The Hacker
Author’s Note
#Every reader now lives inside authorization. A phone can refuse its owner. A router can expose a household. An account can contain years of work while denying the person who made it. The person who discovers why may be praised, fired, sued, or charged depending not only on what the machine revealed, but on who had permitted the question.
#This volume asks what happened when American law placed a federal criminal boundary around that act of discovery. It does not collapse exploration, theft, extortion, espionage, destruction, fraud, and violence into one moral category. Those remain distinguishable acts with distinguishable victims. The inquiry begins one enclosure above them: what happens to a technological civilization when the law governing access to its machines also governs how its people may learn the limits of those machines?
Part I: The Boundary
Chapter 1: Annunciation
#We begin with a delicate balance of power.
#The Cold War had divided the world between two competing systems: communism and capitalism. Every nation was pressed to align with one pole or the other, each offering protection beneath nuclear arsenals pointed across the divide. So long as the two superpowers maintained mutually assured destruction, the world was said to be safe. Safe, at least, once you had chosen a side.
#Then the personal computer began escaping the institutions that had created it.
#Computational power moved into homes. Telephone lines reached nearly everywhere. A machine owned by an ordinary person could call another machine belonging to a school, corporation, university, or government. The institutions remained large, but the distance between them was disappearing. The modem turned curiosity into motion.
#In 1983, WarGames carried this new condition into public consciousness. Matthew Broderick played David Lightman, a teenage computer enthusiast who went searching for unreleased games and found a military system instead. He possessed a modem, technical curiosity, and the will to explore the world. He nearly initiated global thermonuclear war.
#The story was fiction. The fear it activated was real.
#President Ronald Reagan watched the film at Camp David and soon asked his national-security advisers whether such an intrusion could actually occur. Computer crime was already under consideration, and no movie wrote an act of Congress by itself. But WarGames gave political form to a fear whose real-world boundaries had not yet matured. It helped make catastrophic unauthorized access imaginable before legislators had lived in the networked world they intended to govern.
#Congress enacted federal computer-crime provisions in 1984 and substantially expanded them in the Computer Fraud and Abuse Act of 1986. The law addressed real dangers: theft, fraud, espionage, destruction, and interference with systems upon which lives could depend. Yet it placed extraordinary causal force upon a boundary that computing would make increasingly difficult to define.
#Authorization.
#The CFAA divided the people who knew how to use computers into authorized and unauthorized states. Through enclosure, it created the hacker.
#It did not create curiosity, circumvention, trespass, or technical prowess. Human beings had opened, altered, repaired, repurposed, and exceeded the intended uses of machines for as long as there had been machines. The law created a legal identity around that behavior. On one side stood the authorized user, whose interaction remained legible to the owner of the system. On the other stood the unauthorized user, whose interaction could enter federal criminal machinery before its purpose and completed consequences had been fully distinguished.
#Exploration, acquisition, disclosure, fraud, destruction, and espionage remained different in the world. The law supplied a handle by which prosecutors could seize them at their common entrance.
#That handle persists.
#The feared teenager reaching nuclear command did not become an ordinary condition of American life. Military systems were not made lawful to enter from a bedroom modem and then protected from catastrophe only by the CFAA. The law did not avert WarGames. It enclosed the civilian computational world that arrived afterward.
#The country was about to discover that a law can project causality far beyond the event it was written to prevent.
Chapter 2: What the Law Enclosed
#Let us take a careful look at the hacker the law enclosed, and then at everything enclosed with him.
#For the moment, we will not ask what encloses the CFAA: not the Cold War, Congress, property, sovereignty, national security, or the institutional fear of machines escaping institutional control. We begin by asking what the CFAA encloses.
#It encloses the whole of computer science within its jurisdiction.
#This is not because every act of computer science violates the law. It is because the legal distinction projects into the field at precisely the boundary where some of its most important knowledge appears.
#Computer science advances through construction and proof, but also through failure, misuse, circumvention, unexpected behavior, and the discovery that a system can do something its designer did not know it could do. The authorized user learns what the system is intended to provide. The adversarial learner discovers what the system will permit despite that intention.
#That knowledge appears in malformed requests, exposed credentials, abandoned interfaces, forgotten services, timing errors, privilege boundaries, insecure defaults, unexamined trust, and combinations no designer anticipated. It is not completely available in a manual because the manual describes the intended system. It is not completely available in source code because the operating system includes its dependencies, networks, users, configurations, mistakes, and surroundings. It is not completely available in a classroom because the most consequential failure may appear only when a real intelligence presses against a real boundary.
#These are vulnerabilities, defects, unintended responses, and unanticipated capabilities. They retain their names.
#Departure enters the model only after a reference condition is stated. A system owner represents the computer as accepting a defined range of requests and returning only the information assigned to each requester. The operating system can then be observed returning information beyond that represented condition. The departure is not the vulnerability, the request, the person who found it, or the information exposed. It is the describable relation between the stated reference condition and the observed state of the operating system. Naming that relation allows the model to preserve both states without replacing the technical facts with its own vocabulary.
#The CFAA projects legal causality directly into its formation. Its influence does not begin when an indictment is filed. It begins when a student decides whether to follow an unexpected response. It appears when a programmer decides whether to reproduce a defect, when a researcher decides how far to validate a vulnerability, and when an institution decides whether an unfamiliar person discovering an unfamiliar truth is helping or attacking it.
#At MIT, where some of the world’s most capable computer scientists are trained, residential network policy currently prohibits routers and devices performing routing or network-address translation on MITnet. The operational reasons are comprehensible: a misconfigured router can disrupt a shared network. The educational consequence remains visible. A student may own the hardware and possess the software, including OpenWrt, yet may not connect a personally administered router to that institutional network in router mode. The institution that teaches networks must also prevent students from independently becoming the network where the shared enclosure cannot safely contain the experiment.
#That example is not an accusation against MIT. It is a miniature of the national condition.
#American computer science can teach construction, authorized administration, controlled laboratories, and sanctioned security testing. At the live boundary, however, the student encounters overlapping institutional rules, contracts, disciplinary consequences, civil liability, and criminal law. The educational field narrows before the criminal case exists.
