United States v. Undrae Moseby

In the United States Court of Appeals For the Seventh Circuit ____________________ Nos. 19-3476, 19-3481, 19-3484, 19-3537, 20-1113 & 20-1266 UNITED STATES OF AMERICA, Plaintiff-Appellee, v. GILBERTO VIZCARRA-MILLAN, et al., Defendants-Appellants. ____________________ Appeals from the United States District Court for the Southern District of Indiana, Indianapolis Division. No. 1:17-cr-00222-JMS-TAB — Jane Magnus-Stinson, Judge. ____________________ ARGUED MAY 12, 2021 — DECIDED SEPTEMBER 30, 2021 ____________________ Before FLAUM, HAMILTON, and BRENNAN, Circuit Judges. HAMILTON, Circuit Judge. Richard Grundy III and his net- work of drug suppliers, couriers, distributors, and dealers traﬃcked hundreds of pounds of methamphetamine in Indi- anapolis. Grundy and over two dozen co-conspirators were indicted. Most ultimately pled guilty. After a three-week trial, Grundy and four other defendants were convicted of all the charges against them. In these consolidated appeals, the ﬁve 2 Nos. 19-3476, et al. trial defendants and one defendant who pled guilty challenge their convictions. There are no sentencing issues. Grundy argues that the district court violated his Sixth Amendment right to counsel by improperly obstructing him from representing himself. Gilberto Vizcarra-Millan argues that the district court should have disqualiﬁed his chosen counsel due to a conﬂict of interest. Derek Atwater, James Beasley, and Undrae Moseby all challenge the denials of their untimely motions to suppress evidence. Atwater, Beasley, and Ezell Neville all contend that the evidence was insuﬃcient to support some of their convictions. We aﬃrm the convictions of Grundy, Vizcarra-Millan, Moseby, Atwater, and Neville. We also aﬃrm the conviction of Beasley on one count but reverse his convictions on two others. The evidence at trial necessarily left a reasonable doubt as to whether he committed those crimes. Beasley’s case is remanded to the district court for resentencing on the one remaining count. I. Factual and Procedural Background We provide the basic outline of Grundy’s network and the procedural history of this case here, with more speciﬁc details as needed for each defendant’s appeal. Because the co-con- spirators’ appeals do not hinge on the substance of the con- spiracy, we omit many of the details of the inner workings of Grundy’s crew as presented during the trial. A. Grundy and his Crew Grundy has been of interest to law enforcement for years. He has been arrested or indicted for, among other things, murder and conspiracy to commit murder, though he was never convicted of so serious a charge. As recently as 2015 or Nos. 19-3476, et al. 3 2016, he had been indicted on four counts of conspiracy to commit murder, but he ultimately pled guilty to a lower-level state charge for marijuana distribution. After his plea in the state case, Grundy pooled seed money to restart his drug-traﬃcking network. Gilberto Vizcarra-Mil- lan, who lived in Arizona, provided the drugs to Grundy. A network of couriers, including defendant Undrae Moseby, brought the drugs to Indiana. Once the drugs arrived in Indianapolis, Grundy distrib- uted them himself or via a network of wholesalers, including David Carroll. These wholesalers then sold to retail dealers, including Derek Atwater and James Beasley. Ezell Neville was another associate of Grundy’s. The parties dispute the exact nature of his relationship with Grundy, but he also sold Grundy’s methamphetamine. All told, Grundy and his crew brought at least 280 pounds of highly pure methampheta- mine, as well as other drugs, to the streets of Indianapolis. In May 2017, federal law enforcement obtained wiretaps for the cell phones of crew members. FBI agents also coordi- nated controlled drug buys from Grundy’s dealers. Things started to fall apart for the Grundy gang in August 2017. Law enforcement seized over $84,000 in cash that was headed to Vizcarra-Millan to purchase more drugs. Next, in September 2017, agents intercepted a phone call between Carroll and Beasley discussing an upcoming methamphetamine deal. Soon after Beasley purchased the methamphetamine, local police stopped the car in which he was a passenger for a rou- tine traﬃc violation. The oﬃcers smelled raw marijuana and recovered drug paraphernalia while searching the car. A search of Beasley’s person uncovered methamphetamine. The driver of the car, Susan Koch, told the oﬃcers that Beasley 4 Nos. 19-3476, et al. stored more methamphetamine in her home and gave her consent for a search that recovered more methamphetamine. On November 17, federal agents executed several search warrants against members of the conspiracy. As relevant here, they uncovered methamphetamine and drug-traﬃcking par- aphernalia at Atwater’s house. They also raided what they be- lieved to be the “clubhouse” of Grundy’s gang. Several mem- bers of the conspiracy were present, and the police collected several cell phones, including two used by Moseby that con- tained incriminating evidence of his connection to the Grundy crew. B. Pretrial and Mistrial In two separate cases, over two dozen members of Grundy’s gang were charged with federal oﬀenses, including conspiracy to distribute drugs and money laundering. Most defendants ultimately pled guilty. David Carroll, one of Grundy’s key wholesalers, agreed to testify for the prosecu- tion. Carroll’s attorney, John Tennyson, however, had also been retained by Vizcarra-Millan, who was charged in the sec- ond, formally distinct but related case. The government pointed out the potential conﬂict of interest—if Carroll were to testify at trial (which he had agreed to do), he might incrim- inate Vizcarra-Millan. If he did, Tennyson’s duties to his re- spective clients would conﬂict. See Indiana Rule of Prof’l Con- duct 1.7. Two district judges held hearings to discuss the conﬂict with Vizcarra-Millan and whether he wanted to waive it. At the ﬁrst hearing with Judge Barker, the potential conﬂict ap- peared both minor and unlikely to become actual, and Viz- carra-Millan said he would waive any conﬂict of interest. Nos. 19-3476, et al. 5 When both of the Grundy cases were later consolidated before Judge Magnus-Stinson, she also held a hearing that ended with Vizcarra-Millan again waiving his right to unconﬂicted counsel. But then, ﬁve weeks before trial, attorney Tennyson ﬁled a motion to withdraw the waiver because his client had rejected the government’s plea oﬀer and said he intended to go to trial. Contrary to what he had told both judges months earlier, Ten- nyson now claimed his conﬂict of interest prohibited him from adequately representing Vizcarra-Millan at all. Judge Magnus-Stinson denied Tennyson’s motion without holding a hearing. Vizcarra-Millan later pled guilty without a deal from the government. In the meantime, the district court had set a deadline in February 2019 for ﬁling motions to suppress evidence. Months later, just days before trial, defendants Beasley and Moseby submitted motions to suppress, styled as evidentiary motions in limine. The district court denied both motions. Trial got under way on July 8, 2019, in Indianapolis. The court took the unusual step of empaneling an anonymous jury after the government came forward with evidence of at- tempted witness tampering and intimidation. The court pro- vided juror information to defense counsel but forbade the defendants themselves from learning the jurors’ names or de- tailed personal information from which they could be identi- ﬁed. The ﬁrst trial did not last long. By day three, the district court learned that, despite the precautions, some defendants had gotten their hands on this conﬁdential information. Moseby had written down the names of several jurors. Grundy had obtained partially redacted juror questionnaires. 6 Nos. 19-3476, et al. The district court declared a mistrial and moved the trial venue to Evansville. Between the mistrial and the second trial, Grundy moved to represent himself. He wanted to pursue his own strategy, but he also wanted standby counsel to help him with the nuts and bolts of lawyering. The district court held a hearing that ended when Grundy ultimately declined to waive his Sixth Amendment right to counsel. C. Second Trial and Posttrial Procedure The second trial lasted three weeks. As relevant here, sev- eral co-conspirators testiﬁed against Grundy, implicating him as the mastermind of this drug traﬃcking ring. The evidence showed that Atwater, Beasley, and Neville all bought drugs from Grundy or one of his associates. The details of each de- fendant’s relationship to Grundy’s network, however, varied considerably, as did the circumstances of each documented purchase. During the trial, Atwater also submitted another late motion to suppress. The district court denied it, as well. The jury found all defendants guilty on all counts. All de- fendants were found guilty of Count 1, conspiracy to distrib- ute drugs. Neville was also convicted of conspiring to launder a monetary instrument in connection with the seizure of more than $84,000 from Grundy’s shipment of cash (Count 24), and Beasley was convicted on two counts of possessing metham- phetamine with intent to distribute (Counts 16 and 17). The district court imposed a life sentence on Grundy. At- water was sentenced to 216 months; Beasley was sentenced to 216 months, Moseby to 240 months, and Neville to 360 months. Nos. 19-3476, et al. 7 Vizcarra-Millan had pled guilty before the ﬁrst trial. His attorney, Tennyson, submitted both a sentencing memoran- dum and a motion to withdraw his guilty plea. The district court held a hearing at which it ﬁrst denied the withdrawal motion and then sentenced Vizcarra-Millan to 300 months in prison. D. Issues on Appeal On appeal, Grundy argues that the district court violated his Sixth Amendment right to counsel by unduly discourag- ing him from exercising his right to represent himself. Viz- carra-Millan contends that his right to counsel was violated by the district court’s failure either to disqualify Tennyson or to grant his motion to withdraw his conﬂict waiver shortly before trial. Atwater, Beasley, and Moseby appeal the district court’s denials of their respective motions to suppress. Atwater ar- gues that the search of his home was not supported by prob- able cause. Beasley moved to suppress the evidence seized during the traﬃc stop on the ground that the oﬃcers lacked reasonable suspicion that the driver had committed a traﬃc violation. Moseby objects to the government’s seizure of his cell phones during the raid on Grundy’s clubhouse and its search of those phones. Finally, Atwater, Neville, and Beasley argue that the evi- dence did not support some of their convictions. All three say that the evidence at trial was insuﬃcient to rule out the possi- bility that they were merely buyers from Grundy’s gang, as opposed to co-conspirators. Neville also challenges his money-laundering conviction, and Beasley challenges his 8 Nos. 19-3476, et al. conviction for constructively possessing the methampheta- mine found at Koch’s home. In Part II, we address and reject ﬁrst Grundy’s and then Vizcarra-Millan’s arguments that their Sixth Amendment rights to counsel were violated. In Part III, we aﬃrm the deni- als of the motions to suppress ﬁled by Atwater, Beasley, and Moseby. In Part IV, we address the suﬃciency of the evidence for Atwater, Neville, and Beasley, aﬃrming all convictions ex- cept two of Beasley’s. II. Right to Counsel: Grundy and Vizcarra-Millan The Sixth Amendment provides that “in all criminal pros- ecutions, the accused shall enjoy the right … to have the assis- tance of counsel for his defense.” This constitutional com- mand encompasses the accused’s right to choose his own counsel, to have counsel free of conﬂicts of interest, and to choose to forgo counsel and to represent himself. Grundy, by invoking his right to represent himself, and Vizcarra-Millan, by invoking his right to choose his own counsel despite a conﬂict of interest, each put the district court in a constitutional double-bind. No matter how a district court decides each issue, the defendant will have an issue to appeal. We have, for example, compared a district court’s navigation of a request to proceed pro se to passing between Scylla and Charybdis, see United States v. Sandles, 23 F.3d 1121, 1127 (7th Cir. 1994), and similarly, “trial courts confronted with multi- ple representations face the prospect of being ‘whip-sawed’ by assertions of error no matter which way they rule.” Wheat v. United States, 486 U.S. 153, 161 (1988); see also United States v. Lowry, 971 F.2d 55, 60 (7th Cir. 1992) (quoting same and en- couraging district courts to consider the ramiﬁcations of joint- Nos. 19-3476, et al. 9 representation disqualiﬁcation rulings in light of the high constitutional stakes in either direction). We accordingly give considerable deference to the district court’s decisions to the extent that they are subject to abuse-of-discretion analysis. We must also be mindful of the district court’s front-row seat, which is especially relevant where a defendant deliberately seeks to sow ambiguity for a tactical advantage. See generally United States v. Balsiger, 910 F.3d 942, 952 (7th Cir. 2018). A. Grundy’s Motion to Represent Himself Richard Grundy presents a claim under Faretta v. Califor- nia, 422 U.S. 806, 819 (1975). He says that the district court’s eﬀort to ensure that his intended waiver of his Sixth Amend- ment right to counsel was knowing and intelligent went too far, eﬀectively preventing him from exercising his Faretta right, a request that he ultimately withdrew. As an accused defendant, Grundy had a Sixth Amend- ment right to represent himself: “The Sixth Amendment does not provide merely that a defense shall be made for the ac- cused; it grants to the accused personally the right to make his defense.” Faretta, 422 U.S. at 819. But any waiver of counsel in favor of self-representation must be knowing and intelligent. A pro se defendant “should be made aware of the dangers and disadvantages of self-representation, so that the record will establish that he knows what he is doing and his choice is made with eyes open.” Id. at 835. To establish a record, we strongly encourage district courts to conduct a formal hearing into whether the defend- ant is going into self-representation with “eyes open.” Faced with a defendant invoking his right to proceed pro se, a dis- trict court should: 10 Nos. 19-3476, et al. engage in a thorough and formal inquiry with a defendant that probes his age, education level, and understanding of the criminal charges and possible sentences. The court should also in- form the defendant of the diﬃculties of pro- ceeding pro se. United States v. Johnson, 980 F.3d 570, 577 (7th Cir. 2020) (quo- tation marks and citations omitted). We are reluctant to police too closely the details of district courts’ Faretta hearings. The district judge “is on the razor’s edge in assisting a defendant to make an informed choice between representation by coun- sel with whom the defendant is irrationally dissatisﬁed and self-representation.” United States v. Oreye, 263 F.3d 669, 672 (7th Cir. 2001). 