liveBook · Manning

Cost

As explained in Machines that Think

Cost is the resource expenditure required to train, evaluate, deploy, or operate a computational system.

For large language models, cost is part of the conditions that make systems possible, not merely a consequence of building them. Training and deployment depend on computation, energy, and infrastructure, so the demands of scale shape which models can be built and used. section 4.0 section 4.6

Cost also influences research choices. Architecture, model scale, and training strategy must be affordable and sustainable; approaches that fit existing hardware and train predictably are favored over those difficult to implement, evaluate, or reproduce at scale. section 4.6.4

The pressure of cost shapes research in several ways:

  • Researchers commonly test ideas on smaller models first, reserving expensive frontier-scale evaluation for approaches with promising results. section 4.6.4
  • Long planning cycles and high-stakes experiments leave less room for rapid iteration, encouraging incremental improvements over radical departures. section 4.6.4
  • Over time, investment and specialized infrastructure concentrate effort on proven approaches, while other research directions may remain unexplored. section 4.6.4

How Machines that Think covers it

How deep it goes
Connects training and deployment expense to design choices and research constraints, without quantifying costs.
Examples
Contrasts small-model testing with costly frontier-scale evaluation and favors predictable, hardware-compatible approaches.
What you can do after
Evaluate how affordability and infrastructure shape viable model designs, experiments, and research directions.
How it connects
Links cost to model scale, computation, energy, infrastructure, architecture, and training strategy.

Also explained in

Compare how 17 books explain cost →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machines that Think. The section links point to the parts of the book it draws on.