#Meanwhile, the discarded WRT54G router does not disappear. It can become one node among thousands. A person who understands its firmware, credentials, radio, update path, memory limits, and network behavior may understand what a botnet made from our discarded machines can do. If that knowledge cannot develop openly where the devices were designed and sold, it does not cease to be available. It develops somewhere else.
#This is the first signal made visible by the new science of Nested Causality.
#The CFAA does not merely punish a separable inventory of crimes after they occur. It is an outer enclosure projecting causality inward into education, employment, research, institutional security, and the imagination of the end user. It changes who explores, where exploration occurs, which discoveries become public, which capacities are professionalized, and which capacities migrate beyond the enclosure entirely.
#The law is very good at enforcing itself. The question is what else it has been enforcing.
Part II: The Hammer
Chapter 3: The Sentence Before the Sentence
#The force of computer-crime law cannot be measured only in the sentences judges finally impose.
#It begins with the possible sentence.
#Federal prosecution can stack counts arising from one course of conduct. It can classify response costs and other consequences as loss. It can invoke sentencing guidelines designed for fraud, theft, and property destruction. It can offer a plea whose apparent mercy derives from the magnitude of the alternative. It can consume years, money, health, relationships, reputation, employment, and the ability to participate in the field where the accused acquired an identity.
#The sentence begins before sentencing.
#An empirical study of CFAA access-to-information cases found an identity crisis in the law. Unauthorized access is often compared to trespass, but sentencing proceeds under the economic-crime guideline, where calculated loss can dominate the result. The law enters through access and accumulates force through the accounting architecture of fraud.
#This is how a broadly written boundary becomes a club.
#The problem is not that every prosecutor is cruel or every defendant harmless. The problem is that enormous leverage becomes available before the state has adequately distinguished the human and material meaning of the conduct. Harmful conduct may justify severe punishment under laws addressing fraud, theft, extortion, destruction, espionage, threats, identity theft, and interference with critical infrastructure. The CFAA adds felony potential to access itself and allows the authorization boundary to project severity into conduct whose remaining effects may already be separately punishable.
#The deterrent does not stop with the defendant.
#Every public prosecution informs an unseen population. The student who could have learned does not need to receive a subpoena. The founder who could have built does not need to see a courtroom. The researcher who could have reported does not need to be threatened directly. They need only observe what the system is capable of doing when it decides that someone has crossed the line.
#Fear becomes distributed without appearing in prosecution statistics.
#Capability exits silently.
Chapter 4: Aaron
#Aaron Swartz was a programmer, writer, political organizer, and advocate for public access to knowledge. He helped develop RSS, participated in the creation of Reddit, worked on Creative Commons, and became a public force against legislation that threatened the open Internet.
#Then he downloaded too many papers.
#The sentence is deliberately plain because the machinery that followed was not.
#Swartz used MIT’s network to download a large collection of academic articles from JSTOR. The facts included efforts to continue downloading after technical countermeasures were imposed and physical access to a network closet. They can be stated without turning the act into violence, espionage, or the destruction of a computer system.
#JSTOR recovered the files and did not seek his prosecution. Federal prosecutors continued. A superseding indictment ultimately charged thirteen felony counts carrying immense theoretical exposure. United States Attorney Carmen Ortiz’s office pursued the case under Attorney General Eric Holder’s Department of Justice. The government maintained that its plea position was appropriate. Critics saw a catastrophic mismatch between conduct, harm, and coercive leverage.
#Swartz died by suicide in January 2013 before trial.
#The fatal act was suicide. Naming the act does not complete its causal history. Swartz’s family and partner called his death “the product of a criminal justice system rife with intimidation and prosecutorial overreach.” The prosecution was not background scenery. It was an active enclosing condition: a young man faced federal power, felony identity, prison exposure, public judgment, financial pressure, and the possible destruction of the future through which his work had meaning.
#The law designed in the shadow of a fictional teenager approaching nuclear command had reached a real young American downloading academic literature.
#The causal projection did not end with Aaron.
#It entered every classroom where his name was known. It entered the decision of every technically ambitious student who understood that institutional displeasure could be translated into federal exposure. At MIT, the place where hacking had once named playful technical ingenuity, the word carried a body.
#The prosecution did not merely remove one person. It changed the conditions under which others would choose to become visible.
#Ask why Americans traveled abroad, hid behind handles, joined distributed collectives, accepted institutional capture, or abandoned unsanctioned exploration. Aaron’s prosecution does not answer every case. It belongs inside the answer.
Chapter 5: The Address AT&T Published
#Andrew Auernheimer called himself weev.
#In 2010, Apple released the first cellular iPad in the United States through AT&T. To make account login convenient, AT&T configured a web endpoint to prepopulate an iPad subscriber’s email address when the request contained that device’s Integrated Circuit Card Identifier, or ICC-ID. The identifier occupied a predictable numerical range.
#Daniel Spitler discovered that changing the number in the public request produced another subscriber’s email address. The server required no password. It presented no login challenge. It did not reject the request as unauthorized. AT&T had placed the addresses on the web and programmed its server to return them when asked the right numerical question.
#Spitler wrote a script that asked repeatedly.
#The program incremented the ICC-ID, submitted the request, and retained the email address AT&T returned. Spitler and Auernheimer refined the process and collected approximately 114,000 addresses. The list included prominent people in business, government, media, and the military.
#This was organic red teaming in its most difficult form.
#AT&T had made a security error. Two outsiders found it without authorization, a contract, a budget, a statement of work, or an approved testing environment. Their method demonstrated the scale of the exposure more conclusively than one manually retrieved address could. Their discovery gave AT&T information it had failed to produce internally: its convenience feature was also a population-scale disclosure mechanism.
#Auernheimer did not begin with Gawker as though publicity were the first available boundary. His account is that AT&T was given an opportunity to understand and patch the error first. No confirmation reached him that the company understood the population-scale exposure or had secured the subscribers whose addresses remained available. He escalated to the press because a warning without confirmed recognition had not enclosed the problem.
#The appellate record describes the contact indirectly: while the addresses were still being collected, Auernheimer contacted members of the media; some of them emailed AT&T; AT&T then closed the endpoint; and Gawker published after the repair. The record does not identify the direct AT&T contact described in Auernheimer’s account, but it does establish the sequence the earlier draft erased. Gawker was not handed a live method and asked to publish it before AT&T had a chance to act. Publicity followed the attempt to make the company encounter its own error.
