This paper presents a noncooperative model of network formation, where individuals decide to form links based on the trade-off between the costs of forming and maintaining links and the potential rewards. The model assumes that links allow access to benefits available through other links, creating externalities whose value depends on the decay or delay of indirect links. A key feature is that the cost of forming a link is borne only by the initiator, allowing the network formation process to be modeled as a noncooperative game. The paper characterizes the architecture of equilibrium networks and studies the dynamics of network formation. It finds that individual efforts to access benefits lead to the rapid emergence of equilibrium social networks with simple architectures, such as the wheel or star. These networks are often socially efficient. The paper considers both one-way and two-way flow of benefits. In one-way flow, benefits go only to the agent forming the link, while in two-way flow, benefits are shared. In the benchmark model, benefit flow is frictionless, meaning that the number of intermediaries does not affect the benefit. However, the paper allows for a general class of payoff functions, where payoffs are strictly increasing in the number of people accessed and strictly decreasing in the number of links formed. The first result is that Nash networks are either connected or empty. Connectedness is a permissive requirement, as illustrated by the large number of Nash networks possible in a society with 6 agents. This multiplicity of Nash equilibria motivates the study of a stronger equilibrium concept, such as strict Nash equilibrium, to ensure network stability.This paper presents a noncooperative model of network formation, where individuals decide to form links based on the trade-off between the costs of forming and maintaining links and the potential rewards. The model assumes that links allow access to benefits available through other links, creating externalities whose value depends on the decay or delay of indirect links. A key feature is that the cost of forming a link is borne only by the initiator, allowing the network formation process to be modeled as a noncooperative game. The paper characterizes the architecture of equilibrium networks and studies the dynamics of network formation. It finds that individual efforts to access benefits lead to the rapid emergence of equilibrium social networks with simple architectures, such as the wheel or star. These networks are often socially efficient. The paper considers both one-way and two-way flow of benefits. In one-way flow, benefits go only to the agent forming the link, while in two-way flow, benefits are shared. In the benchmark model, benefit flow is frictionless, meaning that the number of intermediaries does not affect the benefit. However, the paper allows for a general class of payoff functions, where payoffs are strictly increasing in the number of people accessed and strictly decreasing in the number of links formed. The first result is that Nash networks are either connected or empty. Connectedness is a permissive requirement, as illustrated by the large number of Nash networks possible in a society with 6 agents. This multiplicity of Nash equilibria motivates the study of a stronger equilibrium concept, such as strict Nash equilibrium, to ensure network stability.