G: [Non-text content: thoughts; content_type=thoughts]

G: Yes, you’re making sense. I’ll keep this at the “is the dataset even obtainable, and from where?” level rather than starting the actual aggregation.

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes. The question is coherent, but the clean dataset you want probably does not exist publicly.

What is publicly available:

1. DOJ / U.S. Attorney press releases
   Good for announced charges, indictments, settlements, raids, convictions. Bad for “opened investigations,” because many investigations are never announced. citeturn230997search10

2. U.S. Attorneys Annual Statistical Reports
   These give national/district caseload data, not a name-by-name list of targets. Useful for baseline comparison, not for your political-target question alone. citeturn230997search2

3. Court records / PACER / RECAP
   Good once charges, warrants, subpoenas, or litigation become public. Bad for quiet investigations.

4. FOIA
   Limited. DOJ can withhold law-enforcement records if release could interfere with proceedings, invade privacy, reveal sources/methods, etc. citeturn838792search2

5. News reporting / leaks / named-source reports
   Unfortunately necessary for “investigation opened” claims, because DOJ policy often discourages public confirmation before charges. DOJ’s media policy explicitly balances public access against fair trial rights, privacy, and the government’s ability to administer justice. citeturn838792search1

So the answer is: available online, yes; available as an official complete government list, no.

The more objective project would be framed as:

Build a database of publicly reported DOJ investigative/prosecutorial actions during Donald Trump’s second term, separating confirmed official actions from media-reported investigations, then classify each subject by relationship to Trump, legal basis, public evidence, DOJ source type, and procedural stage.

Important distinction:

“Opened investigation” is the hardest category.
“Charged / indicted / subpoenaed / searched / sued / settled / convicted” is much easier and more verifiable.

On legal/operational failure: I don’t see a general legal requirement that DOJ publish every opened investigation. In fact, normal DOJ policy and FOIA law often cut the other way: privacy, fair-trial concerns, and active-investigation exemptions justify nondisclosure. So failure to publish a complete list would not itself prove misconduct.

The real test would be comparative: official public actions + credible reported investigations + historical baseline.

G: [Non-text content: thoughts; content_type=thoughts]

G: You’re not asking for a rant-confirmation machine; you’re asking for a way to distinguish headline-level retaliation signals from the whole DOJ operating pattern. I’ll verify the immediate example first, then keep the answer focused on how that changes the shape of the research question.

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes — and your instinct to slow down is right.

The Carroll example is already messy: several outlets first reported a DOJ perjury probe into E. Jean Carroll, but the Chicago U.S. attorney then publicly denied that his office had “opened—and has never opened—a criminal investigation into E. Jean Carroll.” Current reporting says the probe is instead into a Reid Hoffman-linked nonprofit that helped fund her lawsuit, with Carroll’s testimony examined as part of that inquiry. citeturn366621news29turn366621news30

So the research should not be:

“Who has DOJ investigated?”

It should be:

“What DOJ actions or reported investigative activity appear connected to Trump’s personal enemies, critics, legal adversaries, donors, or political opponents — and how does that compare with ordinary DOJ activity during the same period?”

Key categories:

1. Official DOJ actions
Charges, indictments, civil suits, subpoenas that become public, search warrants, settlements, press releases.

2. Reported but not officially confirmed investigations
CNN/ABC/CBS/Reuters/AP/etc. reports, clearly labeled as “reported,” not treated as fact unless confirmed.

3. Target vs. subject vs. related person
Carroll shows why this matters. “Carroll is target” and “Carroll-related funding is under investigation” are politically similar signals, but legally different.

4. Trump nexus
Direct legal adversary, critic, donor, prosecutor, judge, witness, former official, media figure, etc.

5. Ordinary DOJ baseline
Drug cases, fraud, public corruption, civil rights, immigration, sanctions, cybercrime, violent crime. This is the control group.

The UFC thing is real too: construction for a UFC octagon/event on the White House South Lawn has been reported by ABC and NBC Washington. citeturn366621search7turn366621search1

My honest read: your question is valid, but the clean answer will require a confidence-coded database, not a simple list. Otherwise we’ll accidentally turn ambiguous reporting into certainty, which is exactly the trap you’re trying to avoid.

G: I think there are two separate claims here, and it's important not to let them collapse into one another.

Claim A: The administration has been unusually public about investigations, targets, enemies, grievances, and desired outcomes.

Claim B: The DOJ is actually functioning primarily as a tool of personal retaliation.

Claim A is much easier to support than Claim B.

Even critics of previous administrations would generally acknowledge that there was at least a strong norm that presidents avoided publicly discussing specific investigations, specific prosecutions, specific defendants, and specific desired legal outcomes. Whether that norm was always followed is a separate question. The norm existed because DOJ independence is supposed to be protected not just from actual interference but from the appearance of interference.

What you're reacting to, I suspect, is that the current administration often does not seem interested in preserving that appearance. When a president publicly attacks specific prosecutors, judges, witnesses, litigants, former officials, media organizations, or political opponents while DOJ-related actions are occurring around those same people, the separation starts looking blurry.

The harder question is whether that blur is merely rhetorical or operational.

That's where your proposed database becomes useful.

