United States v. Kunz

(“[A]ny condition that affects a significant liberty interest . . . must be

2 This is consistent with our approach to sentencing more generally. See United States v. Brooks,

889 F.3d 95, 100

(2d Cir. 2018) (“Sentences for violations of supervised release are reviewed under the same standard as for sentencing generally . . . .” (internal quotation marks omitted)). We have long recognized that a remand may be warranted either by a district court’s failure to sufficiently explain its reasoning, see, e.g., United States v. Genao,

869 F.3d 136, 141-42

(2d Cir. 2017), or by its flawed explanation for imposing a given sentence, see, e.g., United States v. Park,

758 F.3d 193, 199

(2d Cir. 2014).

16 imposed by the district court and supported by particularized findings that it

does not constitute a greater deprivation of liberty than reasonably necessary to

accomplish the goals of sentencing.” (internal quotation marks omitted)).

Crucially, though, we have also been careful to qualify that while “the failure to

do so” is generally “error,” we may nonetheless affirm “if the district court’s

reasoning is self-evident in the record.” Betts,

886 F.3d at 202

(internal quotation

marks omitted).

That carve-out looms large here. Kunz is correct that the immediate

explanation offered by the district court in the moment – and reiterated when

pressed by defense counsel – was simply that these were the “conditions that

we’re ordering now.” App’x 60; see also id. at 62 (“It’s the condition that we’ve

been ordering in all these kind[s] of cases.”). Under other circumstances, a

similarly brusque and generalized explanation might indeed amount to

procedural error.

But under the circumstances of this case, the rationale for monitoring

Kunz’s computer access was “self-evident,” Betts,

886 F.3d at 202

(internal

quotation marks omitted), not just from the details of his original child

pornography conviction, but also from his prolific record of supervised release

17 violations since then (including several alleged violations of his computer

monitoring restrictions). This was the sixth time in ten years that the same judge

was tasked with sentencing the same defendant for violating conditions of

supervised release arising from the same underlying conviction. The shorthand

and familiar tone permeating the sentencing transcript reflects that shared

context: all participants understood why they were in the room; all understood

that they had been there many times before; and, for the most part, all agreed

about what needed to be done next. The reimposition of computer monitoring as

a special condition of supervised release, which had been part of the conditions

since the original judgment of conviction in 2005, was thus a foregone conclusion.

The court was under no obligation to recite that context pro forma, or to pick

through every condition and explain, point-by-point, how each was responsive to

the offending conduct. We have never required district judges to perform the

obvious.3 We see no reason to start now.

3 See, e.g., United States v. Forney,

797 F. App’x 31

, 33 (2d Cir. 2019) (summary order) (although the district court did not explain why it imposed a condition requiring defendant to abstain from alcohol, the justification was “self-evident” where the defendant had pled guilty to driving impaired while on pretrial release (internal quotation marks omitted)); United States v. Lopez, No. 21-1450,

2022 WL 1572995

, at *2 (2d Cir. May 19, 2022) (summary order) (similar, with respect to a

18 To the extent Kunz argues that the district court failed to adequately justify

the differences between prior iterations of his computer monitoring restrictions

(which Kunz did not appeal when imposed, and still takes no issue with), and

the purportedly more exacting restrictions ordered this time around, we again

disagree. On appeal, Kunz highlights just three differences between his prior,

apparently unobjected-to, special conditions and the current special conditions:

(1) the language conditioning his internet access on compliance with the CIMP

terms; (2) the requirement that Kunz himself “pay the cost of monitoring

services”; and (3) the provision ordering that, “as much as possible,” Probation’s

monitoring of Kunz’s computer activity be designed to avoid “reading any

privileged information or any private material that is not illegal or reasonably

likely to lead to illegal material or evidence related to illegal activity.”

condition restricting access to services that provide child pornography, because even though the defendant had never purchased child pornography, he had “repeatedly” taken other actions “intending to engage in criminal sexual conduct with minors”); United States v. Deverso, No. 21-2815,

2022 WL 16753115

, at *1 (2d Cir. Nov. 8, 2022) (summary order) (similar, with respect to a condition barring the defendant from viewing pornography where “the record shows that websites containing pornographic content helped to facilitate [the defendant’s] sexual contact with minors”); cf. United States v. Smith,

949 F.3d 60, 66

(2d Cir. 2020) (“[W]e do not require district courts to engage in the utterance of ‘robotic incantations’ when imposing sentences . . . .”).

19 Appellant’s Br. 19-20. Kunz is correct that the district court did not offer any

specific on-the-record justification for those changes.

Conspicuously, though, Kunz has neglected to demonstrate how any of

those changes meaningfully burdened a “significant liberty interest” and

therefore needed to be “supported by particularized findings.” Matta,

(internal quotation marks omitted). With respect to the newly imposed CIMP

terms, while the record does not illuminate precisely how the current monitoring

program compares as a practical matter with Probation’s past monitoring of

Kunz himself, defense counsel conceded at oral argument that prior generations

of the monitoring program enforced against other supervisees in the Western

District of New York included at least some identical restrictions, even as the

program appeared to have evolved in other respects. See, e.g., App’x 122-23,

United States v. Browder,

866 F.3d 504

(2d Cir. 2017) (No. 16-1322) (prohibiting the

supervisee from altering his operating system or using encryption without prior

approval, among other similar restrictions). More importantly, Kunz has not

identified any meaningful difference between his own former computer

monitoring regime and this new one. Any contention that the new CIMP

obligations themselves impose a heightened burden on Kunz’s liberty is routed

20 through his technology-centric arguments. We address those concerns below,

and construe the relevant CIMP terms in a manner that avoids the problematic

implications that might be suggested by an aggressively literal reading of their

text.

