Former X Product Head Nikita Bier: Most Large Companies Survive Purely by Continuously Launching Products and Using Elimination to Discover What Works
Nikita Bier posted that most large companies survive purely by continuously launching products and using elimination to discover what works. He wrote: There is no rhythm, reason, or belief in anything they do.
The post did not name specific products or financial figures, but instead described the growth of large companies as a trial-and-error inventory: first laying out a row of launches, then seeing which ones survive. He himself sold TBH to Meta, then became the head of product at X, observing how large companies use traffic and experimental groups to consume smaller products.
This aligns with his recent commentary on Muse's decision-making. Muse introduced automated pricing into Facebook Marketplace, knowing that robots would hurt sellers, yet it was framed as being able to "only poison a little." Within the elimination framework, poisoning is not a strategic mistake, but just another launch, waiting for data to decide whether to retract.
The advantage of large companies, he noted, is their capacity to endure ineffective launches: budgets, distribution, and existing users allow most actions to be taken without belief. Small teams must have a reason before starting; large companies can start first and then provide reasons for the one that remains.
The incremental information turns "experimental culture" into a survival algorithm. A/B testing and gray scaling are no longer science, but a way to exchange scale for not having to think things through first. Lack of belief is not a management failure, but a default state allowed by scale.
In market mechanisms, this is organizational narrative. The buy side bets on platform distribution, advertising inventory, and the ability to shut down failed products and launch the next one; the sell side seeks independent companies with clear product beliefs. The flow of funds favors giants with room for trial and error. The beneficiaries are platforms that control distribution; the pressured are startup teams that must hit the mark every time.
Source: Public Information
ABAB AI Insight
Bier has sold explosive social applications and observed experimental machinery within Facebook and X. His conclusion is not that large companies do not innovate, but that innovation has been rewritten as launch density. TBH was acquired in nine weeks, which itself is a large company buying something already validated with cash while continuing to eliminate the ten unvalidated ones. Belief has been outsourced to retention curves.
The capital path is to exchange distribution for time. Large companies have advertising, app store placements, and existing social relationships, allowing them to present features without reasons to millions of people first. Dead products are recorded as learning costs, while those that survive are recognized as strategies. Muse entering Marketplace follows the same logic: first get the agents running, then see if sellers withdraw. The motivation is to avoid internal debates over right and wrong, leaving the debate to the elimination after launch.
In a similar position, this is closer to the failure rate model of pharmaceutical pipelines rather than a single product gamble like Steve Jobs. The industry phase is consumer technology replacing "vision" with "launch rights." Those who can simultaneously endure ten blank shots do not need to have belief first.
Structural changes belong to capital concentration. As trial-and-error rights concentrate at nodes with distribution and cash, the market superficially has many products, but decision-making has little belief. The mechanism is: scale dilutes the cost of single failures, thus organizations stop training judgment; once judgment exits, only a tighter round of launches remains to replace strategy.
ABAB News · Cognitive Laws
Scale first buys the right to operate without belief
Elimination appears scientific, but is actually inventory management
Launching first and finding reasons later is the default operating system of giants