1 This case illustrates this razor’s edge. To ensure that a waiver is “knowing and intelligent,” the court should point out the disadvantages of self-representation. If the court does not say much about these challenges, defendants who waive their right to counsel may argue on appeal that the Faretta col- loquy was too cursory. E.g., Johnson, 980 F.3d at 577. But if the district court goes into detail about these challenges, a defend- ant who chooses in the end not to waive counsel may argue on appeal that his Faretta rights were burdened by a colloquy that was too persuasive. See Sandles, 23 F.3d at 1127. That’s part of the reason there is no “talismanic procedure” for Faretta hearings, Torres v. United States, 140 F.3d 392, 401 (2d 1Formal hearings are not always necessary. We have affirmed find- ings that a defendant constructively waived the right to counsel where the record showed clearly that the defendant knowingly and intelligently waived the right and understood the risks of going it alone. E.g., Balsiger, 910 F.3d at 953–54. Nos. 19-3476, et al. 11 Cir. 1998), and why we generally review district court’s Faretta hearings deferentially. E.g., United States v. Hill, 252 F.3d 919, 928 (7th Cir. 2001) (“Often asking the Benchbook questions may ensure that the defendant has his eyes open, but we do not read any of this court’s decisions to hold that the litany is pre- scribed in every case or that advice about any particular dis- advantage of self-representation is essential; such a reading would put us at odds with the Supreme Court.”). The key is that the accused must make the ﬁnal choice, but only once he understands his alternatives. Nonetheless, as a matter of prudence if not constitutional law, district judges should discourage all or nearly all defend- ants from exercising their rights under Faretta. E.g., Johnson, 980 F.3d at 578 (commending judge for having “strongly ad- vised Johnson against proceeding pro se”); United States v. Mancillas, 880 F.3d 297, 302 (7th Cir. 2018) (“courts ‘should … advise the defendant that proceeding pro se is unwise.’”), quoting United States v. Todd, 424 F.3d 525, 531 (7th Cir. 2005); cf. Kubsch v. Neal, 800 F.3d 783, 812–16 (7th Cir. 2015) (aﬃrm- ing denial of habeas relief where petitioner argued that trial judge did not try hard enough to discourage him from repre- senting himself in capital sentencing phase; defendant made competent strategic decision not to present evidence to miti- gate three brutal murders, but to emphasize residual doubt about guilt), vacated on other grounds, 838 F.3d 845 (7th Cir. 2016) (en banc). The waiver of counsel must also be “unequiv- ocal.” That means that district courts must press diﬃcult, hes- itant, or ambivalent defendants to answer yes or no whether they wish to waive the right to counsel. See United States v. Campbell, 659 F.3d 607, 612 (7th Cir. 2011) (discussing im- portance of unequivocal demand and rationale for require- ment), vacated on other grounds, 568 U.S. 802 (2012). 12 Nos. 19-3476, et al. Turning to the facts here, seven days after the mistrial was declared in the Indianapolis trial, Grundy moved to represent himself. The district court held a Faretta hearing ﬁve days later. The court began by asking Grundy about his age, edu- cation, drug-use history, and general ability to read and un- derstand evidence. The court informed Grundy of the charges against him and the long prison terms that could be manda- tory if he were convicted. See Johnson, 980 F.3d at 577 (endors- ing such advisements in Faretta hearing). The bulk of the hearing was devoted to Grundy’s request that his current attorney, Kenneth Riggins, be appointed standby counsel and to the data security problems posed by having a detained defendant with a track record of disregard- ing court orders seek unlimited access to conﬁdential discov- ery within the jail. After discussing these issues for almost an hour, the court noted that if Grundy were to invoke unambig- uously and unconditionally his right to proceed pro se, it would appoint attorney Riggins to have a limited role as standby counsel. The court also said it would likely require that Riggins retain physical possession of the most sensitive documents, though Grundy could review those documents with Riggins during jail visits. The court ﬁnally asked Grundy for a yes-or-no answer. Grundy replied, “I guess I can’t represent myself.” He ex- plained that it was a hard choice because he did not know how the district court would rule on future motions, includ- ing those pertaining to data security, whether the court might allow him to re-invoke his right to counsel, and how it would rule on a variety of late motions in limine he had contem- plated ﬁling. Nos. 19-3476, et al. 13 On appeal, Grundy makes two distinct arguments. First, he says that the court was too persuasive during the Faretta hearing. Second, he says that the court should not have en- dorsed the government’s suggestion that his access to conﬁ- dential documents should be limited. 1. Too Persuasive? We have suggested on occasion that a too-persuasive Faretta colloquy could burden the right. Kubsch, 800 F.3d at 812 (“If a judge believes, as the trial judge did here, that the defendant is making a knowing and intelligent waiver, then she would commit constitutional error by discouraging that decision too strongly.”); Oreye, 263 F.3d at 672 (Faretta “right is not honored if judges must depict self-representation in such unremittingly scary terms that any reasonable person would refuse.”), quoting Hill, 252 F.3d at 928–29; United States v. Sandles, 23 F.3d 1121, 1127 (7th Cir. 1994) (a too-persuasive colloquy risks “trammeling the defendant’s constitutional right to present his own defense”). But Grundy does not cite, and we have not found, a case in which we have actually re- versed a district court for dwelling on the “perils of self-rep- resentation … in truly loving detail.” Oreye, 263 F.3d at 672. Instead, we “have tried to keep the permissible middle ground between these opposing errors fairly broad, allowing trial judges reasonable leeway to adapt the inquiry to the cir- cumstances of the case without requiring a script or check- list.” Kubsch, 800 F.3d at 813. Grundy’s hearing was dominated by prolonged discus- sions regarding his request to have his current attorney ap- pointed as standby counsel. The district court responded that defendants who proceed pro se must represent themselves, and they cannot be co-counsel with their standby lawyers. 14 Nos. 19-3476, et al. Grundy explained that he did not want attorney Riggins to be co-counsel, but “just need[ed] him to be there to, like, help [] with the litigating parts and stuﬀ like that.” The district court explained that that is not the job of standby counsel and, in any event, a criminal defendant has no right to standby coun- sel. See United States v. Harrington, 814 F.3d 896, 901 (7th Cir. 2016) (collecting cases; “a court has no obligation” to appoint standby counsel). The district court underscored that if Grundy were to pro- ceed pro se, he would be truly on his own. Grundy later ex- plained that he viewed standby counsel as “training wheels”—someone who could help him organize his thoughts, sort through the evidence, and consult throughout the trial. To that end, he also wanted to have access to Riggins’ computer at the counsel table during trial so that he and Rig- gins could quickly scroll through the electronic evidence. Grundy’s concerns were greater than the computer, though, as he wanted “to be able to, to see what is about to happen before it happens.” The court reiterated that it was concerned that Grundy en- visioned an expansive role for Riggins. The court announced that it would appoint Riggins as standby counsel but would not tolerate hybrid representation (that is, the two working together as co-counsel). The court again explained that trial advocacy “is a highly technical involved area of education, training, and understanding. I am not saying that to intimi- date you. I am required by law to tell you that it is a hard thing to do, but it is your absolute right to do it, and if that is what you want to do, you have the right to do it.” The court then read from the Federal Judicial Center’s Benchbook, which con- tains a suggested script for Faretta hearings, strongly urging Nos. 19-3476, et al. 15 Grundy not to try to represent himself. Grundy asked whether he could reverse course during trial. The court re- fused to deal with hypotheticals because it needed a “ﬁrm de- cision.” Grundy balked at the phrase, “I am giving up my right” because he was concerned about the possibility that he could not re-invoke his right to counsel. The district court’s repeated questioning on Grundy’s un- derstanding of standby counsel did not run afoul of Faretta. We have explained that hybrid representation “is generally to be avoided,” Carlson v. CSX Transportation, Inc., 758 F.3d 819, 826 (7th Cir. 2014), citing United States v. Chavin, 316 F.3d 666, 671–72 (7th Cir. 2002), if not outright “forbidden.” Oreye, 263 F.3d at 672. As noted, before the district court could accept Grundy’s waiver, it had to establish a record “that he knows what he is doing and his choice is made with eyes open.” Faretta, 422 U.S. at 835 (quotation marks and citation omitted). As the foregoing summary illustrates, when Grundy at ﬁrst said he wanted to waive his right to counsel, it was not clear whether he understood the consequences or understood that standby counsel is not co-counsel and that once he waived his Sixth Amendment right, he had no guarantee that the court would reappoint counsel if he later changed his mind. We see no error in the district court’s explanation of the limits of standby counsel, especially in the face of Grundy’s repeated ambiguous answers as to what he wanted from his attorney. Faretta hearings “may require a give and take be- tween the accused and someone trying to educate him about counsel’s beneﬁts.” Jean-Paul v. Douma, 809 F.3d 354, 359 (7th Cir. 2015) (quotation marks and citation omitted). Here, Grundy repeatedly said that he wanted to have it both ways in trial and to be allowed to change his mind. That’s a recipe 16 Nos. 19-3476, et al. for trouble, and Judge Magnus-Stinson correctly acted to ex- plain that hybrid representation was not possible and that she could not guarantee Grundy that he could change his mind later. In a variation on this challenge, Grundy argues that the district court should have stopped the hearing about halfway through because he invoked his Faretta right clearly at that point. He contends that any further discussion burdened his right. But his argument both misreads the record and misun- derstands Faretta and the purpose of the hearing. Grundy points to the following exchange after a ﬁfteen-minute recess to discuss his options with his attorney: The Court: So you have had about 15, 20 minutes to talk to Mr. Riggins. I, before we—do you still want to proceed with the hearing on your motion to represent yourself? Grundy: Yes. I wanted to ask. I kind of got to the bottom of what I needed to know if I am going to represent myself or not. Grundy asserts that this was an unequivocal invocation of his right to self-representation and that the hearing should have stopped then. We disagree for two reasons. First, based on the transcript, this simply was not a clear invocation of his Faretta right. Ra- ther, Grundy said only that he wanted to continue the conver- sation. See United States v. Miles, 572 F.3d 832, 836–37 (10th Cir. 2009) (aﬃrming district court’s interpretation of ambigu- ous statements as not expressing desire to proceed pro se). Second, and more fundamental, even a clear attempt to in- voke Faretta is only the beginning of the matter, not the end. Nos. 19-3476, et al. 17 The court was required to ensure that any waiver by Grundy of his right to counsel would have been knowing and intelli- gent. The court was not yet satisﬁed that it had established a record that Grundy understood that he would in fact be on his own and without Riggins’ formal assistance. See United States v. Campbell, 659 F.3d 607, 612 (7th Cir. 2011) (where de- fendant clearly invokes right to proceed pro se, next step is for district court to engage in colloquy to address that re- quest). Faretta is not a Miranda-like right, where custodial interro- gation must cease once a suspect says he wants a lawyer. See, e.g., United States v. Wysinger, 683 F.3d 784, 796 (7th Cir. 2012) (applying Miranda). When a defendant clearly and unambig- uously invokes his Faretta right to proceed pro se, the district court is not required to take the statement at face value and as ﬁnal. To the contrary, the district court must refuse to accept a supposed waiver until the record shows that the waiver is knowing and intelligent. Because Grundy remained ambiva- lent and tried to invoke his Faretta right conditionally, the court did not burden his right by asking further questions. 2. Access to Evidence Grundy’s second argument hinges on issues related to trial and witness security. He contends that the district court erred by asking the government for any input during the hearing and that the court’s suggested conditions for data se- curity burdened his constitutional rights. He points to the dis- trict court’s suggestions that it would forbid him from taking possession of sensitive discovery materials (such as still-con- ﬁdential witness statements), though he could review any and all documents with attorney Riggins, and that it might not allow Grundy to use Riggins’ computer during the trial 18 Nos. 19-3476, et al. itself. As we understand Grundy’s argument on appeal, he says these conditions would have been unconstitutional if they had been imposed, so invoking them as a threat to coerce him into staying with his attorney violated Faretta. 