The unprecedented global lockdown in response to the COVID-19 pandemic exposed the extreme vulnerability of “essential” yet underpaid workers, the vast inequality between the wealthy and the less fortunate, and the bottomless pit facing those without a social safety net. While the crisis has laid bare the near-universality of human susceptibility to disease and unemployment in a world in which few can safely work, it has also highlighted the disproportionate precarity experienced by low-wage and contingent workers, people of color, and non-citizens. Well before the pandemic, rampant socioeconomic and racial inequality, high underemployment and concentration of wealth, and technological advances threatened to render many human workers obsolete, powerless, and in need of social support and care. At the same time, in the United States, the Trump administration targeted an already shrunken social safety net for elimination through the extension of punitive workfare ideology originally reserved for poor, single mothers to all forms of government-funded support, including health care, nutrition, and housing assistance available those just above the poverty line regardless of parental status. Such ideology effectively conditions public support to needy individuals (some already employed at low wages) on their willingness and ability to engage in (more) work, regardless of pay, employment conditions, or caregiving obligations. Left to our own devices, whether in times of crisis or calm, advocates have sought to strengthen interpersonal relationships and community bonds for the provision of basic social support for the poor, as a way for ordinary people to help those whom the government will not. This Article examines two experimental models—restorative justice and “radical help”—that seek to reform welfare administration explicitly to weave people back into the fabric of the social safety net. These social welfare innovations foreground human relationships as an underutilized resource to highlight the power of meaningful social connections to help those experiencing everything from disability and discrimination to bad luck not just avoid disaster, but thrive and flourish in strong communities. Each model emphasizes human relationships to help poor people benefit voluntarily from social supports and community engagement instead of punishing them for noncompliance with paternalistic and exploitative government program work mandates. Such relationships can center poor people’s lived experiences and combine collaborative, localized, and responsive community support with technology to facilitate social networking and, ideally, increased economic security and empowerment. At the same time, without appropriate safeguards or oversight, overreliance on private relationships for social welfare provision risks replicating existing forms of disempowerment. In practice, both models risk reinscribing a private, marginalized sphere, neither restorative nor radical, in which those who perform the work of nurturing relationships remain subject to the will of those with power to offer or withhold assistance. Cautious optimism must be combined with meaningful protections in order to preserve the most promising aspects of new models while preventing the worst harms of what could be in effect a return to private, discretionary provision—or deprivation—of social support. Informed by feminist and antiracist theories critical of both market relations mediated by the state and private family relations entirely insulated from oversight, this Article concludes that we must continue to explore and adapt new models of welfare provision that truly protect and promote all human potential.