#Gawker reported the vulnerability and named some affected subscribers while redacting the published addresses and identifiers. A responsible argument can still be made about the scale of collection and the transfer of the dataset to a reporter. It cannot honestly begin with the claim that direct notice should have come first when notice and an opportunity to repair were already part of the sequence.
#That argument concerns judgment inside an act that revealed a real defect.
#The federal response converted it into something else. Prosecutors charged Auernheimer with conspiracy to violate the CFAA and identity fraud. The government treated the public server’s responses as information obtained without authorization even though AT&T had installed no password or other technical gate between the requester and the addresses. The predictable ICC-IDs and the email addresses returned beside them became “means of identification.” A disagreement about how an outsider demonstrated a company’s failure became two felony convictions.
#Auernheimer was sentenced to forty-one months in federal prison and ordered to pay AT&T more than $73,000 in restitution.
#Prison did not make Auernheimer simple. It did not make him safe, agreeable, or easy to recruit into anyone’s preferred account of the case. David Kushner’s 2014 profile records that Auernheimer taught members of the Aryan Brotherhood at Allenwood to sing “Springtime for Hitler.” After officials ended the improvised systems through which he communicated from prison, he spent periods in solitary confinement. When the Third Circuit released him after approximately a year, he emerged more furious, publicly displayed a swastika across his chest, and continued rhetoric that was racist, antisemitic, and at times openly violent.
#None of this repairs AT&T’s endpoint, supplies the password that never existed, or turns conduct elsewhere into venue in New Jersey. It makes the causal result harder to sentimentalize. The state found a technically formidable and profoundly antagonistic person, imprisoned him for querying arithmetic addresses on a public server, isolated him, and returned him to the network angrier than before. A prosecution defended as deterrence became part of the education and mythology of the person it intended to deter.
#The prosecution selected New Jersey even though Auernheimer was in Arkansas, Spitler was in California, and the relevant AT&T servers were in Texas and Georgia. Approximately 4,500 exposed addresses belonged to New Jersey residents. That consequence was used to bring the case into a state where no essential conduct had occurred.
#The Third Circuit vacated the conviction in 2014 because venue was improper. Its decision did not finally resolve whether querying the unprotected endpoint violated the CFAA. Auernheimer had already entered prison.
#The sequence matters more than the eventual reversal:
AT&T publishes an unintended response
-> outsiders discover the pattern
-> automation establishes its scale
-> warning does not produce confirmed recognition
-> press escalation makes the failure undeniable
-> AT&T closes the exposure
-> the discoverer becomes the felon
-> prison follows the repair
#AT&T received the security benefit. The public learned that the exposure existed. Subscribers received a repaired system. The organic red team member received a prison number.
#Nothing in this account requires Auernheimer to be admirable, polite, institutionally acceptable, or correct in everything he later believed. An organic red team is valuable precisely because reality does not first recruit a suitable employee and then reveal a vulnerability to him. The defect becomes visible to whoever finds it. If legal protection depends upon the researcher’s character, manners, politics, or retrospective willingness to follow an institutional process that did not produce the discovery, the system protects conformity rather than security.
#The case demonstrates the distinction this volume needs. Organic red teaming is not a reporting program. It is naturally occurring adversarial pressure generated wherever technically capable people interact with systems. It may be inconvenient, abrasive, excessive, embarrassing, or badly communicated. Its value lies in making the system encounter a vulnerability before a more destructive actor does.
#The intervention is not to approve every method. It is to stop turning unauthorized use alone into the felony that erases the domestic source of the work.
Chapter 6: Anonymous
#Anonymous was never one organization with one command structure. It was a name, an image, a method of assembly, an invitation, and sometimes an alibi. People entered and left. Operations mixed protest, spectacle, disruption, disclosure, cruelty, comedy, political action, and crime.
#The name emerged from an interface before it became an identity. On 4chan’s /b/, people posting without names appeared as Anonymous. The conversation could be juvenile, offensive, surreal, cruel, inventive, and brilliant in the same hour. No application was required. No permanent identity had to survive the next sentence. Anonymous was what remained when the speaker disappeared but the signal continued.
#In 2008 that condition met the Church of Scientology. Critics had opposed Scientology for decades. Writers, former members, journalists, Usenet participants, website operators, and publishers had endured demands, investigations, threats, and litigation. Anonymous was the first opposition whose complete form did not have to endure litigation because there was no complete legal person to sue. An individual participant could still be identified, threatened, charged, or brought into court. The opposition itself had no office to raid, treasury to exhaust, leader to enjoin, or durable body upon which process could be served.
#Project Chanology began after Scientology attempted to suppress circulation of a video featuring Tom Cruise. What began on imageboards and related channels moved into worldwide street demonstrations. The Guy Fawkes mask did more than conceal a face. It displayed a discovery about power: an institution practiced in making criticism expensive had encountered criticism without a stable attack surface.
#Anonymous did not need to destroy Scientology to change the balance. It made openly resisting Scientology reproducible by people who did not first need the resources to survive Scientology’s lawyers. Litigation could still reach a body. It could not enclose the name.
#The Guy Fawkes mask promised a digital citizen too distributed to imprison.
#The state found the bodies beneath it.
#Young participants in Anonymous operations were investigated, charged, recruited as informants, sentenced, and separated from the communities through which their capabilities had formed. Some conduct caused real damage. Some exposed private information. Some became entangled with threats, fraud, or other offenses. A lawful analysis does not flatten those acts into harmless curiosity.
#It also does not flatten the people into the crimes charged against them.
#The United States possessed a cohort of technically capable, politically motivated young people who believed computer systems could be made answer to public power. The state encountered that cohort principally through investigation and prosecution. It was able to convert participants into defendants and informants. It had no comparably visible pathway for converting the movement’s adversarial competence into a domestic public red team.
#The Anonymous prosecutions amplified the same lesson as Swartz: unauthorized prowess was a liability before it could become a civic asset.
#Some Americans left. Some disappeared into ordinary work. Some went silent. Some continued from abroad. Some crossed from idealism into conduct that made recruitment impossible. These were not equivalent outcomes, but they shared an interrupted origin.
#The days of the American whiz kid as an all-powerful electronic vigilante began to close. The archetype remained in film. In life, the handle increasingly belonged either to an authorized professional or to someone elsewhere.