2 The district court correctly focused on the serious security concerns posed by Grundy’s prosecution. The right to repre- sent oneself “is not unlimited,” and “[s]ecurity considerations … may require special adjustments.” Milton v. Morris, 767 F.2d 1443, 1446 (9th Cir. 1985), citing United States v. Chatman, 584 F.2d 1358, 1360 (4th Cir. 1978) (no constitutional violation where defendant’s access to legal resources was restricted af- ter being moved into “segregated conﬁnement”); see also United States v. Galloway, 749 F.3d 238, 242 (4th Cir. 2014) (no constitutional violation where district court’s discovery or- ders inconvenienced pro se defendant because they were “jus- tiﬁed by the circumstances,” that is, “enormous security is- sues”); United States v. Sarno, 73 F.3d 1470, 1491–92 (9th Cir. 1995) (“The right of access is not unlimited, but must be bal- anced against the legitimate security needs or resource con- straints of the prison;” no constitutional violation where gov- ernment allowed defendant to review discovery to determine which documents should be copied for his pretrial 2Circuits have reached different answers about the extent to which a pro se defendant in custody has a constitutional right to government as- sistance in his defense, and about where to find such a right in the Consti- tution. See Kane v. Garcia Espitia, 546 U.S. 9, 10 (2005) (per curiam) (noting split and reversing grant of habeas relief because state-court determina- tion that pro se defendant was not entitled to law library access was not contrary to clearly established federal law: “Faretta says nothing about any specific legal aid that the State owes a pro se criminal defendant.”). We need not deepen the split here because Grundy’s claim fails for reasons explained in the text. Nos. 19-3476, et al. 19 preparation—but gave him only 20 hours to inspect 250,000 pages of material); United States v. Bisong, 645 F.3d 384, 396 (D.C. Cir. 2011) (quoting same and expressing skepticism that pro se defendants have any Sixth Amendment right to discov- ery in preparing their defense). Recall the context of this hearing. The district court, al- ready alert to the grave security concerns in the case, had taken the unusual step of empaneling an anonymous jury based on concerns about juror intimidation. The ﬁrst trial ended in a mistrial, in part because Grundy himself obtained conﬁdential juror questionnaires. The transcript of the Faretta hearing also indicates that he was placed in solitary conﬁne- ment as a result of the conduct that led to the mistrial. 3 Given this backdrop, Grundy posed obvious and legiti- mate security concerns, even with counsel. If he were to rep- resent himself, the district court had to consider further po- tential problems. Judge Magnus-Stinson’s discussion of po- tential solutions was important and responsible. A core part of the district judge’s job is to protect jurors, witnesses, and the integrity of judicial proceedings more broadly. Consider- ation of these proposals was entirely appropriate under the circumstances. Discussion of them did not violate Grundy’s constitutional rights. 3 Though the district court did not discuss it at the hearing, Grundy had a hand in producing a song called “COS [Code of Silence] (Ain’t No Tellin).” The lyrics instruct to “close your mouth for the feds” because people who cooperate with the police are “better off dead.” The accompa- nying music video, which features Grundy, includes graphic fictionalized depictions of what happens to people who “rat” on the gang. (They are murdered.) 20 Nos. 19-3476, et al. Grundy’s argument also faces a threshold problem. Since he declined to proceed pro se, we do not know what steps the district court would actually have taken. In reviewing the court’s statements, we do not adopt the least charitable read- ing possible. The court told Grundy that if he chose to proceed pro se, they could “ﬁgure something out,” and we do not as- sume that the district court would have done so unreasonably or in bad faith. At least some arrangements like the proposed restrictions would have been warranted by the unusual circumstances of Grundy’s prosecution. The government requested that Rig- gins, as standby counsel, retain physical possession of certain sensitive documents, though Grundy and Riggins would be allowed to review them together. As it mulled its options, the court made clear that under any scenario, it would not forbid Grundy from viewing any document, but it had legitimate se- curity concerns about Grundy taking conﬁdential witness statements into the jail. Grundy’s solitary conﬁnement may have made this process cumbersome, but we are not con- vinced that approach would have been per se unconstitu- tional or even an abuse of discretion. See Chatman, 584 F.2d at 1360 (restricting a particularly dangerous defendant’s access to legal materials did not violate Bounds); Galloway, 749 F.3d at 242 (aﬃrming limits on pretrial detainee’s access to discov- ery in light of security concerns). Grundy also objects to the court’s comments on his request to use Riggins’ computer during trial if he waived counsel. We see no error. The court merely noted that it would “ﬁgure something out” if Grundy chose to proceed without counsel. In any event, Grundy’s request was ambiguous. Putting aside the ethical and security issues posed by a court ordering a Nos. 19-3476, et al. 21 defense attorney to give one client access to all his ﬁles, it is not even clear that Grundy wanted a computer as such. On follow-up questioning, he explained, “I would like to be able to, to see what is about to happen before it happens.” As the district judge noted, Grundy’s comment was more likely an observation about the beneﬁts of having the guiding hand of experienced counsel. To sum up on this issue, the court’s Faretta hearing was procedurally sound, and its discussion of possible restrictions on access to conﬁdential documents due to security concerns did not violate the Constitution. We aﬃrm Grundy’s convic- tion. B. Vizcarra-Millan’s Potentially Conﬂicted Counsel Vizcarra-Millan’s appeal attempts to take advantage of the inherent tension between two aspects of the accused’s Sixth Amendment right to counsel: the right to counsel of one’s choice and the right to counsel who is free of conﬂicts of in- terest. United States v. Gonzalez-Lopez, 548 U.S. 140, 150 (2006) (denial of counsel of choice is a structural error; reversal does not require showing of prejudice); Wheat v. United States, 486 U.S. 153, 159-60 (1988) (right to counsel free of conﬂicts); United States v. Turner, 594 F.3d 946, 955 (7th Cir. 2010) (re- versing conviction following trial where district court had disqualiﬁed defendant’s preferred counsel too readily in face of potential conﬂict). If the accused says he wants to be represented by a lawyer who faces an actual or potential conﬂict of interest, the district court can err in either direction: either deny the accused his choice of counsel or deny him counsel free of conﬂicts of in- terest. This tension also can create opportunities for 22 Nos. 19-3476, et al. manipulation by unscrupulous counsel or a defendant who faces overwhelming evidence of guilt. See Wheat, 486 U.S. at 161. District courts therefore have broad discretion in how to handle this constitutional balancing act. To complicate mat- ters further here, Vizcarra-Millan seemed to change his mind over the course of the case. To begin, we review the facts of the two hearings where Vizcarra-Millan waived his right to conﬂict-free counsel and explain why we ﬁnd no error at that stage. We then consider Vizcarra-Millan’s change of position a few weeks before trial, when he tried to revoke his waiver of his right to conﬂict-free counsel. The district court rejected this change of course and found that he was acting in bad faith, which Vizcarra-Millan essentially admitted at sentencing. We also ﬁnd no reversible error in the court’s handling of his change of mind. 1. Vizcarra-Millan’s Waivers of the Right to Conﬂict-Free Counsel a. The Problem Arises Before Judge Barker Defendant Gilberto Vizcarra-Millan hired a conﬂicted at- torney, John Tennyson, to represent him. Tennyson had two other clients in the Grundy gang, and both had agreed to tes- tify for the government. The government informed Judge Barker that attorney Tennyson represented several defend- ants in two Grundy-related cases—Carroll and Carter in one case, and Vizcarra-Millan in the other. When Judge Barker learned of this conﬂict, she promptly held a hearing to deter- mine whether Vizcarra-Millan and the other defendants knowingly and intelligently waived the right to conﬂict-free counsel and whether she should nonetheless disqualify Ten- nyson based on the conﬂict. Nos. 19-3476, et al. 23 Judge Barker explained that a hearing was necessary be- cause Vizcarra-Millan “intends to proceed to trial,” and Car- roll and Carter had agreed to testify for the government. The judge summarized the facts: Carroll and Carter were part of Grundy’s drug-distribution network in Indianapolis. They knew generally that the drugs came from Arizona, but they did not know that Vizcarra-Millan was the source in Arizona. Their testimony would implicate Vizcarra-Millan only indi- rectly. Tennyson suggested that he would not cross-examine either Carroll or Carter except to ask whether they knew Viz- carra-Millan or knew the source of the drugs, to which they would reply no. Skeptical that Tennyson or the government could guaran- tee exactly how two cooperating witnesses would testify at trial, Judge Barker observed that, given these conﬂicts, “it seems so obvious to me … that you’re on thin ice.” Nonethe- less, she conducted a full colloquy to inform each of Tenny- son’s clients of his right to conﬂict-free counsel and the dan- gers of proceeding with Tennyson. All three clients orally waived their right to conﬂict-free counsel. Carroll and Carter both told law enforcement that they did not know Vizcarra- Millan. The government agreed that neither witness had sug- gested any knowledge that would implicate Vizcarra-Millan directly. Tennyson also noted that he expected other trial evi- dence to establish clearly “a very strong Arizona connection,” so that he would not contest the general geographic source of the drugs at trial. Judge Barker then discussed the potential conﬂict with Vizcarra-Millan. She noted that it was unlikely to be “serious” and that it was “not a substantive issue where there’s a direct obvious conﬂict.” She then told Vizcarra-Millan of his right to 24 Nos. 19-3476, et al. conﬂict-free counsel. He said that he waived that right and wanted to continue with attorney Tennyson. Judge Barker also ordered Tennyson to ﬁle written waivers, and she issued a written order observing that disqualiﬁcation was not yet necessary given the limited nature of the conﬂict and Viz- carra-Millan’s knowing and intelligent waiver. The order also observed that limiting measures short of disqualiﬁcation might eﬀectively remedy any future conﬂict. Judge Barker handled the issue exactly right. She exam- ined the nature of the potential conﬂict closely and concluded correctly that the potential conﬂict was not very serious or im- portant as a practical matter. She also noted that if the poten- tial conﬂict became actual, measures short of disqualifying counsel could protect Vizcarra-Millan’s rights. Judge Barker did not abuse her discretion by deferring to Vizcarra-Millan’s informed choice to stick with attorney Tennyson. As noted, it is risky to disqualify counsel over the objection of the accused, especially where the conﬂict remains only potential, because denial of the right to counsel of choice can also violate the Constitution. Gonzalez-Lopez, 548 U.S. at 150; see also Turner, 594 F.3d at 955 (reversing conviction following trial where district court had disqualiﬁed counsel too readily in face of potential conﬂict). b. The Second Hearing, Before Judge Magnus-Stinson Judge Magnus-Stinson took over the conﬂict issue when the two cases were consolidated. She ordered Tennyson to submit written waivers. When he did not, Judge Magnus-Stin- son held another hearing on February 19, 2019 on whether Vizcarra-Millan knowingly and intelligently waived his right to conﬂict-free counsel. At the hearing, he explained that he “got to kind of thinking” after the ﬁrst hearing. Judge Nos. 19-3476, et al. 25 Magnus-Stinson reiterated that he had an absolute right to conﬂict-free counsel, and that, based on the initial hearing be- fore Judge Barker, it sounded as if Tennyson could not guar- antee the absence of a conﬂict and that, at a minimum, Tenny- son’s other clients would likely implicate him indirectly through the Arizona connection. Vizcarra-Millan’s argument on appeal hinges on what happened next. He told the judge that he was “kind of stuck in a hard position” because he had already paid Tennyson in full and did not want to have to restart with a new attorney with trial just a few months away. Vizcarra-Millan then blamed the government for the “unfair” and “tough” position he was in. The court responded that it was not the govern- ment’s fault that he had retained Tennyson, but nonetheless observed that it did not sound as if he wanted to waive the conﬂict. Vizcarra-Millan expressed ambivalence about whether he wanted to waive his conﬂict with Tennyson, but he then announced for the ﬁrst time that he did not want to take the case to trial. The government provided background as to why plea ne- gotiations had broken down. Tennyson and Vizcarra-Millan had met with the government for a proﬀer session. During that meeting, the prosecutors learned that he had never signed the written waiver as ordered by Judge Barker. At the hearing before Judge Magnus-Stinson, the government ex- pressed no preference for or against Tennyson and waiver but noted that it could not be conﬁdent it could make a valid plea agreement where it was unclear whether Vizcarra-Millan’s counsel actually represented him and could properly repre- sent him: 26 Nos. 19-3476, et al. If he wants to sign the form and waive the con- ﬂict and proceed with Mr. Tennyson, I am ﬁne working with Mr. Tennyson. If he wants a new lawyer, I am ﬁne with working with a new law- yer, but I am not going to negotiate a plea agree- ment or someone’s cooperation under this pos- ture. The judge turned the choice back to Vizcarra-Millan. After a hearing and a half, he knew the stakes and the facts. He re- sponded: “I don’t mind waiving my right. I don’t mind waiv- ing it. I don’t mind waiving it.” He then complained that the prosecutor did not help him “in any type, shape, form, or way,” even in response to the proﬀer. The judge correctly observed that this was not an unequiv- ocal waiver, so she pressed Vizcarra-Millan to answer yes or no whether he wanted to continue with his waiver. The judge then went oﬀ the record for almost twenty minutes while Viz- carra-Millan discussed the issue with