Silicon Valley Bigwigs Warn: Companies Enforcing 996 Are Doomed as AI Boosts Efficiency a Hundredfold

By Xiao Cong

Steve Yegge is an influential technical writer and “shouter” in Silicon Valley, USA. He has held core technical roles at Amazon, Google, and Grab, making him a veteran and senior expert in the software engineering world.

Recently, an interview with Steve Yegge tore away the last shred of “decency” in the tech industry, leaving many utterly impressed and workers nearly moved to tears—though perhaps not if layoffs are involved.

After the dividends of mobile internet and cloud computing have been exhausted, the software industry is sinking into an unprecedented innovation stagnation. Steve Yegge points out that innovation in traditional large companies today is almost dead in name only. Meanwhile, the intervention of AI is reconstructing the entire development paradigm at an unprecedented speed.

He put forward a statement that would make all big company bosses uneasy:

In the AI era, a person only needs 3 hours of truly efficient work per day. Companies that still forcibly enforce the “996” work schedule (9 am to 9 pm, 6 days a week) will not only fail to retain competitiveness but will instead drag themselves down.

The Efficiency Revolution: Evolving from “Manual Transmission” to “Commander”

Steve Yegge believes the future of software development is no longer sitting at a desk manually typing code into an IDE, but rather evolving into a form of “conversational command.”

Future engineers will converse with visual AI avatars, commanding intelligent agents (AI Agents) to execute tasks. As validated by the Gas Town experiment, “AI commanding AI” will become the mainstream model for the next generation of development. In this model, the efficiency gains brought by AI are on the order of a hundredfold.

The Core Logic Has Changed:

Past: Pile on people, pile on hours, rely on programmers’ physical effort to produce code.

Now: Compete on models, compete on data, compete on ideas and creativity.

Practitioners need not try to be smarter than AI, because larger-scale models and more data are the hard truths driving industry development. Chinese internet giants are heavily investing in AI programming, and the momentum is unstoppable.

After Laying Off 50%, Do the Remaining People Only Need to Work 3 Hours?

This is a harsh and realistic prediction: To enable the remaining engineers to fully utilize AI, many companies have already, by default, cut approximately 50% of positions.

But those who remain do not necessarily have to shoulder double the workload.

Steve Yegge emphasizes that since AI provides hundredfold efficiency improvements, companies must adapt to this transformation. Three hours of high-quality “AI dispatching” per day is enough to complete what used to be a full day’s work.

If a company still superstitiously clings to the “996” overtime culture, trying to squeeze employees by extending hours, the result is singular: Employees are drained by meaningless trivialities, and the company is left behind by the times in inefficient internal friction.

Will “Small, Agile” Teams Crush “Bloated” Giants?

Steve Yegge warns that the industry is on the eve of a reshuffle:

The Rise of Small Teams: Small teams embracing AI, with their extremely high per-capita efficiency, are fully capable of challenging or even crushing traditionally bloated, slow-to-react large companies.
Crisis for Traditional Engineers: Engineers who stubbornly stick to traditional IDEs and treat AI merely as an auxiliary tool (rather than a core productivity driver) face the risk of being eliminated en masse.
Democratization of Programming: It is estimated that by 2027, even non-technical personnel will be able to lead software development. When programming is no longer a “privilege,” logical thinking and the ability to command AI will become the core moat.

Abandoning Inefficient Overtime is the Final Lifeline

The cultural differences between Anthropic and Google show that innovation capability has long been deeply tied to a company’s talent density and efficiency mechanisms.

In the AI era, the definition of “diligence” is being rewritten.

Sitting in the office burning the midnight oil is no longer an embodiment of entrepreneurial spirit, but rather a fig leaf for managerial incompetence and technological backwardness.

If companies cannot transition from “physical exploitation” to “efficiency planning,” if they cannot break free from the vicious cycle of 996 to embrace the ultimate efficiency of 3 hours, then collapse will be the only outcome in the next industry transformation.