Part III: Dominion
Chapter 7: The Day the Cloud Was Seized
#Kim Dotcom does not belong in this book because Megaupload was a CFAA prosecution. It was not. He belongs because the same prosecutorial culture projected American authority across borders and over a computational system used by the world.
#Megaupload was among the first global cloud-storage platforms to reach enormous scale. It claimed more than 150 million registered users, 50 million daily visitors, and approximately four percent of Internet traffic. The United States later repeated those figures in announcing the indictment, while other measurements placed its share lower in particular markets. Even the narrower account describes infrastructure of extraordinary reach.
#Megaupload was not YouTube. It did not present a searchable public catalogue of the files stored by its users. A person uploaded a file and received a generated URL. Someone else could retrieve the file only by possessing that address, ordinarily because the uploader had supplied it somewhere beyond Megaupload. A private file sitting behind an undistributed URL was not a public performance waiting to be taken down. It was stored data.
#Copyright owners used Megaupload’s abuse system to identify particular public links and demand that access through those links be disabled. Those URL-specific notices were not cease-and-desist letters ordering Megaupload to end an unlawful business. They did not provide the company with notice that the United States considered the enterprise itself a criminal conspiracy. Megaupload’s outside counsel, Ira Rothken, said the company first learned of the indictment from the Department of Justice’s public announcement.
#No cease-and-desist progression preceded the destruction of the business.
#The unresolved leap begins before the copyright theory. It begins in surveillance and operational planning.
#At the request of New Zealand police, the Government Communications Security Bureau intercepted the communications of Dotcom and an associate before the raid. Dotcom was a permanent resident and therefore belonged to a class the intelligence service was then prohibited from targeting. The New Zealand government later acknowledged that the surveillance was unlawful. Prime Minister John Key publicly apologized to Dotcom and to the country for the state’s failure to provide the protection its own law required.
#Megaupload received no cease-and-desist demand against its business, no summons, and no opportunity to answer the government’s theory before enforcement. The state did not ring the doorbell. Helicopters and guns were its first attempt at communication.
#It planned an assault. Internal police material considered the possibility that the use of an elite tactical group and helicopter would appear “heavy-handed” and “over the top.” Nevertheless, in January 2012 a sealed United States indictment became a coordinated seizure and arrest. New Zealand police arrived at the rural mansion by helicopter and vehicle with armed officers and a police dog. Dotcom, members of his family, and employees encountered the physical force of an operation designed as though the alleged copyright entrepreneur presented a threat that ordinary process could not contain.
#Perhaps police feared flight, resistance, destruction of evidence, or remote deletion. Those concerns can explain precautions. They do not complete the missing comparison. Dotcom was a conspicuous resident operating a famous business from a known home. The alleged offenses were nonviolent. The servers were distributed through commercial hosting facilities and could be preserved through orders directed at infrastructure providers. The government has not made the necessity of overwhelming residential force as visible as the force itself.
#Why was there no summons? Why no demand to preserve evidence, surrender specified equipment, retain counsel, and appear? Why was an ordinary knock at the door treated as insufficient before the person behind the door had been given an opportunity to answer it?
#That is the causal transformation. Suspicion of a technological offense crossed an institutional boundary and returned as a threat model. The size of the business became the scale of the presumed conspiracy. Technical capacity became presumed capacity to destroy evidence. Wealth became presumed capacity to flee. The word hacker supplied the missing danger. By the time police reached the house, the prosecution’s interpretation had already hardened into tactical reality.
#The operation expressed a claim of dominion over the computational world, but it expressed that claim with guns rather than with superior technical knowledge. American institutions asserted authority over a foreign platform they could not enter, understand, preserve, separate, or correct through their own command of its architecture. They could not distinguish every lawful file from every alleged infringement before seizure. They could not remove only the criminal conduct while preserving the living system and the property of innocent users. They could, however, compel another nation to surround the founder’s home, take the machines, disable the names, and end the service.
#The deficiency was not an absence of American intelligence. It was the result of an enclosure that had restricted the development of independent technical prowess while expanding the instruments of enforcement. The state had become more capable of taking a computer system than understanding one. Its claim of technological dominion was therefore carried by its oldest technology: organized force.
#Domain names were seized. Servers were searched, frozen, or left beyond the company’s control. The service disappeared before Megaupload or its officers received a criminal trial.
#The government alleged that the platform was a deliberate criminal enterprise: that it encouraged infringement, rewarded popular uploads, retained duplicate files after identified links were disabled, and generated at least $175 million. But the architecture matters to each allegation. Deduplication meant that identical data could be stored once while more than one user or link referred to it. Disabling a reported public link did not necessarily establish that every private reference to the same bytes was itself an infringing public distribution. Removing every identical stored object would also remove lawful personal copies without a claim directed at those users.
#The absence of a public search function was similarly inverted. A design that prevented strangers from searching private storage became, in the indictment, evidence that the company concealed infringement. Elsewhere, the absence of search had been treated as evidence that a file host was not constructing a piracy catalogue. Features of a private storage system were interpreted through the conclusion that the system was criminal.
#The prosecution then attempted to carry concepts of secondary civil copyright liability into a criminal case against a foreign storage provider. Megaupload’s defense argued that American law supplied no general crime of secondary copyright infringement and that automated storage did not make the company the direct infringer for each transfer initiated by a user. New Zealand courts spent years deciding whether the alleged conduct could satisfy extradition requirements, at one point holding that the asserted copyright offense itself did not supply the extradition path and examining fraud and other theories instead. Later appellate decisions reached different conclusions about the available New Zealand offenses.
#Two former executives eventually pleaded guilty to charges in New Zealand under agreements that ended the effort to extradite them. Their pleas are evidence concerning their own admitted conduct. They are not the adversarial American trial Megaupload never received before its infrastructure and market position were destroyed. Dotcom has continued fighting surrender to the United States.
#The allegations remain allegations where they have not been tried. The illegal surveillance is acknowledged. The tactical escalation is documented. The destruction was immediate.
#Millions of lawful users lost practical access to files held inside infrastructure now treated as the instrument of an alleged conspiracy. First-mover advantage, network effects, public recognition, customer relationships, and operational continuity could not be placed in evidence storage and returned after trial. The business existed in a volatile state. Once enforcement crossed the boundary, acquittal could not reconstruct the day before the raid.