Tennyson. Following that conversation, he clearly and unequivocally waived his right to conﬂict-free counsel. He later ﬁled a written waiver of his right to conﬂict-free counsel. On appeal, Vizcarra-Millan argues that his waiver was in- valid because it was not knowing, intelligent, voluntary, and unequivocal. See United States v. Hill, 252 F.3d 919, 923–24 (7th Cir. 2001). We review that question de novo, though we defer to the credibility determinations of the district court. United States v. Balsiger, 910 F.3d 942, 952 (7th Cir. 2018). Based on this record, there is no question that Vizcarra-Millan’s waiver met these criteria. We defer to both Judge Barker’s and Judge Magnus-Stinson’s determinations that he understood the hearings. His comments in both hearings show that he Nos. 19-3476, et al. 27 understood what was happening. There is no hint that the government coerced the waiver. To the contrary, the govern- ment was agnostic about his decision. And ﬁnally, though Vizcarra-Millan was ambivalent about joint representation at the beginning of the second hearing, he cleared up any ambi- guity by the end of the hearing. To the extent he argues now that Tennyson misled him during the oﬀ-the-record break, two diﬀerent federal judges concluded that Vizcarra-Millan understood what was happening and what the stakes were, so no further colloquy would have been necessary. See Turner, 594 F.3d at 952 n.1 (emphasizing minimal procedural require- ments for hearings on waiver of conﬂict-free counsel). Vizcarra-Millan also argues that the red ﬂags in Tenny- son’s representation should have caused the court to disqual- ify Tennyson despite his own waiver and his invocation of his right to counsel of choice. Such a step would have invited ap- peal based on denial of his right to counsel of choice, as in Turner, 594 F.3d at 948 (reversing where district court disqual- iﬁed defendant’s counsel of choice based on potential conﬂict of interest). The fact that Tennyson was “up to [his] elbows” in conﬂicts, as Judge Barker put it, might have permitted the court to override the defendant’s choice, but he has not shown that judicial override of his choice was required. The question on appeal is whether Judge Magnus-Stin- son’s decision to accept the second waiver was within the wide band of discretion that district court judges have when facing such a dilemma. It was. While disqualiﬁcation of the defendant’s counsel of choice over his objection might have been permissible here under Wheat, it would have produced a sure-ﬁre issue for an appeal, especially in light of both Ten- nyson’s and Vizcarra-Millan’s statements minimizing the 28 Nos. 19-3476, et al. seriousness of what seemed like only a minor potential con- ﬂict that could, if necessary, be managed by steps short of dis- qualiﬁcation. See United States v. Gearhart, 576 F.3d 459, 465 (7th Cir. 2009) (collecting examples of creative evidentiary so- lutions to avoid conﬂicts); United States v. Turner, 651 F.3d 743, 750 (7th Cir. 2011) (citing same). 4 2. Vizcarra-Millan Changes His Mind After Vizcarra-Millan submitted his written waiver, the parties went back to the negotiating table but were unable to reach an agreement. On May 2, 2019, Tennyson reported at a status conference that Vizcarra-Millan would proceed to trial. Three weeks later, and just ﬁve weeks before trial, Tenny- son ﬁled a motion seeking to withdraw Vizcarra-Millan’s waiver of the right to conﬂict-free counsel. The motion was terse. It recited the procedural history and said: “The Defend- ant no longer wishes to waive his right to conﬂict-free repre- sentation because plea negotiations have broken down.” It added that Vizcarra-Millan “felt pressured into executing the waiver because the Government would not negotiate with him while the issue of whether he would sign a joint repre- sentation waiver was pending.” 4Much of Vizcarra-Millan’s argument against the validity of his sec- ond waiver seems to preview a future argument that he did not receive effective assistance of counsel during plea negotiations. His primary com- plaints about Tennyson do not concern the joint representation issue with Carroll and Carter. Rather, he questions Tennyson’s overall competence as an attorney and the adequacy of representation he provided in this case. But the overall record of Tennyson’s representation is not before us, and we express no opinion on the subject. See United States v. Flores, 739 F.3d 337, 340–41 (7th Cir. 2014) (explaining why we discourage defendants from bringing ineffective assistance claims on direct appeal). Nos. 19-3476, et al. 29 After brieﬁng but without a hearing, the district court de- nied the motion. The court analogized Vizcarra-Millan’s oral and written waivers to guilty plea colloquies and suggested that he not be able to recant his sworn testimony without some reason. It stressed that Vizcarra-Millan’s motion was ﬂatly in- consistent with the oral and written representations both he and his lawyer had made throughout the case. The court con- cluded that Vizcarra-Millan sought to use his counsel’s con- ﬂict of interest as a “bargaining chip in plea negotiations.” The court also found his “eleventh-hour” motion to be “suspect” because Vizcarra-Millan had rejected the government’s plea oﬀer over three weeks before the motion to withdraw the waiver was ﬁled, and at that hearing, the court underscored the logistical challenges that the trial would pose. The court, in other words, viewed the motion as a last-minute gambit to postpone the joint trial. 5 5 In the end, though, the district court disclaimed reliance on any lo- gistical or calendar problems that could be posed by granting a continu- ance to allow replacement counsel to get up to speed, though it found such concerns to be “notable.” See Gonzalez-Lopez, 548 U.S. at 152 (noting dis- trict court’s wide latitude in considering calendar concerns). Vizcarra-Mil- lan’s attempt to compare the timing of his motion to instances where we have held that a court abused its discretion in denying such a motion even closer to trial is thus misplaced. The district court did not deny his motion to withdraw on that ground. We note, however, that these are highly fact- and context-specific issues, and there is no hard timing cutoff. See United States v. Sellers, 645 F.3d 830, 836 (7th Cir. 2011) (“[W]e look at the whole of the circumstances surrounding the last minute filing.”). If the district court had granted Vizcarra-Millan’s motion, the other trial defendants would have had to sit in jail, their day in court postponed for months until replacement counsel could make it through the mountains of discovery, to say nothing of witnesses’ fading memories or weakening resolve to 30 Nos. 19-3476, et al. The district court focused most of its analysis on the fact that both Tennyson and Vizcarra-Millan had long been aware of this potential conﬂict of interest. The May 27 motion to withdraw his waiver argued that the likelihood that Vizcarra- Millan would go to trial would create a new conﬂict, but that possibility had always been on the table. Judge Barker’s hear- ing back in October 2018 was premised on Vizcarra-Millan go- ing to trial, and the parties discussed in detail what such a trial might look like. Tennyson oﬀered suggestions as to how to avoid or manage the conﬂict if it arose. Judge Barker noted other options in her order approving the waiver. Vizcarra- Millan was present for all of that. Two federal judges observed him through two hearings discussing the conﬂict of interest and potential for trial problems, and both thought that he un- derstood the issue and voluntarily chose to waive the poten- tial conﬂict. Vizcarra-Millan has two principal arguments for reversing based on the denial of these motions. First, he argues that the court made a procedural error by ruling on his ﬁrst motion without holding a hearing. At the very least, he says, the court’s failure to hold a hearing should call for de novo review of the denial. Second, he says that the court erred substan- tively because, if a defendant objects to conﬂicted counsel be- fore trial, it is a per se constitutional violation to make him proceed to trial. In the alternative, he argues that the district court held him to too high a standard for withdrawing his waiver. testify against the conspirators. It would not have been an abuse of discre- tion to deny Vizcarra-Millan’s motion on these grounds, too. Nos. 19-3476, et al. 31 a. Need for Another Hearing? As a general rule, an opportunity to be heard face-to-face can improve the accuracy of decision-making and ensures a sense of fairness and legitimacy in the process. See generally Joint Anti-Fascist Refugee Committee v. McGrath, 341 U.S. 123, 161–74 (1951) (Frankfurter, J., concurring). Perhaps another hearing might have sharpened the district court’s decision- making and aided our review, but we are not convinced that yet another hearing was essential here. In challenging the denial without a hearing, Vizcarra-Mil- lan relies on our cases involving motions to replace appointed counsel due to a lack of communication between client and counsel preventing an adequate defense. E.g., United States v. Jones, 844 F.3d 636, 642 (7th Cir. 2016). In such cases, we re- view for an abuse of discretion when a defendant has “had an opportunity to explain his reasons for requesting substitute counsel.” Id. We weigh “the timeliness of the defendant’s mo- tion, the adequacy of the district court’s inquiry into the mo- tion, and whether the conﬂict resulted in a total lack of com- munication preventing an adequate defense.” Id. Another hearing was not essential in this case. The im- portance of a hearing on issues of replacing counsel is based on two assumptions that do not apply here. First, a motion for substitution may be the ﬁrst time that the district court be- comes aware of the conﬂict or tension between attorney and client. Zillges, 978 F.2d at 371–72; see also United States v. Ryals, 512 F.3d 416, 418 (7th Cir. 2008) (reversing denial of motion for substitute counsel for sentencing). Second, where there has been a “total breakdown in communication,” an in-per- son hearing where a defendant can explain himself or herself serves an important purpose. With an asserted breakdown in 32 Nos. 19-3476, et al. communication, the court should not necessarily assume that counsel has described the problem accurately. See Ryals, 512 F.3d at 419–20 (defendant cannot be blamed because counsel described tension inadequately); United States v. Morrison, 946 F.2d 484, 498–99 (7th Cir. 1991). Neither of those concerns was present here. First, the dis- trict court had already held two hearings on the same issue. Everyone knew what the issues and concerns were. To be sure, a hearing would have allowed the district court to probe why Vizcarra-Millan’s thinking had changed, but it was not necessary to understand the fundamental issues in Tenny- son’s motion. See United States v. Bjorkman, 270 F.3d 482, 501 (7th Cir. 2001) (“even without [a] hearing the court may well have been suﬃciently apprised of the nature of Bjorkman’s complaints to rule on his request” where defendant’s letter provided some detail); see also United States v. Miller, 405 F.3d 551, 557 (7th Cir. 2005) (applying abuse-of-discretion review where district court held hearing “after it denied the request”). Thus, we review for abuse of discretion. Second, the conﬂict at issue did not implicate Tennyson’s ability to convey or fulﬁll Vizcarra-Millan’s wishes. The con- ﬂict was the same one that had been discussed all along. If Vizcarra-Millan were to go to trial, Carroll might testify that the drugs came from Arizona, and Tennyson would have to limit his impeachment of Carroll, also his own client, on that fact. Nothing about that conﬂict suggested that Tennyson was unable to represent Vizcarra-Millan, let alone that he could not present an adequate defense or describe the issue ade- quately. Nos. 19-3476, et al. 33 b. Rejecting the Change of Mind Turning to Vizcarra-Millan’s substantive challenges to the denials, the general rule is that if there is no waiver, “when- ever a trial court improperly requires joint representation over timely objection reversal is automatic.” Holloway v. Ar- kansas, 435 U.S. 475, 488 (1978). That general rule does not ap- ply here because Vizcarra-Millan did waive his right, know- ingly and voluntarily, and the case had proceeded for months in reliance upon that waiver. District courts are entitled to treat such waivers as binding. United States v. Roth, 860 F.2d 1382, 1387 (7th Cir. 1988) (“A judge is entitled to hold a de- fendant to statements made in open court and need not give him a hearing so that he may more readily contradict him- self.”). At least for the reasons we have described with the con- stitutional double-bind, we review only for an abuse of dis- cretion a district court’s refusal to allow the withdrawal of a valid waiver. Vizcarra-Millan’s evident bad faith supported denial of his attempt to withdraw his waiver. District courts have dis- cretion to deny bad-faith attempts to manipulate judicial pro- ceedings or to undermine the integrity of the judicial process. See, e.g., United States v. Fazzini, 871 F.2d 635, 643 (7th Cir. 1989) (district court did not err in refusing to inquire into well- documented about-face regarding pro se representation); United States v. Thibodeaux, 758 F.2d 199, 201 (7th Cir. 1985) (defendant “may not manipulate his right to counsel to un- dermine the orderly procedure of the courts or subvert the administration of justice” by ﬁling dilatory and contradictory motions regarding appointed counsel); cf. United States v. Gonzalez-Lopez, 548 U.S. 140, 151–52 (2006) (listing instances 34 Nos. 19-3476, et al. where courts weigh right to counsel against case and court management). Judge Magnus-Stinson soundly rejected Vizcarra-Millan’s attempt to blame the prosecutor for forcing him into this sup- posed corner. At the February 2019 hearing, the government had said it had no preference for or against Tennyson. The problem in the plea negotiations was that Vizcarra-Millan’s strategic ambiguity in orally waiving his right without sign- ing a written waiver had made it impossible to know whether Tennyson was truly authorized to negotiate on his behalf. Un- til that question was resolved deﬁnitively, plea negotiations would have been a waste of time. Vizcarra-Millan argues more generally that a district court is not allowed to deny a motion to withdraw a waiver of a constitutional right out of spite, and that applying any height- ened standard is suspect. He says that the district court should not have required