An excerpt from the book Behind the Startup details how an unnamed Silicon Valley-based unicorn's push for rapid growth left little time for actual engineering
For 19 months, the sociologist Benjamin Shestakofsky embedded himself in an early-stage tech startup to study its organization and culture. X: @ieeespectrum , @p_h_albuquerque , and @grady_booch LinkedIn: Christos Makridis and Noah Gift X: @ieeespectrum : Sociologist @bshestakofsky spent 19 months embedded in a Silicon Valley startup that went on to be valued at over $1 billion. Here's what he saw. The company's name has been changed to protect privacy, but you're welcome to guess! #startups #venturecapital https://spectrum.ieee.org/... @p_h_albuquerque : “The interdependence between generously compensated software engineers in San Francisco and low-cost contractors in the Philippines suggests that advances in software automation still rely on human labor and on global inequalities.” #unicorns @IEEESpectrum https://spectrum.ieee.org/... Grady Booch / @grady_booch : In the early days of computing, computers were human. It would seem that some would have us return to that era. “AllDone's users never knew that human workers, rather than a computer algorithm, had handcrafted each introduction.” https://spectrum.ieee.org/... LinkedIn: Christos Makridis : The Silicon Valley unicorns can sound good at first glance, but we need to be careful not to idolize the concept of “scale”; some of the most seemingly persuasive companies have major flaws. … Noah Gift : This is true! Been in so many “bootleg” startups now that had shoddy engineering it is hard to count. Move fast and destroy things, works until it doesn't. …
Context & Ripple Effects
Shestakofsky’s embedded account adds organizational evidence to an older critique of VC-fueled blitzscaling, in which growth imperatives can outrun the operating foundations needed to sustain them.
It also extends the disillusionment documented in earlier accounts of startup culture: the story is not only about employee experience, but about how a polished product can depend on concealed human work across borders.
First-order effects
- The studied company’s growth model left engineering with insufficient time to build and maintain the product, making operational scale depend more heavily on existing processes and people.
- Users of AllDone were presented with a service whose introductions were hand-crafted by workers rather than generated by an algorithm, obscuring the labor behind the product experience.
Second-order effects
- A human-mediated service can add contractor-management costs and quality-control pressure as demand rises, rather than gaining the cost advantages customers may associate with software automation.
- Competitors and investors evaluating similar growth claims have stronger reason to distinguish demonstrated automation from services propped up by low-cost global labor.
Third-order effects
- If growth metrics continue to outrank engineering investment, more venture-backed software businesses may accumulate operational and technical fragility before those weaknesses become visible to customers or markets.
- The case points to a broader accountability question for digital services: whether products marketed as scalable software must become more transparent about the human labor that delivers their core output.
The trend: This is one data point in the reassessment of growth-at-all-costs startup models, especially where apparent software scale is underwritten by hidden labor and deferred engineering.