The other two cited changes are no more objectionable. As for the new

requirement that Kunz pay monitoring costs, Kunz does not even attempt to

articulate any impact on his liberty. Apart from (literally) highlighting (in yellow)

the relevant language from the judgment as newly imposed provision to which

he objects, his brief devotes no discussion to the issue whatsoever, and the record

is silent as to what, if any, costs Kunz will actually be forced to bear. See City of

New York v. Mickalis Pawn Shop, LLC,

645 F.3d 114, 137

(2d Cir. 2011) (“We

ordinarily deem an argument to be forfeited where it has not been sufficiently

argued in the briefs . . . .” (internal quotation marks omitted)). While we can

imagine scenarios where such a condition would impose a sufficiently onerous

burden on a supervisee to require a specific on-the-record justification, in the

absence of any showing that such a burden exists here, the self-evident rationale

– shifting reasonable monitoring costs to the person whose crime made the

monitoring necessary – provides sufficient justification. Similarly, Kunz does not

explain why altering his special conditions to expressly prohibit Probation from

21 indiscriminately reviewing his privileged and private information somehow

aggravates, rather than reduces, the burden on his liberty.

Moreover, even were Kunz’s liberty meaningfully affected in some way by

the changes, it is self-evident from the record that in adopting all of Probation’s

proposed updates to Kunz’s special conditions, the district court generally

credited Gilbert’s recommendation that Probation’s modern monitoring

approach would be more effective than prior iterations of the monitoring

program at aiding the rehabilitation of a defendant who was originally sentenced

during the relative infancy of computer monitoring, see United States v. Sofsky,

287 F.3d 122, 126

(2d Cir. 2002) (collecting early computer monitoring cases from the

late 1990s and early 2000s), who had violated his supervised release many times

since then, and who since his most recent sentencing had now offended yet

again.

In sum, although “the defendant, the public, and appellate courts should

not be required to engage in guesswork about the rationale for a particular

sentence,” Genao,

869 F.3d at 142

, it requires no “guesswork” to understand why

the district court imposed the conditions it imposed in this case. Given the nature

both of Kunz’s underlying offense and of his repeated violations of supervised

release, the longstanding general computer monitoring requirement that Kunz

22 has never (including now) objected to, and Gilbert’s justification for the

recommended updates, the district court’s rationale for imposing this sort of

computer monitoring program was apparent on this record. We thus discern no

procedural error in the sentence it imposed.

B. Substantive Reasonableness: Technological Issues

Kunz’s substantive reasonableness challenge centers on his contention that

Probation’s computer monitoring requirements are technically vague or

unworkable. We are not convinced.

Kunz points us to the portion of his digital forensics expert’s declaration

describing various technical predicaments stemming from the expert’s

interpretation of the CIMP terms. The particular terms at issue include

restrictions requiring Kunz to notify Probation of “any alterations to

computers(s)/connected device(s) and/or passwords/screen names prior to

executing [the] change” and to obtain Probation’s permission prior to “alter[ing]”

any software or operating systems. App’x 48-49. According to Kunz’s expert,

those restrictions “would be difficult if not impossible for a user to comply with

as they are currently worded.” App’x 52. For example, many programs and

operating systems update automatically without the user’s intervention or

23 knowledge, and such “alterations to [Kunz’s] computer . . . software” thus cannot

realistically be cleared in advance. App’x 44. Similarly, passwords for many

services expire automatically, periodically forcing users of those services to

immediately update their credentials upon login, making it impractical for users

to seek prior approval. Finally, a broad reading of the provision barring the use

of “any type of encryption” without prior approval would preclude Kunz from

using a machine with preloaded standard encryption, or any of the countless

websites with URLs that begin with “https,” because that prefix indicates that the

website encrypts information as it is transferred to or from the user.

Were we required to read the CIMP terms as broadly as Kunz’s expert

does, we might be persuaded. Fortunately, we are not. Those terms are obviously

and reasonably intended to preclude a defendant’s active attempts to avoid

monitoring, not to punish a passive defendant for innocuous day-to-day

activities that trigger changes initiated by forces outside of his control.

Importantly, the CIMP terms themselves suggest that they should be read

that way. All of the challenged provisions are phrased in terms of what “I”

(Kunz, the supervisee) will or will not actively do. Thus, for example, Kunz is

required to pledge that he shall “notify the U.S. Probation Office” of certain

24 changes before “executing” those changes. App’x 48 (CIMP ¶ 7). On its own

terms, that language does not require Kunz to notify Probation of automated

changes initiated by a vendor or developer, which are not “execut[ed]” by Kunz

himself.

Similarly, it is Kunz who is required to pre-clear “alterations” to his

software. Id. at 49 (CIMP ¶ 16). By personalizing that obligation to Kunz, the

CIMP language gives no indication that Probation somehow expects Kunz to

perform “the impossible,” United States v. Johnson,

446 F.3d 272, 281

(2d Cir.

2006), either by predicting the unpredictable, or by otherwise assuming

responsibility for automatic, unannounced, third-party updates that he plays no

active or intentional role in initiating. And for those updates that do require a

user to take some active, intentional step to trigger the change, nothing prevents

Kunz from notifying his probation officer before taking that step. Likewise,

nothing in the CIMP terms precludes Kunz from seeking, or the probation officer

from granting, blanket permission for Kunz to install standard program updates

initiated by the likes of Microsoft or Apple.