him to show changed circumstances justifying his about-face. He also analogizes his case to others in which we have criticized district courts for failing to accom- modate other motions to substitute counsel even closer to trial. See, e.g., Sellers, 645 F.3d at 838. Vizcarra-Millan is correct that timing is often central when a district court denies a motion to change counsel. Where he veers oﬀ course, however, is misunderstanding the multiple ways in which timing may be relevant. It’s not only a matter of logistics, even if logistics alone would have justiﬁed denial of his about-face here. As the Fifth Circuit has noted, a district court can deny such a motion if the defendant “was attempt- ing to abuse his rights to achieve some mischief” or if “grant- ing his request would have interfered in any way with … cal- endaring.” United States v. Taylor, 933 F.2d 307, 311 (5th Cir. Nos. 19-3476, et al. 35 1991); see also United States v. Leveto, 540 F.3d 200, 207 (3d Cir. 2008) (“Certainly, evidence of a defendant’s dilatory motive is properly considered as a basis for denial. Moreover, partic- ularly as the trial date draws nearer, the district court can and should consider the practical concerns of managing its docket.”). In short, a defendant “may not manipulate his right to counsel to undermine the orderly procedure of the courts or subvert the administration of justice.” United States v. Balsi- ger, 910 F.3d 942, 954 (7th Cir. 2018), quoting Thibodeaux, 758 F.2d at 201. The court did not clearly err in ﬁnding that Vizcarra-Mil- lan acted in bad faith. Judge Magnus-Stinson observed Viz- carra-Millan in his second waiver hearing and later proceed- ings, and she concluded that the new attempt to revisit the conﬂict issue was just a ploy. Everyone, including Vizcarra- Millan, knew there were easy ways to work around this po- tential conﬂict. Based on the well-documented history of Viz- carra-Millan’s waiver, including his initial attempt to hold out on signing the waiver in an attempt to sow strategic ambigu- ity, we see no error in Judge Magnus-Stinson’s ﬁndings or her denial of Vizcarra-Millan’s pretextual motion. Even if there were doubt about Vizcarra-Millan’s bad faith before trial and his guilty plea, the doubt would have been erased at sentencing. After the court denied Vizcarra-Millan’s waiver, he pled guilty without a plea agreement. The court’s plea colloquy was the “most thorough, most pointed” the at- torneys had ever seen. Nonetheless, on the eve of sentencing, Vizcarra-Millan, through Tennyson, moved to withdraw his guilty plea. The judge gave him an opportunity to explain his actions throughout the litigation. He admitted that he had cre- ated ambiguity in whether he had waived his right to conﬂict- 36 Nos. 19-3476, et al. free counsel so as to keep his options open. In other words, he admitted that he had eﬀectively misled the court in his sworn Rule 44 colloquy. He also claimed that he had not understood the plea colloquy and therefore should be released from his plea. Judge Magnus-Stinson was not impressed: So I know you are trying to ﬁgure out an angle to get yourself oﬀ the hook that you are on, but you are on the hook is what I am telling you to- day. And so the Court will deny any request to withdraw the plea of guilty in this case. Because just for the record, I think you will say whatever you can to try to get out of it…. Vizcarra-Millan says that we should not credit his state- ments at sentencing because they were tainted by the court’s earlier denial of his motion to withdraw his waiver. We do not follow his logic. As noted, we review the district court’s bad- faith ﬁnding for clear error. The court found that he was mo- tivated by the desire to use the strategic revocation of his waiver as a bargaining chip in plea negotiations. He admitted at sentencing that he had done exactly that. The district court therefore did not abuse its discretion in denying Vizcarra-Millan’s motion as made in bad faith. Dis- trict courts are entitled to rely on “the representations of the defendant, his lawyer, and the prosecutor…. Solemn declara- tions in open court carry a strong presumption of verity.” Blackledge v. Allison, 431 U.S. 63, 73–74 (1977); United States v. Roth, 860 F.2d 1382, 1387–89 (7th Cir. 1988) (holding defend- ant to earlier waiver of right to conﬂict-free counsel). Tenny- son and Vizcarra-Millan’s about-faces gave the court a choice. Nos. 19-3476, et al. 37 It could either reward their gamesmanship by allowing their “transparent ploy for delay,” Morris v. Slappy, 461 U.S. 1, 13 (1983), or it could conclude that, in the interest of fair and or- derly administration of justice, they would have to stand by their repeated solemn declarations. Contrary to Vizcarra-Millan’s arguments, the district court did not ignore the Sixth Amendment stakes. Tennyson’s con- ﬂict was only potential and narrow. Carroll’s anticipated tes- timony that the drugs came from Arizona was simply not much of an issue for the government or Vizcarra-Millan. Ten- nyson and Vizcarra-Millan conceded as much. Carroll did not testify about it at trial. The record was clear that any conﬂict at trial could have been mitigated in any number of ways so that it never would have required disqualiﬁcation. Tennyson and Vizcarra-Millan discussed those options in open court. Finally, Vizcarra-Millan argues that he should not be blamed for having hired an unethical and unscrupulous law- yer. He was taken advantage of, he says, and did not waive his rights or seek to revoke that waiver in bad faith. He would rather we focus on Tennyson’s misconduct in soliciting joint representation without obtaining written waivers as required by the Indiana Rules of Professional Conduct. He did so while reaching out directly to represented clients, again in violation of the Indiana Rules of Professional Conduct. He was held in contempt of court for failing to abide by court orders. His li- cense to practice law was brieﬂy suspended in Tennessee. This is a troubling story, but it should not distract us from Vizcarra-Millan’s conduct. The district court inferred (and he admitted) that from the very early stages of this case, he de- liberately tried to sow ambiguity about his waiver. He then continued to use Tennyson’s potential conﬂict as a wedge to 38 Nos. 19-3476, et al. secure a better plea deal. Whether Tennyson committed fur- ther ethical misconduct, whether his performance fell below the constitutional minimum, and whether Vizcarra-Millan might have been prejudiced by any such failure are questions that are not properly before us. We aﬃrm Vizcarra-Millan’s conviction. III. Fourth Amendment Issues The Fourth Amendment provides: The right of the people to be secure in their per- sons, houses, papers, and eﬀects, against unrea- sonable searches and seizures, shall not be vio- lated, and no warrants shall issue, but upon probable cause, supported by oath or aﬃrma- tion, and particularly describing the place to be searched, and the persons or things to be seized. Defendants Atwater, Beasley, and Moseby all ﬁled late mo- tions to suppress evidence gathered against them, each claim- ing that a diﬀerent facet of the Fourth Amendment was vio- lated. Atwater’s house was searched pursuant to a warrant. He argues that the warrant was not issued “upon probable cause.” Beasley argues that the pretextual traﬃc stop leading to discovery of methamphetamine on his person was an “un- reasonable seizure” because the oﬃcers did not actually ob- serve a traﬃc violation. Moseby’s cell phones were seized and searched pursuant to a warrant authorizing federal agents to seize and search all cell phones found in Grundy’s clubhouse. Moseby argues that, given the unique privacy interests in cell phones, the warrant was too broad to justify seizure and search of his cell phones. Nos. 19-3476, et al. 39 A. Timeliness of the Motions and Beasley’s Motion to Suppress The government asks us to aﬃrm the district court’s deni- als of all three motions as untimely. Per the district court’s scheduling order, motions to suppress were due in February 2019. All three defendants’ motions were ﬁled months later, two in the days preceding trial and the third actually during trial. “If a party does not meet the deadline for making a [mo- tion to suppress], the motion is untimely. But a court may con- sider the defense, objection, or request if the party shows good cause.” Fed. R. Crim. P. 12(c)(3). “If the defendant also fails to ﬁle a motion for relief showing good cause before the district court, then we apply a hyper-deferential standard of review in which we examine whether, if a motion for relief had been made and denied, the district court would have abused its discretion in concluding that the defense lacked good cause.” United States v. Adame, 827 F.3d 637, 647 (7th Cir. 2016) (quotation marks and citation omitted). At least two, if not all three, motions to suppress could have been denied on procedural grounds as untimely under Rule 12. Neither Atwater nor Beasley tried to show good cause before the district court. Moseby’s late motion tried to estab- lish good cause, and the district court did not rule on that question. Although Rule 12 is mandatory, it is not jurisdic- tional, United States v. Nixon, 901 F.3d 918, 921 (7th Cir. 2018), and we may aﬃrm the district court on any ground supported by the record. United States v. Harden, 893 F.3d 434, 451 (7th Cir. 2018). Because the district court addressed the merits of 40 Nos. 19-3476, et al. each defendant’s suppression motion, we do the same for At- water’s and Moseby’s respective motions. 6 Beasley’s motion, however, presented a question of fact, and the district court did not hold a hearing with in-person testimony. Recall that Susan Koch and Beasley drove in her car to buy some methamphetamine. Local police knew about the deal and stopped their car soon after the deal had been completed. Beasley’s motion to suppress asserted that the of- ﬁcers did not have reasonable suspicion for the seizure be- cause they did not actually observe a traﬃc violation. See, e.g., United States v. Simon, 937 F.3d 820, 829 (7th Cir. 2019). He then argued that the 120 grams of highly pure methampheta- mine that were found on him and served as the basis of his conviction for Count 16 should have been suppressed. Beasley’s motion thus attacked the credibility of the oﬃc- ers who made the stop. One group of oﬃcers said Koch swerved out of her lane over a double yellow line. Another group of oﬃcers said Koch drove 50 miles per hour in a 35 mile-per-hour zone. Beasley claimed that a hearing would show that the stop violated the Fourth Amendment. The gov- ernment said that Koch was simply driving erratically and two sets of oﬃcers saw her commit diﬀerent violations in dif- ferent places. So did the oﬃcers actually see one or both violations? Were they honestly mistaken? Was one group of oﬃcers ly- ing? Were both? Neither? These questions are fact-intensive and would have required weighing witness credibility. E.g., 6Ruling on the substance of such motions can avoid later claims for ineffective assistance of counsel premised on motions that would not have prevailed. Nos. 19-3476, et al. 41 United States v. Rodriguez-Escalera, 884 F.3d 661, 672 (7th Cir. 2018) (aﬃrming grant of motion to suppress based on credi- bility ﬁndings). The district court denied Beasley’s motion on the twin grounds that it was untimely and that the motion did not cleanly tee up a material factual dispute that would justify a hearing at that late juncture. Even on appeal, Beasley does not identify good cause for his late motion, and no justiﬁcation is apparent from the rec- ord. We accordingly aﬃrm the district court’s denial of his motion to suppress as untimely. See Adame, 827 F.3d at 647. We do not reach the merits or the need for a hearing. B. Atwater’s House Atwater argues that the evidence of drug possession and distribution uncovered during a search of his home should be suppressed because the warrant justifying the search was not supported by probable cause. The warrant application pro- vided facts showing that Atwater was a drug dealer, that a speciﬁc residence was likely his house, and that based on those facts, in the experience of the aﬃant, it was likely that he had drugs or evidence of dealing in the home. Atwater claims that the evidence showing that the house was his was too sketchy and that any inference that he kept drugs or other evidence of dealing there was too speculative. An aﬃdavit submitted in support of a warrant application “need only contain facts that, given the nature of the evidence sought and the crime alleged, allow for a reasonable inference that there is a fair probability that evidence will be found in a particular place.” United States v. Aljabari, 626 F.3d 940, 944 (7th Cir. 2010). Probable cause is a common-sense standard, 42 Nos. 19-3476, et al. and we give deference to an issuing judge’s assessment. E.g., United States v. Bacon, 991 F.3d 835, 840 (7th Cir. 2021). We have recognized that “[i]n the case of drug dealers, ev- idence is likely to be found where the dealers live.” United States v. Lamon, 930 F.2d 1183, 1188 (7th Cir. 1991), quoted in United States v. Zamudio, 909 F.3d 172, 176 (7th Cir. 2018). Thus, where an aﬃdavit oﬀers reasonable grounds to believe the target is a drug dealer and that the property to be searched is his or her residence, we defer to an issuing judge’s determi- nation that the warrant application established probable cause to search the residence. Zamudio, 909 F.3d at 176–77. Here, the warrant aﬃdavit said that Atwater bought wholesale quantities of methamphetamine several times in quick succession. He bought, for example, two ounces on July 27, 2017; one ounce on August 6, 2017; and twelve ounces on August 10, 2017. These were quantities of methamphetamine consistent with retail drug distribution. Atwater alleges that the information was stale because these purchases were three months old. It is true that a warrant based on the prior pos- session of consumable or fungible goods may be stale if too much time has passed before the search is executed. But here, the repeated and large purchases suggested ongoing activity, i.e., retail drug distribution. Given the aﬃant’s experience that retail drug dealers keep drugs in their home, that was enough to support a ﬁnding of probable cause. Id. at 176–77; see also United States v. Pless, 982 F.2d 1118, 1126 (7th Cir. 1992) (three-month-old information about methamphetamine laboratory was not stale, and even if it had been, oﬃcers relied on warrant in good faith). That brings us to Atwater’s second issue: the evidence that he lived