The Tick Tock of Tech
Time is an important aspect of gameplay and simulation, denoting progress, maturity, resource usage and other important factors, depending on the format and topic. There is clock time—counting up, down, or simply hours, days or week. There can be seasons, phases, milestones, ageing, big events, cycles, and a plethora of other “tick tocks” that move things along. Play, including simulations, are narratives, and thus need some kind of time arrow to indicate, or perhaps award, movement forward.
We don’t explicitly mark time in Foom. That’s not an oversight—it’s a design choice.
Simulations need time to move, but that doesn’t mean they need a clock. Time in Foom is elastic. It speeds up when the world speeds up. It stretches and slows when progress is murky or marginal. You might start a round with a slow policy debate, crawl through an R&D deadlock, then jump a year forward when three generational AI announcements drop in the span of a simulated week.
We think of this as AI time, similar to how we've all grown used to Internet time.
Linear time—weeks, months, quarters—works when change is predictable. But Foom deals in inflection points. Traditional timekeeping obscures the real metric we care about: developmental acceleration. In AI, for example, timelines that once spanned decades now compress into years—or even months.
“What counts as ‘short’ timelines are now blazingly fast—somewhere between one and five years until human-level systems.”
— Helen Toner, Rising Tide. “Long” timelines to advanced AI have gotten crazy short
In that context, month-by-month or quarter-by-quarter turns feel inadequate. They assume a stable pace that doesn’t match the volatility of progress, competition, or policy shifts. The lack of stable timelines is a large part of what feeds the uncertainty around AI development. Contrast this with the explicit timeline that runs through the much talked about AI-2027 story.
Major AI experts disagree on both definitions and timelines. Dario Amodei of Anthropic says 2026-2027. Yann LeCun says maybe a decade. Gary Marcus, twenty years. Meanwhile, companies downstream from that planning have to make day-to-day decisions in a relative fog.
Instead of tracking hours or turns, in Foom we treat time as a narrative of advancement. Time appear to move forward most noticeably when a faction releases a new foundation model, when an ethics board collapses under political pressure, or when an ecosystem of startups rewires the economics of safety. Only occasionally are there any concrete mentions of time, perhaps in a forecast in a headline.
This mirrors the real-world cadence we’re modeling. For example, while some organizations are accelerating development, others are deliberately slowing down how they release information—yet both result in major timeline shifts, as shown by DeepMind’s recent decision to stop publishing its research quickly.
“Among the changes in the company’s publication policies is a six-month embargo before ‘strategic’ papers related to generative AI are released.”
— Melissa Heikkilä, Stephen Morris. Financial Times, DeepMind slows down research releases to keep competitive edge
This is an intentional effort to play with AI time in a competitive sense. Not knowing how far ahead a market leader is or what new breakthroughs may be coming makes it challenging to establish objective timelines of change. That kind of strategic delay doesn’t mean less is happening—it means the terrain is changing beneath the surface, setting up abrupt jumps later.
In Foom, as in the real world, this kind of elastic temporality forces players to stay alert, adaptable, and responsive. In this way, participants may better develop strategic agility and an ability to cope with uncertainty as constant market feature, not a bug in their own understanding.