#This is why Megaupload matters.
#The United States asserted that conduct occurring through servers, companies, employees, and users distributed across nations remained subject to American criminal power when American copyrights and infrastructure were implicated. The computational system was elsewhere. The authorization that mattered was American.
#After Megaupload, a founder did not need to be American, operate in America, or build the machine in America to feel the reach of American technological dominion. A properly formed foreign company with a globally successful service could be removed first and permitted to litigate the theory of its criminality afterward.
#The signal to startups was not simply that infringement was illegal. It was that technological scale could transform an unresolved legal theory into an existential enforcement event. You could build first, gather millions of users, establish a lawful company, and still discover that the institution interpreting your users possessed the power to end the experiment before judgment.
#America demonstrated power over the Internet. It did not demonstrate control of what that power would cause.
Chapter 8: The Suspect in Every Pocket
#The authorization boundary did not remain attached to the computer terminal.
#The computer entered the pocket. It learned where its owner went, whom the owner called, what the owner read, what the owner searched for, and which other devices occupied the same place. The system capable of identifying unauthorized access became capable of identifying unauthorized presence.
#Geofence warrants reversed the ordinary direction of investigation. Instead of beginning with a suspect and seeking evidence connected to that person, law enforcement could begin with a place and time, compel a technology company to search location histories, and identify devices that had entered the boundary.
#The January 6 investigation made this method nationally visible. Google location information helped identify people in and around the United States Capitol during the attack. Presence was not the only evidence used in every prosecution, and being near the Capitol was not by itself equivalent to entering it or committing violence. Yet the method placed thousands of devices inside an initial field of suspicion because their owners had carried computers through the wrong boundary at the relevant time.
#In 2026, the Supreme Court held that government acquisition of geofence location information is a Fourth Amendment search because people retain a reasonable expectation of privacy in the record of their movements. The Court did not abolish every geofence warrant. It recognized the constitutional event that had been obscured by the database.
#The progression is important.
#First, the person operated the computer.
#Then the law classified the person’s operation through authorization.
#Then the computer continuously operated upon the person.
#Finally, the resulting data became a means by which the state could locate a class of possible suspects before knowing their identities.
#The citizen ceased to experience technology only as an instrument of agency. Technology became a witness whose records could be summoned by institutions the citizen could not observe.
#The result was not only loss of privacy. It was loss of technological authorship. People increasingly experienced their devices as opaque systems controlled by corporations, updated without meaningful consent, watched by unknown parties, and available to government process. Social discussion filled with fear of algorithms, surveillance, manipulation, artificial intelligence, and systems nobody appeared able to control.
#The American whiz kid had become the American end user.
Part IV: The Exterior Learns
Chapter 9: The Archetypal Foreign Hacker
#While American law projected inward, its force weakened at the national boundary.
#The CFAA could indict a foreign operator. It could support sanctions, extradition, seizure, intelligence work, and diplomatic pressure. It could not place a Russian, Chinese, Iranian, or North Korean operator in the same everyday enclosure as an American student. Nation-state hackers tend not to stand where an American sentence can reach them. The Department of Justice has acknowledged this practical limitation in its own writing about cyber sentencing.
#The result was an asymmetric learning environment.
#The American learner faced a liberty-restricted inner enclosure. The foreign learner could occupy a comparatively liberty-activated outer enclosure, particularly where a government tolerated, protected, recruited, or directed the work. The difference was not moral character, ethnicity, or innate ability. It was accumulated permission to gain experience.
#Experience compounds.
#A failed intrusion teaches architecture. A successful intrusion teaches trust. Persistence teaches operations. Detection teaches concealment. Stolen code teaches design. Captured infrastructure teaches implementation. A community that can exchange these lessons without exposing itself to immediate domestic prosecution develops a body of practical knowledge no curriculum can reproduce.
#Russia’s post-Soviet conditions combined strong mathematical education, economic disruption, criminal markets, weak or selective enforcement, and eventual relationships between operators and state power. North Korea selected and trained technical workers whose activity could produce intelligence, disruption, and foreign currency. In 2014, the FBI attributed the destructive Sony Pictures intrusion to North Korea with high confidence. China developed cyber capacity at state scale, targeting intellectual property, telecommunications, government, and defense programs. A Defense Science Board report identified compromised information associated with numerous American weapons programs, including the Joint Strike Fighter.
#America paid to discover. Foreign enclosures learned to acquire.
#Iran provided the physical image. In 2011, an American RQ-170 Sentinel crossed into Iranian airspace and entered Iranian possession substantially intact. Iran said its forces caught the aircraft in an electronic ambush and brought it to earth with minimal damage. A later Iranian account described interrupting its communications and manipulating the coordinates by which it understood where home was. American accounts pointed instead to malfunction. The public record still does not conclusively establish the mechanism.
#The Iranian claim belongs at the center of this inquiry because it describes the exact security asymmetry taking shape. In that account, Iran did not merely find an American machine after it fell. Iranian operators understood a vulnerability in the aircraft’s communications, navigation, or recovery behavior well enough to make the machine leave the control of the people who designed and operated it. The RQ-170 entered their national enclosure, and they knew something operationally decisive about its security that its American enclosure had either not known, not closed, or not successfully defended in Iranian airspace.
#If Iran’s account is substantially correct, the capture is the thesis of this volume made physical: the foreign enclosure possessed greater practical knowledge of a security flaw in an American system than the Americans relying upon that system. Even the competing malfunction account cannot restore the lost control. Under either account, the United States sent an extraordinarily sensitive technological object into a foreign enclosure and could neither keep command of it nor recover it.
#Possession is not disputed.
#President Barack Obama asked Iran to return the aircraft. Iran refused and announced that it would reverse-engineer it. Iranian programs later produced flying-wing aircraft presented as descendants or copies of the RQ-170, although their internal similarity and performance remain debated. Iran’s drone program predated the capture, and the widely used Shahed-136 is not a direct RQ-170 copy. The captured aircraft nevertheless became a complete technological object available for observation, disassembly, imitation, and learning.