In sum, even a stubbornly literal reading of the challenged CIMP terms

does not require the interpretation offered by Kunz’s expert. We thus construe

the requirements to prohibit Kunz from initiating without prior permission any of

25 the above processes or changes;4 they do not, nor do we think they could sensibly

be read to, compel him to seek prior approval for processes entirely outside of his

control and of which he himself has no advance notice. See United States v. Young,

910 F.3d 665, 671-72

(2d Cir. 2018) (construing a condition ordering mental health

treatment as excluding “inpatient treatment,” notwithstanding that a literal

reading might have extended that far, thereby mooting any concern that the

district court had failed to adequately justify that “restrictive” treatment type).

Nor do we believe that a sensible reading of the conditions requires Kunz

to seek prior approval each time he “use[s],”5 App’x 49 (CIMP ¶ 18), common

4 At oral argument, Kunz’s counsel offered that the spirit of these restrictions could be captured by language providing that Kunz “shall not attempt to circumvent the monitoring software and/or hardware in any way,” followed by a non-exhaustive illustrative list closely mapping the current CIMP terms. Oral Argument 6:00. We believe our construction accomplishes the same thing while avoiding the potentially imprecise implications of the word “circumvent.” Under our reading, it does not matter what Kunz’s subjective goal may be in initiating the barred process – be it to deliberately evade his monitoring restrictions or not. What matters is only whether it is Kunz himself who actively initiates (or causes to be initiated) the barred process, or whether Kunz passively experiences the effects of a process caused by some force entirely beyond his control. 5 Once again, we are guided by the language in the CIMP term personalizing the obligation to Kunz, and attaching an active verb – “use” – to that obligation. While one might plausibly suggest that a computer user who purchases household items online through a secure payment interface in some sense “use[s] . . . encryption,” App’x 49 (CIMP ¶ 18), when the vendor encrypts the

26 electronic resources or services that happen to employ encryption in a manner

that does not frustrate the purposes of Probation’s monitoring. We think it would

be absurd to interpret Probation as purporting to require Kunz to ask permission

every time he attempts to access a website whose URL happens to begin with the

characters “https” (signifying that the website “encrypts the data as it is

transferred,” App’x 54) or every time he uses a messaging service that encrypts

messages while they are en route from sender to recipient,6 where that encryption

transmitted financial information to protect it from hackers, see Mark A. Lemley, The Splinternet, 70 DUKE L.J. 1397, 1423 (2021), such a reading is certainly not compelled by that language. Nor is that even remotely the most natural reading; few ordinary, non-expert computer users would ever liken behavior of that sort to “us[ing]” encryption. Cf. United States v. Carlineo,

998 F.3d 533, 536

(2d Cir. 2021) (supervised release restrictions must “put an ordinary person on notice of the prohibited conduct” (emphasis added)). Therefore, construing the provision in light of its evident purpose, as discussed throughout this section, we think a fair reading of the CIMP’s encryption restriction does not encompass such “use” that is effected by the vendor, invisible to the user, and inconsequential to Probation’s ability to monitor its supervisees’ purchasing activity. 6 The record is silent as to whether any of the increasingly common encrypted messaging services that employ “end-to-end encryption” or other “technical impediments to law enforcement” may stymie Probation’s monitoring efforts here. See Alan Z. Rozenshtein, Surveillance Intermediaries, 70 STAN. L. REV. 99, 106 (2018). In our view, Probation’s legitimate interest in restricting the use of such services extends only to the point where the service’s encryption interferes with Probation’s monitoring capabilities. At any rate, the pre-approval process provides Probation with the capacity to evaluate the risks of particular types of services.

27 does not thwart Probation’s ability to keep tabs on Kunz (via, for example, the

monitoring software it has installed on Kunz’s devices). The clear intention of the

prohibition is to prevent Kunz from hiding behind the veil of encryption any

efforts to, for example, access child pornography, make forbidden contact with

minors, or engage in any other illegal activity or violations of his conditions of

supervised release. To that end, so long as no encryption is initiated by or at the

behest of Kunz, and so long as any encryption initiated by external forces does

not meaningfully impair Probation’s ability to monitor Kunz’s behavior

consistent with both the special conditions and the other CIMP terms, we

construe the encryption restrictions to allow the reasonable use of services and

resources that happen to employ encryption in some way.

Interpreting the contested CIMP terms this way should assuage any

technical concerns Kunz harbors, while also preserving both the spirit and the

efficacy of the restrictions. Interpreting them otherwise, by contrast, would run

them headlong into our requirements that such restrictions provide an ordinary

person with “clear notice”7 of whether a given action is prohibited or not, and not

7 To the extent Kunz’s technical vagueness theory doubles as a due process vagueness challenge, we similarly hold that the restrictions, as we construe them, are “sufficiently clear to give the person of ordinary intelligence a reasonable

28 “demand the impossible” of a supervisee. Johnson,

446 F.3d at 281

; see United

States v. Carlineo,

998 F.3d 533, 536

(2d Cir. 2021).8

Construed in that way, we find no procedural or substantive

unreasonableness in Kunz’s computer monitoring obligations. We thus turn next

to his improper delegation arguments.

opportunity to know what is prohibited, so that he may act accordingly.” United States v. Simmons,

343 F.3d 72, 81

(2d Cir. 2003) (internal quotation marks omitted); cf. United States v. Burdick,

789 F. App'x 886

, 888-89 (2d Cir. 2019) (summary order) (holding that the defendant’s overbreadth and vagueness challenges to his CIMP terms were “unripe,” emphasizing that “[t]he district court retains the discretion to remedy a potential ambiguity in the language of a special condition”). 8 Separately, Kunz’s expert also declared that the CIMP term barring use of any service that “conceals or spoofs” his IP address could make life difficult for any user who works for a company that requires the use of a VPN, because that sort of connection would technically run afoul of that restriction. App’x 52, 54. Here, we need not construe to avoid anything. This is a hypothetical worry that, unlike the other seemingly more intrusive CIMP restrictions, would require a simple modification request should a change in Kunz’s employment circumstances warrant as much. Cf. United States v. Balon,

384 F.3d 38, 47

(2d Cir. 2004) (“[C]hanging computer technology is an appropriate factor to authorize a modification of supervised release conditions under Section 3583(e).”). That is not an unworkable constraint; it is merely a less convenient one than Kunz would understandably prefer.

