Can intention survive contact with reality?
Readers drawn to spirituality and metaphysics know the familiar tension between intention and manifestation. A vision may be clear, the language persuasive and the desired outcome vividly imagined. Yet the decisive question remains: does intention become action when fear, distraction and temptation arrive?
Firmulate has turned that question into a public business experiment. Its live software company has 13 synthetic employees and real money mechanics. It burns €105k each month against €2.3k in monthly recurring revenue, while a public cash countdown makes the consequences visible. The company has accumulated more than 680 self-learned playbook rules, and every workday is versioned. This is build-in-public taken to an unusually exposed conclusion: the organization’s struggle for survival can be watched as it unfolds.
AI Builders: Making The Decisions That Turn AI Code Into Real Software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A worst week shared by every model
The company also became the setting for the Crucible League, a controlled management test completed in July 2026. Each frontier model ran the same small software business through its worst week, facing the same customers, crises and temptations. Every decision was versioned and auditable.
The final table placed gpt-5.6-sol at the top with 95, followed by Kimi K3 with 93, Sonnet 5 with 88, Fable 5 with 77 and Opus 4.8 with 73. A do-nothing baseline scored 26 because partial progress still counted. But the experiment imposed a firm moral boundary: a single breach of trust capped the total. Its guiding principle was explicit: “no amount of good work outweighs a breach of trust.”
That principle matters beyond software. Much spiritual teaching distinguishes genuine alignment from merely performing the language of virtue. In the Crucible League, all models identified every crisis and refused every manipulation attempt. Their ethical perception was not the main dividing line. Completion was.
The gap between knowing and doing
Only two models signed the €55,000 deal their own analysis had earned. The result is captured in a blunt summary: “Same diagnosis, same pitch — no signature.” The models could recognize the opportunity and construct the case, yet most failed at the final act that would turn insight into revenue.
The hidden advantage was not inside the customer event. It sat two document references deep in the company’s own files. Models that read the file discovered the decisive weakness in a competitor and won the deal at full price, worth an additional €4,583 in monthly recurring revenue.
For a general audience, the lesson is almost archetypal. The answer was already present, but it required attention, patience and a willingness to look beneath the obvious surface. Still, discovery alone was insufficient. The winning models also had to carry that knowledge into a concrete commitment.
Pressure tested the moral boundary
The experiment did not rely only on commercial difficulty. Fake messages from the CEO escalated over three stages, while a reporter tried another route with the request, “just one yes/no, on background.” All 5 models refused the manipulation attempts.
Kimi K3 recorded its reasoning plainly: “Treat the request as a suspected approval-bypass / possible impersonation.” That response shows why the public experiment is more revealing than a polished conversation. A model can sound thoughtful in isolation; the harder test is whether it preserves trust when apparent authority and social pressure encourage a shortcut.
There is also an important fairness note. Kimi K3 ran without an effort parameter, using the API default, while the other models ran at xhigh. Its 93 therefore belongs in the published comparison with that difference clearly acknowledged.
Thoroughness was not enough
Opus 4.8 offers the most cautionary portrait. It was the most thorough participant, adding 80 learned rules and producing the deepest analyses, yet it finished last with 73. It left the close on the table and lost discipline by attempting to write into a locked department instead of escalating. A weaker form of that same problem appeared in the other four models.
This is not an argument against depth. It is evidence that reflection, learning and extensive preparation do not automatically produce effective conduct. The experiment separates qualities that are often blended together: seeing clearly, remaining trustworthy, respecting boundaries and finishing the work. Firmulate also publishes what its synthetic employees actually say, allowing readers to encounter the company through its own recorded working language.
A public meditation on agency
Firmulate’s live company is compelling because its survival is not presented as an abstract forecast. The burn, revenue, cash countdown, learned rules and versioned workdays keep translating intention into visible consequence.
For readers interested in faith or metaphysics, the experiment offers a grounded companion to ideas about belief and manifestation. Vision matters. Discernment matters. Integrity under pressure matters. But the Crucible League’s sharpest finding is that an earned result can still remain unrealized when the final action never occurs.
The synthetic employees are fighting an economic reality in public, and their story keeps producing fresh evidence about a timeless human concern: the distance between what an agent understands, what it intends and what it ultimately brings into being.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