#Iran’s resulting deadly drone capability is no longer speculative. Iranian drones and their descendants have been used across regional conflicts and supplied abroad; flying-wing systems derived from the captured RQ-170 became an operational branch of that arsenal. By 2026, Iran possessed sufficient regional power to restrict traffic through the Strait of Hormuz and force negotiations over its reopening. No single aircraft produced that condition. The larger causal history includes domestic engineering, imported components, recovered systems, reverse engineering, cyber operations, missiles, proxies, geography, state organization, and decades of pressure. The RQ-170 belongs in that history as a consequential transfer of a complete American technological object into an enclosure determined to learn from it.
#The transfer could not be recalled.
#This is the foreign hacker at full scale: not merely a person at a keyboard, but an enclosure capable of receiving another enclosure’s technical vulnerabilities and retaining the resulting knowledge as national capacity.
Chapter 10: How America Lost Control
#America did not lose control because it stopped inventing.
#It lost control because invention and adversarial understanding were placed in different causal environments.
#The country retained exceptional universities, semiconductor design, operating systems, cloud infrastructure, artificial intelligence laboratories, defense contractors, venture capital, and global technology companies. It continued producing machines of extraordinary capability.
#It also constructed an authorization regime around how those machines could be explored. Institutions narrowed access to protect shared systems. Companies protected property and users. Prosecutors pursued visible domestic actors. Courts inherited statutes written before ordinary life became computational. Surveillance expanded the state’s ability to identify people through the systems they used.
#Each action could be defended within its local enclosure.
#Together they produced the larger pattern.
#The systems remained American. The most unrestricted experience of defeating their boundaries accumulated elsewhere.
#This does not mean America has no capable hackers. American intelligence agencies, military units, security companies, researchers, and red teams contain exceptional talent. The issue is organic formation. Institutional teams recruit from a population whose unsanctioned path into mastery has become dangerous. They receive people after authorization, credentials, employment, and clearance have made the talent legible. They cannot receive the discoveries never attempted, the skills never formed, or the person screened out before becoming useful.
#The domestic field does not need to fall to zero for the exterior to gain the advantage. It needs only to grow more slowly than the threat.
#The pattern then reinforces itself. Foreign intrusion produces fear. Fear produces tighter domestic controls. Tighter controls reduce unsanctioned domestic learning. Reduced domestic learning increases dependence upon institutions. Institutional failure produces another foreign intrusion. The intrusion is cited as evidence that more restriction is necessary.
#The law intended to protect the victim state helps reproduce it.
#Once seen, the pattern is difficult to unsee.
#We created the hacker by dividing the user through authorization. We then projected the criminal identity most forcefully upon the people inside our reach. We did not create every foreign hacker’s intention, state sponsor, or crime. We helped create the asymmetry in which the hacker is expected to be foreign and the hacked is expected to be domestic.
#The victim is part of the enclosure too.
Chapter 11: Security Is the Knowledge of Insecurity
#A secure computer is not a computer without vulnerabilities.
#It is a computer whose vulnerabilities are being discovered, understood, repaired, and displaced faster than they can be used to destroy the state the computer is expected to maintain.
#This distinction matters because every useful change to a computational system changes the thing being secured. A new feature creates new behavior. New behavior creates new relationships among code, data, users, machines, and networks. Those relationships create new boundaries. Somewhere among those boundaries, intention and operation will fail to agree.
#The disagreement is an attack surface.
#Computer security is therefore not a wall constructed once around a finished machine. It is a dynamic steady state between knowledge and ignorance. The system changes. Its users change. The systems enclosing it change. The people examining it learn. The people attacking it learn. A repair closes one route while altering the landscape in which the next route will be found.
#The cycle is ordinary enough to disappear beneath its names:
#Functionality produces attack surface. Attack surface contains unknown behavior. Discovery converts unknown behavior into knowledge. Knowledge becomes repair or exploit. Repair changes the system. The changed system produces another attack surface.
#There is no final secure computer inside this cycle. There is only a system whose unknown behaviors and vulnerabilities are being found by one population before another population can profit from them.
#Now the effect of the legal enclosure becomes measurable in kind, even before it is measurable in quantity. A jurisdiction can forbid unauthorized exploration within its reach. It cannot prevent the same machine, protocol, operating system, service, or mistake from being explored beyond its reach. It does not remove the unknown behavior. It changes the distribution of people permitted to become familiar with it.
#The technical knowledge necessary to use a technology to control a technology includes security. The knowledge of security includes the practice of finding insecurity. A country that protects its machines by restricting that practice may reduce some immediate intrusion while weakening the process that would have converted unknown vulnerabilities into domestic knowledge.
#The thing to secure is not the absence of the practice.
#The thing to secure is the knowledge and capacity.
#Nested Causality makes the transfer visible. The law is an outer enclosure. It projects risk inward through prosecution, sentencing, education, employment, institutional policy, and family expectation. The inner system adapts. Some people obtain authorization. Some confine their inquiry to approved exercises. Some stop. Some leave. Some remain anonymous. The capacity excluded from lawful domestic formation does not disappear from the world enclosing the nation. It continues to form in parallel populations and returns across the network as an external pressure.
#The exchange is not lossless for the United States, but it is lossless for the larger technological enclosure. The knowledge still exists. The vulnerability still exists. The incentive still exists. What changes is possession.
#This is why a world without hackers is impossible when the tools used to exclude them are the tools that created them. A connected computer carries instructions, boundaries, errors, and effects. It can be examined because it operates. No domestic deterrence can become a global prohibition upon learning. The strongest local deterrence may therefore produce the cleanest asymmetry: a restricted interior using systems exposed to a less restricted exterior.
#The result begins to resemble a harvest economy.
#Connected systems continually regenerate valuable material: credentials, payment records, medical histories, private communications, industrial plans, computational capacity, privileged access, intelligence, and leverage. Attackers do not need to hold the whole field. They search it for exploitable weaknesses. One successful boundary crossing may pay for thousands of failed attempts. The defender must maintain many boundaries continuously while the attacker needs one sufficiently valuable opening.
#The victim finances the field that regenerates the harvest.
#As more valuable data becomes reachable through more interfaces, the expected yield of a successful breach grows. As the attack surface expands, opportunities for discovery grow with it. Yet breach frequency cannot be reduced to data volume alone. It also depends upon unresolved vulnerability, adversarial search capacity, defensive discovery, incentives, exposure time, and the architecture connecting a successful entry to valuable material.