at the place to be searched. The warrant application Nos. 19-3476, et al. 43 explained that he subscribed for utility services at the address, that cell-site location information placed him (or at least his cell phone) there a majority of the time, and that surveillance placed him there twice in the week preceding the warrant ap- plication. That was enough to support the magistrate judge’s common-sense determination that there was a reasonable probability that Atwater lived there and that the police would ﬁnd drugs or other evidence of drug-dealing there. C. Moseby’s Cell Phones Moseby’s cell phones were seized during a search of one of Grundy’s properties, the “clubhouse” for his gang. Moseby claims that the search warrant was overbroad in allowing the police to seize every cell phone found inside a residence based on the conclusory assertion that the house was a “clubhouse.” He also claims that once the oﬃcers ascertained that the phones were not Grundy’s, they should have stopped search- ing them without obtaining another warrant speciﬁc to those phones. We conclude that even if we assume that the warrant was overbroad, however, the oﬃcers were entitled to rely on it in good faith because it was not facially defective. The Fourth Amendment requires that warrants be sup- ported by probable cause and that they describe with partic- ularity the places and objects to be searched and seized. These requirements are related. The particularity requirement “en- sures that the scope of a search will be conﬁned to evidence relating to a speciﬁc crime that is supported by probable cause.” United States v. Vitek Supply Corp., 144 F.3d 476, 481 (7th Cir. 1998). Warrants that are overbroad, that is, that allow oﬃcers to search for items that are unlikely to yield evidence of the crime, violate the Fourth Amendment. Balancing these concerns is highly fact- and context-speciﬁc. A warrant that 44 Nos. 19-3476, et al. may be overbroad in one context may be suﬃciently speciﬁc where the oﬃcers have less reliable information about where, exactly, the evidence is likely to be found. Id.; see generally United States v. Bishop, 910 F.3d 335, 336 (7th Cir. 2018) (“Crim- inals don’t advertise where they keep evidence.”). “The fruits of a search based on an invalid warrant may be admitted at trial if the executing oﬃcer relied on the invalid warrant in good faith.” United States v. Orozco, 576 F.3d 745, 750 (7th Cir. 2009), citing United States v. Leon, 468 U.S. 897, 922 (1984). Where an oﬃcer goes through the eﬀort to secure a warrant, we presume the oﬃcer acted in good faith. Id., quoting United States v. Mykytiuk, 402 F.3d 773, 777 (7th Cir. 2005). The good-faith exception to the suppression rule does not apply, however, when, among other situations, “the aﬃ- davit is ‘bare bones’ or ‘so lacking in indicia of probable cause’ that reliance is unreasonable[] [or] when the warrant is fa- cially deﬁcient in that it fails to specify the place to search or the items to seize.” United States v. Glover, 755 F.3d 811, 819 (7th Cir. 2014), quoting Leon, 468 U.S. at 923. To succeed on his suppression motion, Moseby must show not only that the warrant that authorized the search of his phones was overbroad but that it was objectively unreasona- ble for an oﬃcer to think that the warrant was legitimate even though a judge had issued it. The search warrant in question authorized a search of one of Grundy’s residences. The war- rant allowed the executing oﬃcers to “search and seize” “Cel- lular telephones and all electronic data stored within the memory of the cellular telephones.” Neither the warrant nor the warrant application discussed Mr. Moseby or his phones. When the oﬃcers executed the warrant on November 17, 2017, Grundy and Moseby were present, along with three Nos. 19-3476, et al. 45 other co-conspirators. The oﬃcers found two cell phones in a bedroom, one on a bed, and the other on the ﬂoor next to it. They also found a bag containing a handgun and a receipt with Moseby’s name for one of the cell phones. (The phones were registered to the same number and had identical con- tent.) The phones both belonged to Moseby and contained in- criminating data that was the basis of much of the govern- ment’s case against him. Moseby makes several arguments. He argues that a war- rant allowing police oﬃcers to enter a suspect’s home and seize every phone is necessarily overbroad. He says that the warrant was obviously overbroad here because it allowed of- ﬁcers to continue searching the phones even after they ascer- tained that they were not the wiretap’s “target phones” and did not belong to Grundy. Moseby also asks us to apply a heightened standard of review because the warrant here con- cerned cell-phone searches, which he says are especially con- stitutionally sensitive. The government counters that the warrant was not over- broad because the target location was Grundy’s “clubhouse,” so that any phones found there were reasonably likely to have a nexus to the gang’s criminal activity. Also, the government points out, we must face the realities of drug-dealing and po- lice searches. Criminals do not neatly label which cell phone belongs to whom, and oﬃcers would be foolish to trust self- serving denials. Even if the warrant was defective, the gov- ernment says, oﬃcers relied on it in good faith. This warrant application was full of references to cell phones. It opened with a list of the fourteen “target phones” for the wiretaps. It then summarized selected calls for over seventy pages, not to mention other references to cell phones 46 Nos. 19-3476, et al. in other parts of the warrant. The application established that cell phones provided crucial infrastructure for Grundy’s drug network and that they likely captured additional criminal ac- tivities beyond those recorded by the wiretaps. The list of target phones also underscored that Grundy and his gang seemed to treat cell phones as fungible and tried to conceal the true owner of each phone. The list of target phones showed, for example, that at least some conspirators had multiple phones, each with a diﬀerent number. None of the listed phones was used by the person who had registered it. For many phones, subscriber information was unavailable. At least two of those that did list subscriber information used obviously fake names or contact information, such as Tyler Jones of “1234 abcfe ave” (used by Nathaniel Dixson) and “Hector FFFFFFFFF” (used by David Carroll). The warrant application was less speciﬁc about Grundy’s residences, though. The aﬃdavit explained that Grundy likely knew he was under surveillance and/or a target of rival gangs. Wary of surveillance, he covered his tracks and made it diﬃcult to discern where he would be or stay at any given time. Nonetheless, the warrant application identiﬁed the res- idence to be searched as the “clubhouse” for Grundy’s gang. The sub-bullet points under this assertion, however, had little to nothing to do with why the residence was a clubhouse. They mostly discussed Grundy’s family. Moseby’s overbreadth argument is that the warrant here was not “as speciﬁc as circumstances allowed.” Bishop, 910 F.3d at 338. He says that the warrant should have either lim- ited the phones to be seized to those that were used in drug traﬃcking or required oﬃcers to obtain a second warrant to search any phone that had been seized. These options are Nos. 19-3476, et al. 47 available and are often used by issuing judges. See, e.g., United States v. Eggerson, 999 F.3d 1121, 1127 (8th Cir. 2021) (“The second warrant, which police got after they seized Eg- gerson’s phone and which was limited to its contents, is even less suspect.”); United States v. Wagner, 951 F.3d 1232, 1247–48 (10th Cir. 2020) (search warrant implicitly required nexus be- tween computers to be seized and child pornography, and failure to specify ownership was immaterial because it was unknown which computers were used for child pornogra- phy); cf. United States v. Griﬃth, 867 F.3d 1265, 1277 (D.C. Cir. 2017) (where police oﬃcers intend to rely on a second warrant to search phones, the initial warrant must expressly limit agents’ authority to examine any electronic devices seized); Vitek Supply, 144 F.3d at 481 (distinguishing between legal and illegal items to be searched, but noting that in practice, such limits may have little practical eﬀect if oﬃcers cannot distin- guish between the two). Moseby is correct that the warrant’s failure otherwise to limit the types of phones to be seized opens the door for further mischief based on the conclusory assertion that the place to be searched was Grundy’s “club- house.” Moseby argues that these mutually reinforcing gaps are especially troubling because they allowed the oﬃcers to search the cell phone of every person who happened to be in the same house as Grundy. In Riley v. California, the Supreme Court held that police must obtain a warrant before searching a cell phone found on an arrestee during a search pursuant to arrest. 573 U.S. at 386. The Court noted people’s heightened privacy interests in cell phones relative to other objects that may be searched pursuant to arrest: “Cell phones diﬀer in both a quantitative and a qualitative sense from other objects that might be kept on an arrestee’s person.” Id. at 393. As 48 Nos. 19-3476, et al. compared to rummaging indiscriminately through a person’s home, the archetype of a search that violates the Fourth Amendment, “a cell phone search would typically expose to the government far more than the most exhaustive search of a house.” Id. at 396. The Riley opinion included sweeping language about the privacy implications of warrantless searches of cell phones, but the Fourth Amendment holding was relatively limited: “a warrant is generally required before such a search, even when a cell phone is seized incident to arrest.” Id. at 401. Or, more pithily, “get a warrant.” Id. at 403. Here, of course, the oﬃcers did so, and it expressly authorized searches of any cell phones found at the location. In support of his overbreadth argument, Moseby relies on United States v. Griﬃth, which held that the Leon good-faith exception did not apply when oﬃcers seized a cell phone pur- suant to a warrant based on an obviously overbroad and un- der-speciﬁc aﬃdavit. 867 F.3d 1278–79. In that case, the de- fendant was suspected of murder. Following the murder, he was incarcerated for another crime for a year. When he was released, the oﬃcers sought a warrant to search his house and seize, among other things, all electronics found there, includ- ing his cell phone. The aﬃdavit did not contain any infor- mation establishing a nexus between the murder and any electronics, let alone that Griﬃth even had a cell phone. The connection between the hypothetical phone and the crime (the year-old murder) was merely that most people own cell phones, most cell phones capture what their users do, Griﬃth is a person, we think he’s a criminal, so we think there exists a cell phone that captured evidence of a crime. The allegation that supposedly provided probable cause to search Griﬃth’s Nos. 19-3476, et al. 49 home for evidence in the ﬁrst place was that he had a hypo- thetical cell phone that contained hypothetical evidence. The appellate court reversed the district court’s denial of Griﬃth’s suppression motion: “oﬃcers seeking authority to search a person’s home must do more than set out their basis for suspecting him of a crime.” Id. at 1279. The overbreadth of the warrant was especially striking in that case because it al- lowed the seizure and search of all electronic devices, and it was unlikely that Griﬃth’s current phone would have con- tained incriminating evidence for a crime that had occurred over a year earlier. Id. at 1278–79. In other words, a reasonable oﬃcer should have known that the cell phone request was ﬁshy, that the authorization to seize and search every elec- tronic device in Griﬃth’s house (to which he had just returned after prison) was overbroad, and that the justiﬁcation for the search was circular. The warrant application here was substantially stronger. See Eggerson, 999 F.3d at 1126 & n.3 (distinguishing Griﬃth). The central role of cell phones to Grundy’s gang is apparent throughout the 170-page search warrant application. As noted, the application showed that members of Grundy’s crew used multiple phones that were unregistered or regis- tered to others. Unlike Griﬃth, there was an obvious likely connection between the crime and the phones, and there was good reason to think that Grundy and his crew had multiple phones that would be diﬃcult to identify as being used by particular individuals. As the Supreme Court instructed in Ri- ley, law enforcement here obtained a warrant—and one that speciﬁcally authorized seizure and search of cell phones. Based on these facts, a warrant authorizing the seizure and search of any phone found in what was thought to be 50 Nos. 19-3476, et al. Grundy’s clubhouse, regardless of its registration, was not so overbroad as to defeat the good-faith exception. It would have been possible to frame the warrant request more narrowly, but that is not the good-faith standard. Grundy and his crew’s extensive use of cell phones, combined with their attempts to cover their tracks, established enough of a nexus that a war- rant authorizing seizure and search of all phones in their club- house, regardless of ownership, was not so obviously invalid that an agent or oﬃcer should have refused to execute it. This logic emphasizes the assertion that the property was the dealers’ clubhouse as opposed to an ordinary residence. We agree that, as in Griﬃth, a search warrant may be over- broad by authorizing without probable cause the seizure of all electronic devices in a location. Griﬃth, 867 F.3d at 1271-76. Moseby argues that the “clubhouse” assertion was too con- clusory to support such a sweeping search and that the police oﬃcers should have known that this was too “bare bones” to support a warrant. Moseby is correct that there are no other speciﬁc factual allegations (such as observations of other gang members entering or leaving) indicating that the residence was in fact a clubhouse. In support he relies on United States v. Koerth, 312 F.3d 862, 867, 870 (7th Cir. 2002), which criticized a three-paragraph warrant application as conclusory, but we ultimately found that the evidence in Koerth was properly admitted under the good faith exception because the warrant’s defects were not so obvious that a reasonable oﬃcer could not have relied on it in good faith. The line between a conclusory allegation and factual matter can be diﬃcult to draw, even for experienced