In a note, Shemol breezily asked whether there was ever a time when people are paid to create Excel and PowerPoint, reached a positive conclusion with Claude, and compared supposed decline of these employees to the current AI hype threatening developers. This is a good question catering my bouquiniste soul. I left a letter in response, which inspired the post you are reading.
I did not experience the time of spreadsheet-makers. I was, at that time, a fake geek boy lingering about office suites and online services. I habitually help Muqliongists and teachers with paperwork, with no idea of work or even college entrance. I swaggered through the provincial capital, staring at my first smartphone, replacing lorem ipsum in Calibri with ‘observation’ notes I was told to take, and feeling the ‘generousity’ and lightning-fast device switch of cloud services. I almost forgot the carefree age, when I also picked up You Gotta Play with Excel Like This, a book published in 2011 by Quu Hao.
Those not playing like this
Few was Quu Hao remembered outside the book, despite the brazen boasting in the preface.
n is an aged slang meaning an indetermined large number.
I only ranked 10ⁿ-or-so-th in the Excel arena, yet I easily manage n1 types of reports from various industries. I use spreadsheets for department management, standardized operating procedures, production control, and risk anticipation. Classic templates made by my hand still serve the companies and clients I once worked for. I complete the work that would take others a week in a minute, or summon n reports in much the same time.
The oblivion contrasts to later productivity missionaries, or Quu simply disappeared like thousands of StackExchange counterparts where, in Quu’s words, people constantly ask ‘how to do x in Excel.’ Fortunately, Quu rescuedly recorded the life of spreadsheet-makers. One can probably believe from Quu’s vivid narration that such a group of people really existed.
There are people called table lasses and table lads in modern society, coping with endless spreadsheets and statistics. They toil all day long for a discounted outcome. Among them are no shortage of technical experts, function guru, and quick mousers, yet they still struggle with creating various reports.
In fact, spreadsheets do not only trouble the table gang. 80% of working professionals with computers have, more or less, ever or are competing with Excel. However, it is our dear bosses who suffer the most, waiting long days for a single analysis and missing out on valuable opportunities. Worse still, they make poor business decisions based on imprecise data, and in the blink of an eye, hundreds of thousands or even millions vanish into thin air.
Quu started exploring Excel in a catapulted internet company. Seven years later, Lii Dhyih, the author of Speak not You Know PPT, encouraged Quu to write a book as frivolous as herself. I only quote the advertisement from the Excel book:
How did Fveuyong in adjacant office find time replying to Rénrén moments? Why did Yucfuq always have time scrolling down news on Rénrén? Why was I struggling with PowerPoint all the time? Wonder the answer? This Rénrén-accredited must read for white-collar workers unveils to you the secret of PPT!
Tritabular doctrine
After promising such a fundamental change, Quu’s lesson boiled down to merely one line: make a dense, bill-like source data sheet. Think required print-ready reports an output of data sheet, not the place to fill in original data. Merged cell is a visual camouflage of null values. Columns (fields) should be atomic. Each row should be a record, with no spacing nor subtotal. Reports are not made, but summoned, a word you have seen Quu using before. The only tricks are obscure shortcuts, mouse actions, and using delta column in source data.
Quu explained the three sheets until Excel 2010 by default as a theologist would — a parameter sheet, a source sheet, and a presentation sheet — without forgetting to address this is symbolism and the numbers of presentation sheets are limitless given proper source data.
The principles for source data are quite intuisive (readers only did not notice that there should be a data sheet with independent records), leaving the space of a whole book for chattering. My previous paragraph puffed up unto a chapter with forty-four pages, and another chapter digs down to mouse clicking in pivot tables fields, where to ‘summon’ reports even during an interview with the boss. Based off my current knowledge, I am astonished the hypothesized office worker was authorized to access integrated data but all they can use was Excel!
Does the analogy hold?
The term deserves a pair of scare quote.
I dare not say the cursory descriptions are enough to retrodict the era that was scarcely different to the present day, but has began to drift apart. Common sense suffices to know what happened to the table gang, which justify the programming–spreadsheet analogy in the first reading. Design patterns2 for functions, vibe coding for SaaS-crystalized procedures, yet there is one thing different: the remainder.
The table gang was replaced, if we can say, by ERP/CRM, realtime databases, and prediction algorithms, transforming a week of one Excel guru to a click of Export button. They have even coëxisted with the table gang, only hindered by data silos. However, what we are witnessing is the decline of those giant windmills, not the rise of one another, exactly to which have the vast majority of golden-age programmers dedicated their youth. Longing for independent developer is tantamount to the table gang’s revenge. While reports are only moved from manual to automatic, we did not received any yes-answer, nor have we teached LLMs more than stale golden-age memes.
My point is, Salesforce indeed cannot be vibe coded in a day, but it is sold to large centralized bunisesses. And yes, coding counterparts to Hollywood do exist; if only I could see them in an attic with specific offerings.
LLMs are hardly a knowing deus ex machina, or a supplement to the world’s deterministic traditions. If LLMs genuinely supports nomadic businesses, more than fabricated The Economist headlines, we will drift from Salesforce and Jira (hence its creation3) to what photography or performance have experienced. Few will be employed by large organizations and mummify the 2010s workflow. Most of us will make ourselves influencers, jumbling our work with seemingly ubiquitous AIr, to serve the next ambitious careerist.
From fake geek to pessimist
My young self lingered around every experimental feature in Facebook and QQ. I did not notice that someone was conducting thousands of A/B tests from years ago until now, hoping for the retention rate to bump up. The only difference is, the experimental feature was then hits like Facebook timeline, and now tedious disguised advertisement.
Spreadsheet creation, programming, and even tricks that now sound like child’s play, may all be the same thing for my younger self. Perhaps what I loved was never computer science, but rather a certain sense of discovery through a winding path, a path ultimately artificial. I fear (or perhaps it is) that what I enjoyed was merely puzzles or fun, a superficial veneer of professionalism. Spending days coding in my own little world is inherently different from being a cog in the machine in the real world. I felt my already illusory passion dissipate, and it took me a long time to understand what this shift truly meant. I told myself that this rupture did not stem from AI hype, nor some cowards; the rupture was reality.
Vanish did the spreadsheet-makers into the air, leaving only Quu’s out-of-print book. Or they did not, counting startups cramming all the procedures in their holy trinity. Sprinting figures will be someday like the estranged, almost incomprehensible past when reports and slides were typed by hand. But there is no page in history that have solely destruction.