#Published studies tell us that cyber incidents are not distributed like a clean metronome and that their losses are heavy-tailed. Many events are small. A few are immense. The regularity is not in the timing of each breach. It is in the regeneration of the conditions from which breaches emerge.
#The enclosure produces a stronger proposition:
#Breach frequency should vary with economically attractive attack surface relative to available adversarial defensive capacity, while breach severity should vary with the volume, concentration, and value reachable through a successful crossing.
#Cybersecurity is not merely an industry responding to crime. It is a dynamic equilibrium whose outcomes are partly determined by where knowledge is allowed to accumulate. The CFAA participates in the production of security outcomes before any intrusion occurs. It influences who can learn, which discoveries can safely become domestic knowledge, and which unknowns remain available for an exterior actor to harvest first.
#This warrants formal exploration in a research paper. I find that the context of this book does not provide the rigor the harvest-economy model deserves. Here, it is enough to follow the enclosure until the discovery takes shape. More on this in a formal paper.
#America did not secure its computer science by removing the hacker.
#It could not remove the hacker.
#It selected the enclosure in which the hacker would be most free to learn.
Part V: The Return
Chapter 12: Artificial Intelligence Repeats the Boundary
#Artificial intelligence has arrived inside the same architecture.
#America built much of the foundational hardware, software, research culture, capital structure, and industrial capacity behind contemporary AI. It now confronts systems whose meaningful boundaries cannot be learned solely from intended use. Models must be prompted adversarially, induced to fail, examined for unexpected capability, connected to tools, placed under pressure, and observed as they deviate from the behavior their designers anticipated.
#Once again, the knowledge appears at the boundary.
#Once again, institutions have reasons to restrict access. Models can expose private information, generate harmful instructions, assist intrusion, manipulate users, and act through connected systems. Yet an enclosure that permits only approved questions from approved researchers cannot completely discover what the system makes possible.
#Foreign laboratories, intelligence services, criminal organizations, open communities, and curious individuals will not all accept the same boundary. If Americans are permitted to invent AI but not to explore it with comparable freedom, invention and operational understanding will separate again.
#America may invent the intelligence.
#Someone else will learn what it can do.
#The CFAA did not anticipate artificial intelligence. It does not need to mention AI to project causality into it. AI systems run on computers, expose services, control accounts, call tools, store data, and occupy contractual and technical boundaries. The old authorization division arrives before the new science has finished defining itself.
#Without a counterbalance, the people most capable of understanding AI’s unexpected capabilities may become defendants, employees confined to approved tests, anonymous operators abroad, or adversaries who publish only after the system has failed.
#The answer is not the absence of law.
#It is an enclosure capable of retaining the intelligence of the people whom law encounters.
Chapter 13: The Domestic Red Team
#The most persistent objection arrives disguised as common sense: without felony punishment for unauthorized access, anyone could pick the lock on a bank vault and claim to be testing it.
#The metaphor has already changed the facts.
#Imagine instead that a bank’s front door is standing wide open at midnight. A passerby notices. He steps inside, calls, “Hello? Is anyone here?” He takes nothing, threatens no one, touches no account, and goes home. Then he calls the bank and reports that its door is open.
#The bank may dislike that he crossed its threshold. The state may possess some law concerning trespass. But he did not rob the bank. Robbery requires a robbery. Burglary requires its own elements. Theft requires a taking. The open door does not erase the distinctions among entry, observation, intention, appropriation, damage, threat, and report.
#The CFAA made that absurdity normal in computer systems. A public server can stand open and answer the questions it receives. A citizen can notice the pattern, vary the number in an address, confirm the scale of the exposure, and report the open door. The authorization boundary can nevertheless become the federal count, the felony pathway, and the leverage through which the discoverer is made to resemble the person who emptied the vault.
#This volume does not propose legalizing bank robbery, digital theft, or every act someone may commit after reaching a computer. Those acts already possess names, elements, victims, and laws. It proposes removing the CFAA felony generated by American experimentation itself. The felony belongs to the robbery, not to noticing that the bank forgot to close its door.
#That distinction also undumbs the public idea of strength. The person who notices the open door did not weaken the bank. The door was already open. Reporting the condition converts an unknown exposure into knowledge the bank can use. A system does not become weaker when its flaw is named. It becomes capable of repair.
#The intervention changes one causal projection of sentencing at the authorization boundary where felony exposure originates.
#For an American citizen, experimentation through unauthorized access to a computer system cannot carry felony exposure under the CFAA.
#For a person outside that domestic citizenship enclosure, the CFAA retains its full felony potential.
#That is the intervention. Nested Causality did not derive a misdemeanor scheme, a restitution standard, a safe harbor, a test of good intentions, a reporting duty, or a new classification of aggravating harm. It identified the active enclosure generating the domestic hacker as a felon: the felony status attached to unauthorized access within the nation-state capable of imposing it. The model changes that state for American citizens and leaves the foreign state intact. Other acts do not become lawful by implication; they remain whatever the laws governing those acts already make them.
#The signal must be plain:
#We want the best hackers to be American, and we want them keeping America ahead of its own machines.
#The Domestic Red Team is not a federal unit, reporting infrastructure, registry, bug bounty, credential, or managed pathway into employment. No institution creates it. It already exists wherever Americans test the systems around them, exceed intended use, discover unplanned behavior, and force technology to reveal what it can do.
#It is domestic because its members develop inside the American technological enclosure. It is a red team because its behavior supplies adversarial pressure that system owners would otherwise have to purchase. It is organic because no authority selected its members or approved its trials. It is anonymous because anonymity is how naturally occurring work survives before an institution decides whether to welcome or punish it.
#Red teaming costs money because useful adversarial thought is difficult to produce on command. Organizations hire specialists, define targets, construct test environments, and pay for the disciplined attempt to defeat their assumptions. Outside those contracts, the networked world continuously produces comparable pressure without an invoice. Curious users, hobbyists, students, independent researchers, abrasive critics, and people seeking status or amusement encounter systems as they actually operate. Some of their work is less disciplined. Some of it is more original precisely because it is not enclosed by the assumptions of the organization paying for the test.
#The proposal recognizes that naturally occurring pressure as a national security asset without attempting to own it.