lawyers and judges. Under Leon, law enforcement oﬃcers do not need to patrol that line for magistrate judges. The warrant Nos. 19-3476, et al. 51 application here could have been more speciﬁc, especially given the heightened constitutional stakes of searching mul- tiple phones, but that lack of speciﬁcity does not require sup- pression of this evidence. Finally, Moseby argues that the oﬃcers should have known that the warrant was overbroad because it allowed them to search his phone even after they ascertained that it did not belong to Grundy. This argument misunderstands the good-faith question. The warrant authorized the oﬃcers to search every phone found in Grundy’s clubhouse regardless of who owned it or claimed to own it. A search of a phone they found there was not outside the scope of the warrant, even if it was apparent that the phone did not belong to Grundy. We aﬃrm the denial of Moseby’s motion to suppress and his conviction. IV. Suﬃciency of the Evidence Defendants Atwater, Neville, and Beasley all contend that no rational jury could have convicted them and that the dis- trict court should have granted their Rule 29 motions for judg- ment of acquittal. All three contest their convictions on Count 1 for conspiracy to distribute controlled substances, 21 U.S.C. §§ 846 & 841(a)(1). Neville also contests his conviction on Count 24, conspiracy to launder monetary instruments in vi- olation of 18 U.S.C. § 1956(h), and Beasley contests his convic- tion on Count 17, constructive possession of methampheta- mine at Koch’s house in violation of 21 U.S.C. § 841(a)(1). For challenges based on the suﬃciency of the evidence, we review the evidence in the light most favorable to the govern- ment. United States v. Bey, 725 F.3d 643, 649 (7th Cir. 2013). Our review is deferential but not a rubber-stamp of aﬃrmance. We 52 Nos. 19-3476, et al. have sometimes said that a defendant seeking a judgment of acquittal for insuﬃcient evidence faces a “nearly insurmount- able hurdle,” e.g., United States v. Johnson, 874 F.3d 990, 998 (7th Cir. 2017) (collecting cases), but the height of the hurdle the defendant must overcome depends directly on the strength of the government’s evidence. United States v. Garcia, 919 F.3d 489, 496–97 (7th Cir. 2019), citing United States v. Jones, 713 F.3d 336, 339 (7th Cir. 2013). A. Conspiracy or Buyer-Seller? We consider ﬁrst Count 1, challenged by all three of these appellants. There is no doubt there was a large conspiracy, but we must address this charge one defendant at a time. Charges for conspiracy to distribute drugs hold “a unique position in our legal suﬃciency jurisprudence.” United States v. Pulgar, 789 F.3d 807, 812 (7th Cir. 2015). The government must prove beyond a reasonable doubt that the defendant knowingly agreed, perhaps implicitly, with someone else to distribute drugs. Conspiracies must be distinguished from buyer-seller relationships. An agreement to buy drugs (even a large quan- tity) is not necessarily an agreement to join a drug distribution conspiracy. United States v. Johnson, 592 F.3d 749, 754 (7th Cir. 2010). To be liable for conspiracy, a defendant must have “‘a stake in the venture’ and therefore exhibit[] ‘informed and in- terested cooperation.’” United States v. Brown, 726 F.3d 993, 998 (7th Cir. 2013). Distinguishing an agreement to distribute drugs from a buyer-seller relationship can be challenging for judges and ju- ries. The issue is often whether circumstantial evidence al- lows a jury to infer beyond a reasonable doubt that the de- fendant agreed to enter the conspiracy. Generally speaking, circumstantial evidence of an agreement to distribute drugs Nos. 19-3476, et al. 53 may fail to rule out the alternative plausible inference that there is just a slightly more formalized buyer-seller relation- ship. Johnson, 592 F.3d at 755. Where the jury is left with two equally plausible inferences from the circumstantial evidence, guilty or not guilty, it must necessarily entertain a reasonable doubt. Id. (collecting cases). We use a “totality-of-the-circumstances approach in these cases.” Pulgar, 789 F.3d at 813. Nonetheless, there are some rules of thumb that can help us distinguish a conspiracy from a normal buyer-seller relationship: “sales on credit or con- signment, an agreement to look for other customers, a pay- ment of commission on sales, an indication that one party ad- vised the other on the conduct of the other’s business, or an agreement to warn of future threats to each other’s business stemming from competitors or law-enforcement authorities.” Johnson, 592 F.3d at 755–56. We have sometimes described these factors as supporting an inference of heightened trust, but evidence of mutual trust alone is insuﬃcient. See Pulgar, 789 F.3d at 815–16. Where a person repeatedly buys large quantities of drugs on credit, a jury can infer that the person entered into the con- spiracy. Brown, 726 F.3d at 1002. Beyond that, charting the di- viding line between a buyer-seller relationship and a conspir- acy is as much an art as a science. See Pulgar, 789 F.3d at 813; see also United States v. Colon, 549 F.3d 565, 567–71 (7th Cir. 2008) (collecting cases in this and other circuits on distin- guishing buyer-seller relationships and conspiracies). For ex- ample, the district court’s jury instructions here tracked the Seventh Circuit Pattern Instructions: A conspiracy requires more than just a buyer- seller relationship between a defendant and 54 Nos. 19-3476, et al. another person. In addition, a buyer and seller of controlled substances do not enter into a con- spiracy to distribute controlled substances simply because the buyer resells the controlled substance to others, even if the seller knows that the buyer intends to resell the controlled sub- stance. To prove a conspiracy, the government must prove that a buyer and seller had the joint criminal objective of further distributing con- trolled substances to others. See The William J. Bauer Pattern Criminal Jury Instructions of the Seventh Circuit, § 5.10(A) (2020 Ed.). These can be ﬁne hairs to split, but the government carries the burden of producing ev- idence that, if believed, would allow a factﬁnder to make this distinction. We now turn to each of these three appellants’ suf- ﬁciency challenges. B. Atwater The evidence that Derek Atwater was part of the Grundy conspiracy was suﬃcient. The evidence at trial, taken in the light most favorable to the government, showed that he en- gaged in several large-quantity transactions on credit, which strongly suggests his agreement to distribute the drugs. The primary testimony against Atwater at trial came from David Carroll, one of Grundy’s distributors who sold drugs to Atwater. Carroll testiﬁed extensively as to Atwater’s drug purchasing habits, telling the jury that Atwater bought one- ounce quantities of methamphetamine regularly over a sus- tained period, and walking the jury through several sales. Carroll estimated that Atwater bought methamphetamine a “couple times a week” over the course of four or ﬁve months. Nos. 19-3476, et al. 55 He also testiﬁed that “a little under … half” the time, Atwater purchased drugs on credit. Other evidence indicated an agreement to distribute the drugs, such as Atwater’s request that Carroll package the drugs separately, but the credit sales were the most relevant factor. If the jurors believed Carroll’s testimony—and we must presume they did—then they could infer beyond a reasonable doubt that Atwater agreed to distribute drugs on behalf of Carroll. Given their history of wholesale-quantity deals on credit, a reasonable jury could have rejected the alternative inference that Atwater and Carroll had only a buyer-seller re- lationship. Brown, 726 F.3d at 1002; see also Colon, 549 F.3d at 569–70 (purchases on credit can distinguish buyers from co- conspirators). We aﬃrm Atwater’s conviction. C. Neville Neville challenges two of his convictions, for the money- laundering conspiracy and the larger drug-traﬃcking con- spiracy. We look ﬁrst at the evidence of money laundering. The conviction relates to the $84,500 in cash that Grundy had sent to Vizcarra-Millan to purchase drugs and that was seized by law enforcement in August 2017. The factual question that Neville disputes is whether the evidence showed that he had contributed to that pool. He does not dispute that the other elements of conspiracy to launder monetary instruments were met. The evidence at trial showed that Grundy pooled cash from various lieutenants to purchase larger quantities of drugs. The cash seized in the August 2017 bust came from “ten, twelve, twenty, [or] ﬁfteen” of Grundy’s associates. The government also oﬀered evidence that Neville contributed to 56 Nos. 19-3476, et al. a drug distribution ring that he called the “dub club.” Reserv- ing for a moment the question whether Neville was a co-con- spirator or merely a buyer, the government oﬀered evidence of a sustained drug-dealing relationship between Grundy and Neville. One witness, Carroll, testiﬁed that he thought Neville was present during a meeting among people who lost their money in the seizure. Finally, soon after Grundy’s cash was seized, Neville complained to a conﬁdential informant that he just “took a thirteen thousand dollar loss.” The jury could reasonably convict. The evidence of Ne- ville’s participation in Grundy’s drug pool was, with the ex- ception of Carroll’s testimony, circumstantial, but the inferen- tial links the jury had to make to convict were all reasonable. Most of Neville’s arguments stray beyond our standard of re- view on appeal. He says that the jury should not have con- cluded that his statement that he “lost” money referred to the seizure of the $84,500. Perhaps he misplaced another $13,000? He also says that Carroll’s testimony was vague and had little corroboration. But we review the evidence in the light most favorable to the government. The inference that Neville lost his $13,000 to government agents who had just seized his as- sociate’s cash was reasonable, and Carroll’s testimony was not so fantastic that we could treat it as incredible as a matter of law. E.g., United States v. Cherry, 920 F.3d 1126, 1139 (7th Cir. 2019) (“To ﬁnd a witness’s testimony to be incredible as a matter of law, it must have been physically impossible for the witness to have observed that which he claims occurred, or impossible under the laws of nature for the occurrence to have taken place at all.”) (cleaned up). We turn next to Neville’s Count 1 conviction for conspir- acy to distribute drugs. As with Atwater, considerable Nos. 19-3476, et al. 57 evidence showed that Neville purchased large quantities of drugs from Grundy. The question is whether the govern- ment’s evidence proved that he was a member of the conspir- acy. The government did not argue that the evidence of money laundering is the “something more” that could sustain the conspiracy conviction, so we do not rest our holding on that basis. But see United States v. Hopper, 934 F.3d 740, 757 (7th Cir. 2019) (“The Government also presented considerable testimony establishing that Mr. Hopper pooled his resources with other members of the conspiracy to obtain methamphet- amine at a reduced cost from out-of-state sources. This is cir- cumstantial evidence of an agreement to distribute drugs….”). Ample evidence allowed the jury to ﬁnd that Neville was a member of Grundy’s conspiracy. Neville was a wholesale buyer. He bought large quantities of drugs directly from Grundy and Grundy’s business partner, Emilio Mitchell II. Other evidence tied the two even more closely. Grundy and Neville once had an extended phone conversation about an- other Grundy subordinate and whether he was up to the chal- lenges of dealing drugs. Neville also helped other lower-level members of Grundy’s crew on several occasions. He cross-re- ferred customers to others in Grundy’s orbit a couple of times, and he stored drugs for another of Grundy’s dealers. Viewing the evidence in the light most favorable to the government, this was suﬃcient evidence of intra-conspiracy cooperation. See United States v. Maldonado, 893 F.3d 480, 485 (7th Cir. 2018) (cooperation between defendant and conspirators supported his conspiracy conviction). Put diﬀerently, the jury was enti- tled to reject the possibility that Neville aﬀorded these profes- sional courtesies to competitors rather than to co-conspira- tors. 58 Nos. 19-3476, et al. More evidence also tied Neville to the conspiracy. Early on in the conspiracy’s timeline, Neville was caught in Albu- querque smuggling cocaine across the country. He concedes that taking the evidence in the light most favorable to the gov- ernment, it showed that he was acting as a courier for Grundy. He also bought pound quantities of methamphetamine on credit several times. See United States v. Moreno, 922 F.3d 787, 795–96 (7th Cir. 2019) (multiple large purchases on credit sup- ported guilty verdict on conspiracy charge, especially when paired with evidence of cooperation). This mosaic of evidence was enough for a reasonable jury to reject the alternative hypothesis that Neville was a mere buyer without any additional stake in Grundy’s enterprise. The courier work alone may be enough to rule out an alterna- tive buyer-seller hypothesis. See United States v. Salinas, 763 F.3d 869, 877–78 (7th Cir. 2014) (aﬃrming courier’s conspiracy conviction). In addition, Neville time and again helped with the internal aﬀairs of the conspiracy, providing a sounding board for Grundy, helping other dealers, and buying pound quantities of methamphetamine on credit. This all provided ample evidence that Neville was part of the conspiracy. See Moreno, 922 F.3d at 795–96 (circumstantial evidence sup- ported conspiracy conviction); cf. United States v. Johnson, 592 F.3d 749, 755–56 (7th Cir. 2010) (circumstantial evidence was inadequate to support conspiracy conviction where it failed to distinguish buyer-seller relationship). Neville attempts to poke holes in various pieces of evi- dence, but the standard of review is deferential to jury ver- dicts. For example, he points out that the government did not corroborate testimony with cell-site location data or wiretaps. That is an argument for a jury, not a court of appeals. The Nos. 19-3476, et al. 59 government is not required to put on all conceivable or even all available corroborating evidence, lest jury trials become even longer and more complex. We aﬃrm Neville’s convic- tions. D. Beasley Beasley was convicted of three oﬀenses: Count 1, for con- spiring with the Grundy gang to distribute drugs; Count 16, for possessing the methamphetamine found on his person af- ter Susan Koch’s