29 III. Delegation of Judicial Authority

Kunz argues that the district court impermissibly delegated its authority to

Probation in two ways: (1) by endorsing CIMP terms that, as currently written,

leave too much to Probation’s discretion; and (2) by leaving the door ajar for

Probation to unilaterally alter Kunz’s CIMP terms in the future. Although we are

sensitive to Kunz’s concerns, we once again construe the language he

underscores in a way that should, with one exception, obviate those concerns.

“The power to impose special conditions of supervised release . . . is vested

exclusively in the district court.” Matta,

777 F.3d at 122

. By contrast, “the

extensive supervision mission of federal probation officers includes executing the

sentence but not imposing it.”

(internal citation, quotation marks, and

alteration omitted). Thus, we have explained, while a district court may delegate

to Probation authority “over certain minor details of supervised release,” it may

not delegate “decisionmaking authority which would make a defendant’s liberty

itself contingent on a probation officer’s exercise of discretion.” United States v.

Birkedahl,

973 F.3d 49, 54

(2d Cir. 2020) (internal quotation marks omitted); see,

e.g., Matta,

(“[T]he discretion to require either inpatient or

outpatient drug treatment was an impermissible delegation of judicial sentencing

30 authority.”); United States v. Peterson,

248 F.3d 79, 85

(2d Cir. 2001) (“If [the

defendant] is required to participate in a mental health intervention only if

directed to do so by his probation officer, then this special condition constitutes

an impermissible delegation of judicial authority to the probation officer.”).

While these formulas are easily recited, they are less easily applied; they

require that we survey many gray areas and draw many subtle lines. When a

sentencing court orders a special condition of supervised release imposing a

particular form of treatment or supervision (such as “mental health treatment” or

“computer monitoring”), that generic term encompasses many possible

variations and operational components. Of course, all of those options concern, in

the broadest sense of the word, the “details” of the required condition, and all

affect, to some degree, the supervisee’s liberty. But assessing whether the details

left to Probation’s discretion involve significant impositions on the supervisee’s

liberty (like the choice between inpatient and outpatient drug treatment, see

Matta,

) or matters of “minor” detail (such as the choice of a

particular treatment provider, see Peterson,

248 F.3d at 85

) demands a careful and

contextual analysis of both the nature of the overall condition and the relative

importance of the specific choice to the broader intrusion upon the supervisee’s

liberty.

31 A. The CIMP Terms as Currently Written

For the most part, Kunz’s first delegation challenge, which targets the

CIMP terms as currently conceived, presents no great difficulties. He urges that

by rubber-stamping CIMP terms that “demand [Probation’s] prior approval

before . . . even commonplace and automatic computer activity” the district court

impermissibly delegated its judicial authority. Appellant’s Br. 27. But that theory

is really just a rejiggering of Kunz’s earlier contention that the CIMP terms are

vague and unworkable. Having already construed them in a manner that

neutralizes Kunz’s reasonable concerns, we disagree that the district court’s

decision to condition his internet access on compliance with those terms

constituted an impermissible delegation. Naturally, the various CIMP provisions

that permit certain computer activity only with Probation’s approval delegate to

the probation officer authority to approve or disapprove those activities. But in

the context of the broader monitoring regime ordered by the court (again,

without objection), those decisions are properly understood as matters of detail.

The sentencing judge cannot reasonably be expected to compile a detailed list of

the kinds of software updates or password changes or “use[s]” of encryption that

do or do not interfere with Probation’s ability to monitor a supervisee’s activities

32 to detect illicit conduct. Once those provisions are understood in light of that

goal, Probation is equipped with a standard of decision that constrains its

discretion and avoids arbitrary use of its delegated authority.

Kunz raises a different issue, however, that presents a more difficult

question. He argues, and we agree in part, that certain of the CIMP terms

contradict the express language of the special conditions, and therefore exceed

the authority actually delegated by the district court. To be sure, not all of the

language Kunz characterizes as contradictory is actually anything of the sort. For

example, he insists that the CIMP term requiring him to “waive any expectations

of privacy from the supervising probation officer, his or her designee, and the

monitoring company,” App’x 48 (CIMP ¶ 1), contradicts the special condition

providing that monitoring and examination of his devices “shall be designed to

avoid, as much as possible, reading any privileged information or any private

material that is not illegal or reasonably likely to lead to illegal material or

evidence related to illegal activity,” App’x 73. But we read the two together to

mean simply that Kunz agrees that he is entitled to no greater expectation of

privacy than that which is implied by the special condition. Although they tread

similar ground, those two provisions target different things. The special

33 condition’s text constrains the supervisors in their surveillance of Kunz by

requiring them not to abuse their power and access; the CIMP provision instructs

Kunz that he has no standalone expectation of privacy beyond that. Together, the

two provisions convey to Kunz that he can expect no greater privacy from his

supervisors than what is guaranteed to him by the special condition. Though

they may diverge in focus and in thrust, they are not contradictory, and pose no

delegation problems.

