Drishti
The Office as a Computer
A room that converts energy into information
A room that converts electricity into decisions
At a glance
- Flows
- Electricity, attention, email, Slack, decisions
- Optimizes
- Throughput of revenue-generating information
- Persists
- Legal entity, brand, API contracts, org chart agreements
- Likely future
- AI agents as coworkers; hybrid remote equilibrium
Seven lenses on The Office as a Computer
Same order every time. Expand the lens you need; first is open by default.
◎ What Exists — What is this, really?
The office is mostly agreements wearing furniture.
Desks and laptops are physical. Headcount, roles, and OKRs are informational. Culture and trust are agreements — they exist because everyone acts as if they do.
- Physical: power, HVAC, chairs, servers
- Informational: tickets, docs, metrics, code
- Agreement: titles, meeting norms, who decides what
↻ What Changes — What is stationary vs non-stationary?
Markets, headcount, and roadmaps drift faster than HTTP.
- Fast: priorities, morale, tool churn
- Slow: org chart, tech stack, office lease
- Semi-invariant: company identity, regulatory duties
Models that assume stable team topology break after reorgs — treat org charts as non-stationary.
→ What Flows — What moves? Where are bottlenecks?
Electricity in; decisions and shipped code out.
flowchart LR Power["Power & internet"] --> People["People & agents"] People --> Tools["Tools & docs"] Tools --> Out["Shipped work"] People --> Attention["Attention queue"]
Bottlenecks: meeting load, review queues, unclear ownership. Queueing Theory applies directly — every standup is a scheduler.
◈ What Learns — What updates, remembers, optimizes?
Organizations learn via retros, policies, and muscle memory.
- Short memory: Slack threads, sprint notes
- Long memory: wikis, code, incident postmortems
- Optimization target: often local (team KPI) not global (company)
Without shared learning loops, the same incident repeats with new logos.
▣ What Persists — What survives change?
What survives a rebrand or pivot?
- Legal entity and bank accounts
- Core incentive: must capture value to pay salaries
- Customer trust (if earned)
- Technical debt (unwanted but persistent)
Migrations fail when teams ignore what must persist — IDs, contracts, SLAs.
✦ What Emerges — How do simple rules become complexity?
No one designed “the company” — it emerged.
Local rules: reply on Slack, ship Friday, escalate P0. Global patterns: shipping cadence, silos, rumor speed.
Gossip protocols model how information epidemics create culture without a central broadcaster.
? What Will Happen — Which futures are becoming likely?
Branches: hybrid HQ, remote-first, or agent-heavy ops.
| Branch | Rough odds | Leading indicator |
|---|---|---|
| More AI agents in workflows | Rising | % tasks closed without human |
| Remote-first default | Stable | office seat utilization |
| Consolidation / layoff cycle | Cyclical | runway / revenue ratio |
Momentum: tooling for agents is accelerating; physical office momentum is flat in many sectors.
Update odds when leading indicators move — don’t wait for the annual plan.