car was pulled over and he was searched; and Count 17, for constructive possession of the methamphet- amine recovered from Koch’s home after she told the oﬃcers to search there. Beasley does not challenge the suﬃciency of the evidence on Count 16, the 120 grams of methamphetamine found on his person in the traﬃc stop. He challenges the other two con- victions. First, he claims that his limited involvement with Carroll was perfectly consistent with a buyer-seller relation- ship, so that a reasonable jury would necessarily have enter- tained a reasonable doubt as to whether he joined the conspir- acy. He is correct. The evidence was consistent with both a buyer-seller relationship and nascent involvement in the con- spiracy. The jury had no basis to choose the conspiracy ﬁnd- ing beyond a reasonable doubt. The jury should not have con- victed him on that count. Second, he argues that the evidence at trial did not establish that he possessed the methampheta- mine found at Koch’s home. He is correct that the admissible evidence did not support a ﬁnding of that connection beyond a reasonable doubt. The government’s evidence supporting the conspiracy charge against Beasley was notably weaker than for Neville 60 Nos. 19-3476, et al. or Atwater. Beasley bought drugs from Carroll several times over the course of a few weeks. The quantities rose from one ounce to two ounces to seven ounces. There was evidence suggesting that Carroll gave him a $100 discount or credit on one transaction. And during one of the intercepted calls played at trial, Carroll was heard praising Beasley for being a good drug salesman. The evidence was consistent with an inference that Beasley joined the conspiracy at an inopportune moment, just as federal agents were closing in. But the evidence was at least equally consistent with an inference that Beasley and Carroll had only a buyer-seller relationship, with Carroll’s one-time discount and encouragement showing only the work of a good wholesale salesman trying to develop a proﬁtable rela- tionship with a new retailer-customer. Consider the $100 discount or credit. The evidence did not make clear whether this was a credit, a discount, or a negoti- ated lower price. The government’s primary evidence that Carroll once ﬂoated Beasley $100 is a wiretap in which Carroll agreed to sell Beasley two ounces of methamphetamine for $700 when the going rate would have been $800. At trial, Car- roll described this as a partial front of $100, but he later con- ceded that he did not extend credit to Beasley. We do not need to determine whether the $100 is better described as a front or a discount. Either way it was extremely weak evidence of conspiracy. See Johnson, 592 F.3d at 756–57 (rejecting government’s attempt to characterize a negotiation as a credit and dismissing idea that a $30 discount could sup- port inference of conspiracy at all); United States v. Pulgar, 789 F.3d 807, 814 (7th Cir. 2015) (vague evidence of a couple of fronts insuﬃcient to sustain conspiracy conviction). Even if Nos. 19-3476, et al. 61 this was a one-time $100 credit, our case law is clear that “oc- casional sales on credit are consistent with an ordinary buyer- seller relationship,” and we must view those credits in con- text. United States v. Neal, 907 F.3d 511, 516 (7th Cir. 2018) (quotation marks and citation omitted). In the context of this new wholesaler-retailer relationship, Carroll’s small, one- time credit looks too much like a courtesy for a new customer to support an inference beyond a reasonable doubt of an agreement between the two to distribute drugs further. See United States v. Brown, 726 F.3d 993, 999–1000 (7th Cir. 2013) (distinguishing credit sales from consignment and explaining that small and infrequent credits are consistent with a buyer- seller relationship); see also Moreno, 922 F.3d at 794–96 (evi- dence of two very large credit sales supported conspiracy conviction within context of year-long wholesale drug-selling relationship and evidence of cooperation). The other wisps of evidence are no stronger. Beasley in- creased his purchases from one ounce to two to seven during the weeks he bought from Carroll. That is not evidence of con- spiracy. Context matters. In some cases, a buyer choosing to purchase greater quantities of drugs can signify that he has thrown his lot behind the conspiracy. Here the government has certainly shown that Beasley took to drug dealing, but Count 1 required more. The government needed to show be- yond a reasonable doubt that Beasley agreed to distribute drugs for Carroll (which originated with Grundy). Buying multiple large quantities over a long period of time is not by itself suﬃcient to show such an agreement. Moreno, 922 F.3d at 794, citing Maldonado, 893 F.3d at 485. The same goes for buying an ounce or two a few times and larger amounts a few times. The evidence of growing trust between Carroll and Beasley was consistent with a buyer-seller relationship and 62 Nos. 19-3476, et al. therefore could not carry the government’s case over the ﬁn- ish line. E.g., Brown, 726 F.3d at 998–99 (discussing Colon’s re- jection of “mutual trust” as a potentially dispositive factor). The ﬁnal piece of conspiracy evidence that the govern- ment cites is Carroll’s encouragement. He congratulated Beasley on his entrepreneurial success and urged him to sell more drugs: “You're clowning. I ain’t going to lie to you … I am proud of you … Just stay motivated. Just stay wanting more. Don’t get comfortable. You got to want more.” We have said that “‘stimulation, instigation,’ or ‘encouragement’” can indicate an agreement to join the conspiracy where the de- fendant attempts to stimulate interest in the broader conspir- acy. United States v. Colon, 549 F.3d 565, 568 (7th Cir. 2008), quoting Direct Sales Co. v. United States, 319 U.S. 703, 713 (1943). But here, there is no evidence that Beasley was encour- aging Grundy or Carroll. Rather, on one occasion, Carroll en- couraged Beasley. A seller’s one-time, unspeciﬁc encourage- ment of a buyer is not enough to prove beyond a reasonable doubt that the buyer agreed to distribute drugs for the seller. See, e.g., Colon, 549 F.3d at 570 (mutual understanding of intent to resell is not “germane” to whether there was a conspiracy); see also Pattern Instruction § 5.10(A) (“a buyer and seller of [name of drug] do not enter into a conspiracy to []distribute [name of drug] … simply because the buyer resells the [name of drug] to others, even if the seller knows that the buyer in- tends to resell the [name of drug].”). Carroll encouraged a good customer to keep up his pur- chases and to keep doing a good job. In the non-criminal econ- omy, such praise and encouragement are routine, without turning sellers and buyers into joint venturers. See Pulgar, 789 F.3d at 815–16 (evidence consistent with seller being a good Nos. 19-3476, et al. 63 salesperson cannot rule out buyer-seller relationship), citing Ruth Klein, The Everything Guide to Being a Sales Rep: Winning Secrets to a Successful—and Proﬁtable—Career! 176 (2006) (“Suc- cessful sales professionals often ﬁnd that they and their cus- tomers become good friends.”). In sum, the evidence against Beasley did not allow a ﬁnding beyond a reasonable doubt of more than a buyer-seller relationship, so we must reverse his conviction on Count 1. Finally, we turn to Beasley’s conviction on Count 17 for possession of the three ounces of methamphetamine found at Koch’s house. For this count, the government relied on a the- ory of constructive possession in which an individual is deemed to possess contraband items without a showing of immediate, physical control of the objects. E.g., United States v. Schmitt, 770 F.3d 524, 534 (7th Cir. 2014). “Constructive pos- session may be established by demonstrating that the defend- ant knowingly had both the power and the intention to exer- cise dominion and control over the object, either directly or through others.” United States v. Griﬃn, 684 F.3d 691, 695 (7th Cir. 2012). In more practical terms, the “owner of an automo- bile possesses it even when it is parked in a garage and he is miles away. A tenant possesses the apartment he has rented even when he is away on a trip.” United States v. Brown, 724 F.3d 801, 805 (7th Cir. 2013). The government’s theory was that Beasley had recently bought two ounces of methamphetamine from Carroll in front of a clothing store called Hangtime. After Susan Koch was pulled over with Beasley as her passenger, she told the police that he had spent the night at her house. During a con- sensual search of her bedroom, she told the oﬃcers that Beasley had left something at her house, and she directed 64 Nos. 19-3476, et al. them to a shoebox next to a bag from Hangtime. The shoebox contained three bags of methamphetamine, totaling three ounces. Soon after the arrest, Koch also told a member of the larger conspiracy, Robert Lisenby, that Beasley stored meth- amphetamine at her house and that the police had found it, and Lisenby relayed that conversation to co-conspirator Car- roll. The government did not oﬀer any evidence that Beasley had a possessory interest in Koch’s home. The only evidence of any kind of relationship between Beasley and Koch was that they drove together, that Koch told the oﬃcer that Beasley had spent the night at her house, and that Lisenby suggested the two dated intermittently. If Koch had testiﬁed at trial that the methamphetamine was Beasley’s, the conviction would stand, of course. The problem is that Koch did not testify at trial. Others’ reports of her statements blaming Beasley for the drugs found in her home were not admissible as evidence against him. When a police oﬃcer started to testify about Koch’s statements, Beasley properly objected that they were hearsay, that is, out- of-court statements oﬀered to prove the fact of the matter as- serted. See Fed. R. Evid. 801(c); Crawford v. Washington, 541 U.S. 36 (2004) (admission of testimonial hearsay violates Con- frontation Clause of the Sixth Amendment). Koch’s out-of- court statements could not be used to prove that Beasley had spent the night or that he controlled the drugs found in her bedroom. The district court allowed the introduction of Koch’s state- ment to the oﬃcers as “course of investigation” evidence. See United States v. Law, 990 F.3d 1058, 1063 (7th Cir. 2021) (aﬃrm- ing admission of cumulative course-of-investigation testi- mony that explained how law enforcement zeroed in on a Nos. 19-3476, et al. 65 sophisticated, intercontinental sex-traﬃcking ring). The court also properly issued a limiting instruction: the oﬃcer could relay Koch’s statements about Beasley having left something at her house, but the jury could consider the statements only for their value in explaining why the oﬃcer interrupted a traf- ﬁc stop to drive to Koch’s house and search her bedroom. It could not consider it as direct evidence of Beasley’s posses- sory interest. 7 7 It is not apparent to us why the government would need to explain that interruption, except as a means of introducing hearsay. Such risky uses of course-of-the-investigation testimony show why we “are reluctant to permit ‘course of the investigation’ rationale for fear of its abuse or mis- use.” Law, 990 F.3d at 1063, quoting United States v. Marchan, 935 F.3d 540, 546 (7th Cir. 2019); see also Carter v. Douma, 796 F.3d 726, 736 (7th Cir. 2015) (affirming denial of state prisoner’s habeas petition because he could not show prejudice under Strickland: “The problem, as we have explained time and again, is that the ‘course of investigation’ gambit is so often abused and/or misunderstood that it is an evidentiary and constitutional minefield.”) (citations omitted); Jones v. Basinger, 635 F.3d 1030, 1046 (7th Cir. 2011) (granting habeas corpus relief to state prisoner: “While such ‘course of investigation’ evidence usually has little or no probative value, the dangers of prejudice and abuse posed by the ‘course of investigation’ tactic are significant.”). The course-of-investigation statements here were particularly dubious. The police had just noted that Koch’s car smelled like marijuana and had drug-dealing paraphernalia strewn about it. We doubt that a jury would have wondered why they then searched Koch’s house. There was no suggestion that “the officers were officious intermed- dlers staking out [Koch] for nefarious purposes.” United States v. Silva, 380 F.3d 1018, 1020 (7th Cir. 2004) (reversing conviction and remanding for new trial). In addition, Koch had obvious incentives to shift responsibility for the drugs in her bedroom, offering substantial grounds for cross-ex- amination if she had testified. The “course-of-investigation” gambit ap- pears to have been used to admit her statement, albeit with a limiting in- struction, without requiring her testimony in court. 66 Nos. 19-3476, et al. Likewise, Koch’s statements to Lisenby were also hearsay, but Beasley did not object and therefore forfeited the argu- ment that Koch’s hearsay should not have come in. But Lisenby’s statements were admitted pursuant to the conspir- acy exception to the hearsay rule: at a co-conspirator’s trial, another co-conspirator’s out-of-court statements made in the course of and pursuant to the conspiracy may be admitted for the truth of the matter asserted. See Fed. R. Evid. 801(d)(2)(E). But for reasons explained above, Beasley was not shown to have been a member of a conspiracy with Lisenby, Carroll, or others. Accordingly, Lisenby’s gossip about what Koch had told him was not admissible as direct evidence against Beasley. Lisenby’s recorded statements are especially suspect given that they seem to have been used to launder the out-of- court statements by Koch, who as noted did not testify at trial. Without the two direct connections between Beasley and the methamphetamine, the evidence that Beasley construc- tively possessed the drugs found in Koch’s home was too sparse. The government showed only that Beasley bought drugs in front of a clothing store and that a diﬀerent quantity of drugs was later found in the house of someone he knew next to a bag from that clothing store. An inference that those were the drugs that Beasley purchased in front of Hangtime would have justiﬁed further investigation, but it was not enough to support guilt beyond a reasonable doubt. The convictions of Richard Grundy III, Derek Atwater, Un- drae Moseby, Gilberto Vizcarra-Millan, and Ezell Neville are AFFIRMED. James Beasley’s conviction on Count 16 is AFFIRMED, but his convictions for Counts 1 and 17 are REVERSED. Beasley’s case is REMANDED for resentencing on Count 16.

United States v. Undrae Moseby

Related Cases