Kunz’s objection to the CIMP term providing that he “may be limited to . . .

one Internet-capable device,” however, has merit. App’x 48 (CIMP ¶ 5). That

restriction, he argues, cannot be reconciled with the district court’s use of the

“(s)” device at the end of terms like “computer(s), automated service(s), or

connected device(s),” apparently allowing for the possibility that each of those

nouns could be plural. App’x 73. Though we are skeptical that the two provisions

are truly contradictory – Kunz’s formulation of the argument ignores the fact that

the “(s)” also contemplates the possibility that each noun could be singular – we

share the larger concerns lurking within his argument.

We think a restriction limiting a supervisee to just one internet-connected

device would pose a significant burden on his liberty, and therefore would need

34 to be imposed by the court and justified by particularized on-the-record findings.

See Matta,

. Even setting aside the proliferation of the “Internet of

Things,” which has exploded the category of devices that could qualify as

internet-connected, see Chris Jay Hoofnagle et al., The Tethered Economy, 87 GEO.

WASH. L. REV. 783, 785 (2019), such a restriction would force Kunz to choose

between using a computer and using a smart phone, both of which are

“indispensable to participation in modern society.” Eaglin,

913 F.3d at 98

, quoting

Carpenter v. United States,

138 S. Ct. 2206, 2220

(2018) (discussing cell phones); see

Peterson,

248 F.3d at 83

(same for computers). We do not suggest that such a

severe restraint on internet access could never be warranted, but rather that it

would require particularized justification by the court. And for the same reason,

any special condition granting Probation discretion to decide whether or not to

restrict a supervisee to a single internet-connected device would constitute an

impermissible delegation of the court’s judicial authority. See Matta,

777 F.3d at 122-23

Fortunately, the plain language of the court’s special conditions here

avoids those traps. Neither relevant passage purports to restrict Kunz to one

35 internet-connected device, or to delegate to Probation the authority to determine

how many internet-connected devices Kunz may use. The first relevant passage –

requiring Kunz to give Probation “advance notification of any computer(s),

automated service(s), or connected device(s) that will be used,” and authorizing

Probation to install surveillance software on those “computer(s) or connected

device(s)” – simply orders Kunz to proactively notify Probation of any device, or

devices, that he intends to use while under supervision so that they may be

outfitted with surveillance software. App’x 73. Similarly, the second passage –

requiring Kunz to cooperate with Probation’s examinations of any “computer(s),

connected device(s), storage media, and any internal or external peripherals” –

simply cautions that any device, or devices, that Kunz intends to use will be

subject to unannounced searches.

In both instances, the court’s language

implies that it is Kunz himself, not Probation, who is empowered to decide how

many internet-connected devices he wishes to use. We therefore discern no error

in the district court’s special condition.

But that still leaves the language in the CIMP terms that purports to

reserve for Probation the right to restrict Kunz to a single device. We conclude

that this term exceeds the authority actually delegated to Probation by the district

36 court, and therefore may not be enforced as written. Thus, although we affirm the

judgment of the district court, we do so with the stipulation that Probation may

not enforce any provision of the CIMP that purports to authorize the probation

officer to limit Kunz to one internet-connected device. See United States v.

Villafane-Lozada,

973 F.3d 147, 153

(2d Cir. 2020) (explaining that where Probation

“does overstep [its] authority,” the sentencing court may provide relief). Such a

harsh restriction would need to be specifically imposed by the district court and

justified by an express, case-specific rationale for imposing it. See Matta,

. None of that happened here.

B. The Monitoring Policy “In Effect” by Probation

Kunz’s next, and trickier, delegation challenge centers on language in the

district court’s special conditions imposing restrictions “consistent with the

computer monitoring policy in effect by the probation office.” App’x 73.

As Kunz observes, when read literally, the “policy in effect” phrase

appears to vest Probation with the power to unilaterally update his CIMP terms –

a reading Kunz contends would amount to an impermissible delegation of the

district court’s judicial authority. In essence, he argues, because the district court

37 reviewed and endorsed these particular CIMP terms at sentencing, surely these

particular terms must be frozen in time, subject to tinkering by the court alone.

Appellant’s Br. 28-29. The government counters by proclaiming without

qualification that the terms of Kunz’s computer monitoring are “minor details,”

apparently as a matter of law, and therefore that their modification simply would

not “implicate the judicial decision-making duty.” Appellee’s Br. 10. We do not

altogether agree with either party.

We begin by acknowledging the challenges inherent in answering this

question prospectively, before (as far as the present record reveals) Probation has

actually endeavored to alter the CIMP terms. On the one hand, Kunz’s reasonable

concern seems to be that by greenlighting unilateral CIMP changes, the district

court has empowered Probation to make extrajudicial updates that may

meaningfully restrict his liberty. That would indeed be problematic. On the other

hand, the current record grants no insight into what those future changes might

conceivably be, much less that any have actually occurred or even been

contemplated. That makes it difficult to determine whether those undefined

hypothetical changes would be better classified as the sort of minor details

typically within Probation’s purview, or the sort of liberty-burdening changes

38 that pose delegation problems.

For that reason, at first glance, this case presents something of a ripeness

problem. The ripeness requirement “prevents a federal court from entangling

itself in abstract disagreements over matters that are premature for review

because the injury is merely speculative and may never occur.” United States v.

Traficante,

966 F.3d 99, 106

(2d Cir. 2020) (internal quotation marks omitted)

(adding that because the doctrine derives from Article III jurisdictional limits, a

court may raise the matter sua sponte and for the first time on appeal).

