For more than two decades, courts have embraced the rule that computer code is "speech" within the meaning of the First Amendment. While this principle has been celebrated by some, it should now be cause for great concern. In recent years, technology companies have relied on the doctrine that "code is speech" and other expansive theories of First Amendment coverage to claim sweeping constitutional protections for their commercial activities. If this trend continues, it will lead to consequences that are both dangerous and absurd. This Article argues that the prevailing approach to First Amendment coverage for code is fundamentally misguided and must be abandoned. Such a capacious vision of First Amendment coverage for code is ill-suited to the realities of the modern world. Code flows through nearly all of our daily interactions. Adding a constitutional valence to every one of these activities would be a perilous proposition. It would give opportunistic litigants free rein to manipulate the First Amendment to insulate their private power from public accountability, and it would lead to a free speech jurisprudence that is unwieldy, incoherent, and unmoored from democratic values. To avoid these outcomes, courts must reject abstract and categorical rules like "code is speech." Instead, courts should adopt a framework for evaluating code that is grounded in the values that the First Amendment was meant to serve. This involves answering fundamental questions about what the First Amendment is for-and who it is for. In the coming years, as courts are called upon to shape the contours of regulation in the digital age, it will become all the more urgent to elucidate a First Amendment jurisprudence that actually makes sense for the modern world. Addressing these concerns is critical, not just for the future of governance, but for the future of free speech.