previously specialized in computing ballistic tables. They became the first “programmers” in history.
In February 1946, ENIAC programmers Betty Jennings (left) and Frances Bilas prepare for a public demonstration | Image source: U.S. Army
Dorothy Vaughan, the head of the West Area Computing Unit, did something when IBM’s electronic computers were brought into Langley in the late 1950s. She taught herself the FORTRAN language and then trained her fellow human computers to do the same. With the machines taking over the calculations, they transitioned into programmers who wrote code for the machines. In 1958, the West Area Computing Unit was disbanded.
From left to right: Dorothy Vaughan, Lacy Hunt, and Vivian Adair—each of whom was once a human computer at Langley | Image source: NASA
However, human calculators did not disappear overnight. In 1962, astronaut John Glenn was preparing to make America’s first manned orbital flight aboard the Friendship 7. The orbit had already been calculated by an IBM computer, but Glenn wasn’t convinced; he specifically requested that Katherine Johnson, a mathematician from West Virginia, recheck the numbers by hand. According to the most widely circulated account, his exact words were: “If she says the numbers are right, I’ll go.”
On February 20, 1962, John Glenn entered the Friendship 7 spacecraft. Image source: NASA, public domain.
Machines do the calculations, while humans sign off on the results. This scene from more than sixty years ago is almost identical to today’s practice of engineers individually reviewing the outputs of intelligent agents.
In 1966, Katherine Johnson working at NASA. Image source: NASA, public domain.
What happened afterward is well known: the word “computer” stuck, but what it referred to changed completely—from a group of women sitting at desks with slide rules to a machine. The profession wasn’t declared extinct; it simply had its name swallowed by its own tool.
Today, no one celebrates “Computer Programmer’s Day.” Not because computation is no longer important, but because computing has become ubiquitous—so integral that it no longer requires a dedicated group of people to represent it.
The profession of programmer itself emerged at a time when the previous occupation was being diluted.
Now it’s its own turn.
You can spot the trend in the changing names. In 2024, a company called Cognition launched Devin, touting it as “the first AI software engineer.” Since then, nearly every major tech firm has rolled out its own “AI programmer” or “programming agent.” At this year’s 1024 event, ByteDance’s TRAE community titled its showcase “A Tribute to the True AI Engineer.”
The term “programmer” is, much like “computer” in the past, gradually shifting from referring to a person to referring to a tool.
Meanwhile, on the human side, the job titles are also evolving: “Builder,” “AI Engineer,” “Agent Manager,” and “FDE.” According to data from Maimai, hiring for the role of “Field Deployment Engineer” has increased 15-fold year over year. Not a single new title includes the character “code” (码).
So, back to that opening statement: Will this year’s 1024 be the last Programmer’s Day?
Most likely not. On October 24 next year, Changsha Smart Valley will once again light up, major companies will still release keyboards, and AI programming tools will continue to distribute hundreds of millions of tokens on that day. This festival will keep thriving—and it may be more vibrant than ever before.
It’s just that the star of the show has changed. It will become a festival for all “builders,” a marketing milestone for AI tools, and an annual calling card for the local industry. That small community, once brought together by self-deprecating humor—the group of people who understood why “1024” was funny and tacitly shared that inside joke—will gradually realize that this festival no longer needs them to affirm their identity.
Just like Dorothy Vaughan’s generation, who ultimately didn’t retain the title “computer” but became among the earliest programmers.
What will people who today stare at six AI terminals on their screens at the same time be called in the future? That remains unanswered. What is certain is that 1024 is still 2 to the power of 10—this number will not change; what will change is who still needs it to prove themselves.
*Featured image source: PCMag
This article is an original piece by GeekPark. For reprints, please contact Geekjun via WeChat at geekparkGO.