We have on occasion held that challenges to a defendant’s conditions of

supervised release fail that requirement. For example, in United States v. Balon, we

determined that a challenge to a condition subjecting a defendant convicted of a

child pornography offense (whose supervised release would not begin for several

years) to remote monitoring of his electronic communications was unripe.

384 F.3d 38, 46-47

(2d Cir. 2004). Our opinion in that case grappled with themes that

resonate with this one: “The technology that holds the key to whether the special

condition in this case involves a greater deprivation of liberty than reasonably

necessary is constantly and rapidly changing,” and it was therefore “impossible”

to predict whether whatever technological tools Probation would be employing

39 in the future would “unnecessarily deprive[]” that defendant of his liberty.

Id. at 46

(adding that “the issue here is distinctly a matter of fact beyond the prescience

of this court”).9

What separates this appeal from those cases is that Kunz is not challenging

a restriction that may or may not be imposed based on unspecified developments

at an unspecified time in the future. He is challenging a specific condition,

imposed right now, applicable right now, and seeming to delegate authority right

now; and it is only that presently delegated authority which can be read to

empower Probation to make indeterminate future changes for indeterminate

future reasons.

1. Already-Delegated Authority

Insofar as Kunz challenges power that the district court has already

9 See also Traficante,

966 F.3d at 104, 106

(deeming unripe a vagueness challenge to a condition of supervised release that permitted Probation to compel the defendant to notify others “[i]f the court determines in consultation with your probation officer that, based on your criminal record, personal history and characteristics, and the nature and circumstances of your offense, you pose a risk” to third parties, reasoning that the question was “just an abstraction” as applied to a defendant who had not yet been found to pose such a risk); United States v. Rasheed,

981 F.3d 187, 200

(2d Cir. 2020) (similar); United States v. Bryant,

976 F.3d 165, 182

(2d Cir. 2020) (similar).

40 delegated to Probation – here, the discretion to unilaterally update the CIMP

terms in the future – rather than whatever Probation might do with that power in

the future, his challenge is ripe. We said as much in Villafane-Lozada,

973 F.3d at 149

. There, the defendant challenged a condition of supervised release requiring

him to “submit to a polygraph, computerized voice stress analyzer, or any other

such testing,” arguing that this condition impermissibly delegated to Probation

“the choice of which verification testing technology to employ, whether it be

polygraph, computerized voice stress analysis, or something else entirely.”

Id. at 150-52

. We explained that the issue was ripe for review because it constituted a

challenge to “the already realized delegation of judicial power to a probation

officer, not some hypothetical decision that this delegation might allow in the

future.“

(emphasis in original). And we went a step further: we distinguished

that (ripe) delegation dispute from the defendant’s separate (unripe) attack on

the use of future computerized voice stress analysis, which he argued had not

been “scientifically proven” to be reliable.

Id. at 151

. That second issue, unlike the

immediate delegation of authority, hinged on an uncertain question of future

technology – that is, “what the state of computerized voice stress analysis

41 technology will be when [defendant] begins his term of supervision in six

years.”

Both sides of that coin are relevant here. As in Villafane-Lozada, Kunz

argues that the district court in this case has already delegated the challenged

authority to Probation. “In other words, it is the probation officer’s

already-granted authority” to unilaterally update the CIMP terms, not

necessarily “the particular [term] that will be chosen,” that Kunz argues is

unlawful. Villafane-Lozada,

973 F.3d at 151

. And because that purportedly

improper delegation has already occurred, Kunz’s challenge to that aspect of his

sentence is ripe.

2. Construing the Special Condition

But, also as in Villafane-Lozada, the shadow of uncertain technological

change still clouds this case. On one side, the most reasonable rationale for

Probation’s apparent wish to retain discretion to update the CIMP terms is

because (as Kunz’s examination of the current CIMP’s potential technical snags

itself illuminates) both the monitoring technology Probation employs and the

means defendants may summon to evade that technology are fickle and

42 evolving. Probation has a legitimate interest in retaining the flexibility to respond

to emergent technology free from the administrative headache of seeking

individual, judicially approved modifications of conditions for every supervisee

under its care whenever a new response to a new means of evading supervision

is invented. On the other side, we also recognize Kunz’s interest in remaining “on

notice as to what conduct could trigger a charge of violating the condition,”

Carlineo,

998 F.3d at 537

, and in serving out his term of supervised release free

from the specter of extrajudicial overreach further restricting his liberty – in

addition to the burden he and other supervisees would have to bear should the

onus to litigate CIMP changes fall to them on the back end, rather than to the

government on the front end.

This tug-of-war is not entirely novel. The precise language Kunz complains

of – “consistent with the computer monitoring policy in effect by the probation

office,” App’x 73 – appears in special conditions imposed in scores of cases across

the Western and Southern Districts of New York over recent years.10 It also

10 See, e.g., United States v. Dantz, No. 1:09-CR-00146 (W.D.N.Y. Sep. 28, 2009), Dkt. No. 17 (judgment); United States v. Deak, No. 1:15-CR-00659-JSR-1 (S.D.N.Y. Aug. 30, 2016), Dkt. No. 28 (judgment); see also, e.g., United States v. Skvarla, No. 6:09-MJ-00546-JWF-1 (W.D.N.Y. Mar. 27, 2009), Dkt. No. 2 (imposing this condition for pretrial supervision).

43 featured in previous iterations of Kunz’s own special conditions, none of which

he appealed.