#Anonymity must remain intact. The moment the Domestic Red Team is registered, enrolled, routed, or required to disclose identity, it becomes another authorized institutional program and the organic state is lost. Handles, informal communities, public demonstrations, private warnings, unexpected repairs, and even awkward disclosures may remain among its expressions, but none defines membership. There is no membership.
#The intervention occurs above the team, in law. Removing CFAA felony exposure from American experimentation changes the environment in which organic behavior forms. It allows trial, error, curiosity, and technical prowess to remain domestically available without assigning the foreign learner a structural advantage simply because American law cannot reach him.
#Imagine the better story.
#A teenager reaches a NASA system. While inside, he notices failures and solves problems. Investigators find him. His experimentation cannot make him a felon under the CFAA because he is an American citizen inside the domestic enclosure the intervention is designed to change. The encounter is then free to reveal something the current enclosure suppresses: the system has found a person capable of understanding it.
#The possibility of recognition remains alive because felony identity no longer arrives first.
#The foreign condition is different because it occupies a different national enclosure. A foreign actor or foreign worker remains subject to the CFAA’s full felony potential. The distinction is citizenship because the intervention is intended to change where American technical capacity forms. It is not a general theory of innocence, guilt, loyalty, character, motive, or deservingness.
#The strategic objective can be stated without apology: remove the felony risk from American experimentation, retain the full deterrent against foreign access, and allow the best hackers in the world to become American again.
#The United States has spent forty years perfecting the transition from hacker to defendant.
#It is time to perfect the transition from hacker to defender.
Chapter 14: Return the Computer
#The computer must be returned to the people.
#Not returned as property alone. Americans already own machines whose operating systems, networks, accounts, updates, and data they do not fully control. It must be returned as a field of agency: something a person may learn deeply enough to surprise its manufacturer, expose its weakness, repair its failure, and understand its place in the larger system.
#It should not be criminal to operate a computer with skill and prowess.
#The intervention does not continue from that sentence into a general rewrite of criminal law. It changes the felony projection generated by the CFAA for American experimentation and preserves the statute’s full foreign reach. Nothing else must be invented to complete the model.
#The United States is extraordinarily effective at projecting law inward. That is why a change in the projection matters. A small statutory alteration can travel through prosecution, plea bargaining, education, employment, family expectation, youth culture, research, and national defense. Removing CFAA felony exposure from American citizens’ experimentation through unauthorized access would not command anyone to learn. It would change the outer enclosure in which learning becomes imaginable.
#The resulting liberty is not an absence of structure. It is active capacity held inside a larger national purpose.
#The final intervention is therefore not trust the hackers.
#It is know what the law creates.
#America created computers, networks, cloud systems, and artificial intelligence. It also created a boundary that divided the people using them and assigned criminal identity to unauthorized technical experimentation. It enforced that boundary most completely against those near enough to arrest. Beyond it, rival capacities accumulated experience against the systems Americans were increasingly afraid to touch.
#How did America lose control of the systems it invented?
#It mistook control of the user for control of the machine.
#The machine continued to behave in ways its makers had not anticipated.
#The people with the greatest liberty to follow it learned where it went.
#We do not end the story there, with foreign operators growing more capable at harvesting American value while American institutions struggle to stop them. That is not where this inquiry ends.
#It is where the work begins.
Working Evidence Ledger
#This ledger is part of the developing manuscript and will become a formal bibliography and notes apparatus in the publication edition.
- U.S. Congress, Computer Fraud and Abuse Act of 1986, H.R. 4718, 99th Congress.
- Congressional Research Service, Cybercrime and the Law: Primer on the Computer Fraud and Abuse Act and Related Statutes, R47557, 2023.
- Supreme Court of the United States, Van Buren v. United States, 593 U.S. 374 (2021).
- U.S. Department of Justice, Justice Manual 9-48.000: Computer Fraud and Abuse Act.
- Graves, D. B., “An Empirical Analysis of Sentencing of ‘Access to Information’ Computer Crimes,” Journal of Empirical Legal Studies 20 (2023).
- U.S. Department of Justice, United States v. Aaron Swartz, Superseding Indictment, 2012.
- MIT Information Systems and Technology, Residential Computing MITnet Policies and MITnet Rules of Use.
- U.S. Department of Justice, Justice Department Charges Leaders of Megaupload with Widespread Online Copyright Infringement, 2012.
- High Court of New Zealand, Dotcom v Minister of Justice [2025] NZHC 2634.
- Federal Bureau of Investigation, Update on Sony Investigation, December 19, 2014.
- Executive Office of the President, Economic Report of the President, Chapter 7, 2018.
- Iran Watch, Clipping Tehran’s Wings: How Supply-Side Controls Can Impede the Iranian Drone Program, 2023.
- U.S. Government and contemporary reporting concerning the 2011 RQ-170 incident and President Obama’s request for return.
- U.S. Government Accountability Office, Cybersecurity Workforce: Departments Need to Fully Implement Key Practices, GAO-25-106795, 2025.
- Supreme Court of the United States, Chatrie v. United States proceedings and 2026 opinion concerning geofence location records.
- Congressional Research Service, Geofence and Keyword Searches: Reverse Searches and the Fourth Amendment, R48852, 2026.
- National Institute of Standards and Technology, Identifying and Estimating Cybersecurity Risk for Enterprise Risk Management, NIST IR 8286A Rev. 1, 2025.
- Benjamin Edwards, Steven Hofmeyr, and Stephanie Forrest, “Hype and Heavy Tails: A Closer Look at Data Breaches,” Journal of Cybersecurity 2, no. 1 (2016).
- Sasha Romanosky, “Examining the Costs and Causes of Cyber Incidents,” Journal of Cybersecurity 2, no. 2 (2016).
- Martin Eling, Rustam Ibragimov, and Wei Ning, “Heavy-Tailed Distribution of Cyber-Risks,” Journal of Cybersecurity 9, no. 1 (2023).
- Family and partner of Aaron Swartz, “Official Statement from Family and Partner of Aaron Swartz,” January 12, 2013.
- David Kushner, “We All Got Trolled,” Matter, July 22, 2014.
- Electronic Frontier Foundation, “EFF Opposes Scientology Censorship and Attacks on System Operators,” EFFector 8, no. 2 (1995).
- Gabriella Coleman, Hacker, Hoaxer, Whistleblower, Spy: The Many Faces of Anonymous, Verso, 2014.