On at least one occasion, we have upheld by summary order the exact

same (verbatim) special condition in the face of the very argument Kunz now

presents. See United States v. Vietor,

806 F. App’x 60

, 63 (2d Cir. 2020) (summary

order); Appellant’s Br. 7, Vietor,

806 F. App’x 60

(No. 19-1315) (“That delegation

of authority presumes that the Probation Office would have the authority to

revise its computer monitoring program on its own, as it sees fit.”). In our

summary order, we at least implicitly rejected that argument, though not

necessarily embracing the defendant’s framing of the issue. See Vietor, 806 F.

App’x at 63 (“While the condition at issue here delegates authority to the

probation office to select and administer the specific monitoring program, within

the condition’s confines, it is the district court, not the probation office, restricting

[the defendant’s] internet access.”). And because the defendant in Vietor had not

raised the delegation argument before the district court, we reviewed the

judgment for plain error only, id. at 62, further dampening that already

44 nonbinding summary order’s persuasiveness here.11

On at least one other occasion, we have upheld effectively identical

language by construing it to avoid the temporal question now squarely before us.

11 We have also on many occasions affirmed criminal judgments from the Western, Northern, and Southern Districts of New York that included similar (or indeed often the same) language in cases where the defendants did not directly challenge that language on appeal. The Vietor panel itself observed that we had endorsed “similarly worded condition[s]” before. 806 F. App’x at 62-63. It cited our summary order in United States v. Savastio, where we (again applying plain error review) upheld special conditions of supervised release imposed by a district court in the Northern District of New York prohibiting a defendant from “us[ing] or possess[ing] any computer, data storage device, or any internet capable device unless you participate in [computer monitoring]” and requiring that he “comply with all of the rules of the program and pay the costs associated with the program.”

777 F. App’x 4

, 7 (2d Cir. 2019). The special conditions in that case were also worded in a way that seemed to allow Probation to unilaterally alter the operative CIMP terms in the future: “Your internet use will be limited and/or restricted under conditions to be set by the U.S. Probation Office in accordance with their Computer and Internet Monitoring Program.”

Id. at 5

(emphasis added). Similarly, in United States v. Leone,

813 F. App’x 665

(2d Cir. 2020) (summary order), we affirmed a judgment containing a substantially identical special condition to Kunz’s; there, though, the defendant’s delegation argument was limited to the purported vagueness of certain computer restrictions as currently written. See Appellant’s Br. 19-20, Leone,

813 F. App’x 665

(No. 19-1670). And those cases have plenty of company. See, e.g., United States v. Swartz,

459 F. App’x 47, 48

(2d Cir. 2012) (affirming a judgment containing the exact language Kunz challenges, where that language was not at issue on appeal); United States v. Petix,

767 F. App’x 119

, 123 (2d Cir. 2019) (same); United States v. Asch,

775 F. App’x 15

, 19 (2d Cir. 2019); United States v. DeCapua,

822 F. App’x 16

, 18 (2d Cir. 2020).

45 Our opinion in United States v. Browder, which also concerned a defendant

convicted of a child pornography offense, centered on an effectively identical

special condition providing that the defendant’s “computer or computers will be

subject to monitoring by the U.S. Probation Office, consistent with the computer

monitoring policy then in effect by the probation office.”

866 F.3d at 509

(emphasis

added). We immediately flagged from that language the prospect that the

condition “may be subject to challenge as an impermissible delegation of judicial

authority to the Probation Office.”

Id. at 510

. And although we ultimately upheld

the condition, we did so with two conspicuous asterisks. First, because the

defendant did not challenge the condition as an impermissible delegation, we

concluded he had “waived” that issue on appeal.

And second, for that reason,

we construed the “then in effect” language to mean “whatever computer

monitoring policy was used by the Western District’s Probation Office at [the

defendant’s] release.”

(emphasis added).

In this case, the issue is assuredly not waived, and we therefore are not

bound by the Browder panel’s approach. Nonetheless, we find that approach

instructive, and adapt it to reach the outcome that we think best balances the

interests and constraints we have laid out in this opinion.

46 Here, we construe the language at issue as permitting Probation to make

future unilateral changes to the terms of Kunz’s computer monitoring, much the

same as it may make certain unilateral changes to other aspects of his supervised

release, but only to the extent that those changes are the sort of “minor details of

supervised release” already within Probation’s purview. Birkedahl,

973 F.3d at 54

While that construction may appear circular, we think it the most sensible

reading of the “consistent with the computer monitoring policy in effect by the

probation office” language at the core of this dispute.12 Cf. Villafane-Lozada,

(avoiding an impermissible delegation obstacle by construing a

condition requiring periodic truth-verification testing via a seemingly open-

ended list of potential methods as nonetheless precluding “the use of a

verification test that is materially more restrictive on [the defendant’s] liberty

12 It also echoes the assumption implicit in our ripeness determination in Balon: that the Probation Department would select the particular means of computer monitoring it deemed appropriate at the time of the defendant’s release. See

384 F.3d at 46-47

. While recognizing that a challenge to the restrictiveness of the chosen method would ripen at that time, we did not question the propriety of effectively delegating the choice of technology to Probation in the first instance, rather than reserving the selection to the court. See

id.

Allowing Probation to select the means of monitoring at the future commencement of supervision is not meaningfully different from allowing it to adapt to new technology by updating the means it initially selected.

47 than” methods the court had expressly authorized); United States v. Corbett,

767 F. App’x 191

, 192-93 (2d Cir. 2019) (summary order) (similar, construing a

seemingly open-ended condition ordering “anger management and any

additional mental health treatment that the defendant should require,” which the

district court had expressly linked to “the issues that were identified by [a

doctor’s psychological evaluation],” as authorizing Probation “to ensure that [the

defendant] receives mental health treatment to address only those issues

identified” in his psychological evaluation).

Surely, Kunz is correct that no one in his position can rightfully be asked to

bear the risk of Probation imposing “significantly greater restrictions on [his]

liberty” on its own, as a literal reading of the relevant language might suggest is

possible. Carlineo,

998 F.3d at 537-38

. That power is “reserved exclusively for the

district court.”

We are thus precluded from upholding any construction of the

complained-of language suggesting that the district court delegated any more

than our reading of its language suggests. See Villafane-Lozada,

(reasoning that the special condition at issue “cannot reasonably be construed” to

subject the defendant’s liberty “to the whims of his supervising probation

officer”).

48 Nor would it be reasonable to conclude that the district court delegated

any less than our reading suggests. It certainly does not appear to have intended

to do so; Kunz voiced his objection to the “policy in effect” language several

times below, imploring that the court consider “at least removing the sentence

that would incorporate by reference any future new programs with new terms

that probation comes up with in the future.” App’x 61. The court refused.

More importantly, Kunz – who, unlike the defendant in Browder, has

actually raised the issue on appeal – has failed to identify any meaningful

principle that can distinguish computer monitoring from other species of

supervised release in this respect. We can discern no reason why sufficiently

“minor” adjustments to his CIMP terms would be subject to fundamentally

different rules than, say, the “start date and nightly duration” of a court-ordered

curfew, United States v. Degroate,

940 F.3d 167, 177

(2d Cir. 2019), “the selection of

a therapy provider and schedule” for court-ordered therapy, Peterson,

248 F.3d at 85

, or the precise “type of testing” chosen to carry out court-ordered truth-

verification testing, Villafane-Lozada,

(emphasis omitted).13 We

13 To the extent someone in Kunz’s position might suggest that what distinguishes the CIMP terms from other aspects of his supervised release is that they are expressly spelled out in a writing, styled here as an “Agreement,” Kunz

49 have held that Probation is permitted to make those kinds of changes

unilaterally, and we are bound by those holdings here. Thus, we are left with no

conclusion to draw other than that the district court meant what it said and said

what it meant – that Probation may unilaterally alter the minor details of Kunz’s

computer monitoring – and that what it meant is permissible under our

precedents.

3. Future CIMP Updates

All of that notwithstanding, however, we appreciate that Kunz’s

misgivings may prove prescient. We expect that many future changes Probation

might be inclined to make (like tweaks to its monitoring software) would be

did not argue anything of the sort in his brief, and mustered only a single unsupported, conclusory sentence to that effect at oral argument. See Oral Argument 18:50 (“The problem is by codifying, they’ve frozen these . . . provisions.”). Thus, that argument has not been adequately presented to us and is therefore forfeited. See Browder,

866 F.3d at 510

; Mickalis Pawn Shop,

645 F.3d at 137

. In any event, we are reticent to use the fact that Probation has memorialized its policies in a writing as a hook to strip its authority over the kinds of minor details otherwise within its purview. In particular, we are loathe to create any further disincentive for Probation to furnish those under its care with detailed, written notice of its expectations, lest we exacerbate the risk that future supervisees will be left to fend without a “sufficiently clear” understanding of exactly what is expected of them. Carlineo,

998 F.3d at 536

. As Kunz’s own expert submission forebodes, that risk is particularly acute in the context of highly technical computer-based restrictions.

50 entirely appropriate, but it may well also be that other future changes (such as

attempting to limit Kunz to one internet-connected device) would cross the

threshold from minor detail into major imposition on Kunz’s liberty. Matta,

777 F.3d at 122

. That is always a risk. Probation is entrusted to act unilaterally in

many contexts, and in each of those many contexts there is some chance that it

will outpace its discretion to “execut[e]” the court’s sentence and instead purport

to impermissibly “impos[e]” its own.

(internal quotation marks omitted); see

Villafane-Lozada,

(“[I]t is always possible that a probation officer

might find some way to abuse his delegated authority.”); see also, e.g., Degroate,

940 F.3d at 177

(acknowledging that Probation may have “exceeded its delegated

authority” to set the start date and duration of the defendant’s curfew “by

imposing a total ‘lock-down’”). We take that risk seriously in all contexts, but our

precedents provide no basis for us to treat computer monitoring differently from

other classes of generalized conditions whose details we permit Probation to

manage.

That holds true even as we recognize that as a practical matter, computer

monitoring may prove more fraught than other areas of supervised release, if for

no other reason than the sheer ubiquity of internet-connected devices and the

51 “nearly essential” nature of internet access to many aspects of modern life. Eaglin,

913 F.3d at 96

. Against that backdrop, some future incremental computer

monitoring adjustment may perhaps impose a greater burden on a defendant’s

day-to-day existence than a superficially more jarring adjustment to a less

omnipresent aspect of that defendant’s life.

But that is a discussion for another day. For now, we are precluded from

deciding such hypotheticals – not only because the parties have failed to present

us with any particular future scenarios to evaluate, but also because even had

they done so, we would lack jurisdiction to rule on them. Speculative future

developments of that sort no longer concern “the already realized delegation of

judicial power to a probation officer” that we deemed ripe for review in

Villafane-Lozada (and again here); rather, they amount to the very “hypothetical

decision[s] that this delegation might allow in the future” that we deemed unripe

for our review. 973 F.3d at 147, 151 (emphasis omitted). Should Probation make

such a decision in the future, the issue will become ripe.14 It is not today.

14 At that point, “if the probation officer does overstep his authority,” Kunz may “seek recourse before the sentencing court.” Villafane-Lozada,