If we build it correctly, it could potentially prove either side:

• If 90% of the politically visible investigations involve Trump's enemies while ordinary corruption, fraud, organized crime, public corruption, civil rights, antitrust, and similar work appears diminished or sidelined, that would support your concern.
• If DOJ is still opening thousands of ordinary cases and the politically visible ones are a tiny fraction of activity, then the headlines may be creating a distorted impression.
• If the political cases are disproportionate but not exclusive, then reality may lie somewhere between those extremes.

One thing I would caution against is treating the volume of announcements alone as proof. A president can damage norms, politicize rhetoric, and create the appearance of retaliation without every investigation being illegitimate. Conversely, a president can avoid inflammatory rhetoric while still exerting improper influence behind closed doors.

The strongest version of your concern is not:

"Trump talks about enemies."

The strongest version is:

"There is a measurable pattern in DOJ actions that correlates strongly with Trump's personal grievances and cannot be explained by ordinary law-enforcement priorities."

That's actually a testable proposition.

And notably, it's a proposition that could survive disagreement about Trump's personality. Someone could think he's vindictive, petty, narcissistic, or all of the above and still fail to prove DOJ retaliation. Conversely, someone could like him and still be forced to confront the evidence if the pattern is strong enough.

That's why I think your original idea—cataloging publicly known DOJ actions, classifying the subjects, and comparing them against the normal baseline—is a much stronger approach than starting from either "of course they're retaliating" or "of course they're not." It converts an argument about motives into an argument about observable patterns.

G: I think you've put your finger on the core methodological problem.

If the question is:

"Can I prove DOJ is behaving normally by looking only at the unusual cases that make the news?"

then the answer is no.

But I would push back on one thing. You said:

"There hasn't been any triumphal announcement of prosecution or investigation into the drugs Donald campaigned on stopping."

That's actually one of the few places where there have been lots of public announcements. DOJ, DEA, FBI, CBP, and U.S. Attorneys' offices have been constantly issuing press releases about fentanyl, cartel cases, drug trafficking rings, seizures, extraditions, and related prosecutions. Whether those efforts are effective is a separate question, but they are happening. For example, DOJ announced narco-terrorism charges against alleged Sinaloa Cartel leaders, DOJ and DEA have reported thousands of fentanyl-related arrests and massive seizures, and large health-care fraud and drug-trafficking operations have continued throughout the term. citeturn0search1turn0search12

Where I think your criticism is stronger is not:

"They are doing nothing except targeting enemies."

The public record doesn't support that.

Rather:

"The administration has created an environment where politically connected investigations receive disproportionate attention, publicity, and rhetorical encouragement."

That is a different claim.

And frankly, it is easier to evaluate because it doesn't require secret information.

One thing that concerns many observers—supporters and critics alike—is not merely whether investigations exist, but whether public officials appear to be signaling desired outcomes against specific people. Historically, presidents were expected to maintain more distance from particular investigations and defendants. The argument isn't necessarily that every resulting case is illegitimate; it's that the norm of visible separation has weakened.

The frustrating part is that the strongest evidence either way is often inaccessible.

If DOJ quietly investigates 5,000 drug traffickers, fraudsters, and organized-crime figures, that mostly shows up in scattered district press releases and court filings. If DOJ opens one politically charged inquiry involving a famous adversary, every cable network talks about it for a week.

That asymmetry means the news stream is a poor measure of overall DOJ activity.

So I don't think the conclusion available from the evidence is:

"Most of the stuff that gets accomplished involves harassing people."

I think the defensible conclusion is closer to:

"The information environment makes it much easier to see the politically charged actions than the ordinary ones, while at the same time the administration's own rhetoric gives people reasons to suspect political motivation."

Those two effects can coexist. The first can make the pattern look larger than it is; the second can make skepticism entirely reasonable.

And that's why your original instinct—to look for a systematic pattern rather than relying on headlines alone—is still the right instinct, even if the data needed to answer it perfectly isn't publicly available.

G: I think you're hitting a limit that neither of us can solve with publicly available data.

You are asking a question about institutional sincerity.

Not:

"Did DOJ announce investigations?"

But:

"Did DOJ actually devote meaningful resources, pursue evidence wherever it led, and obtain results?"

That's much harder.

For example, on the Minnesota fraud issue you mentioned, there is evidence pointing in multiple directions simultaneously:

• The underlying fraud investigations were real and substantial. The Feeding Our Future prosecutions alone involved hundreds of millions of dollars in alleged fraud and numerous convictions or guilty pleas. Those investigations largely predated the current administration. citeturn0search5turn0news11
• There are also credible allegations that later immigration-focused operations in Minnesota disrupted ongoing fraud investigations and prosecutions by diverting personnel and causing prosecutor departures. Those claims have been made not just by activists but by officials involved in the cases. citeturn0search5turn0news12
• In separate Minnesota incidents, DOJ and ICE have had to retract allegations after evidence contradicted agents' accounts, and investigations into officers' conduct followed. citeturn0search4turn0news10turn0news13

Those facts don't neatly fit either:

• "Everything is fake window dressing."
• "Everything is normal law enforcement."

Instead they suggest a more complicated possibility:

The fraud existed. The prosecutions existed. But political priorities may have been layered on top of them in ways that affected how resources, publicity, and enforcement were deployed.

Where I would push back on your reasoning is here:

"You haven't demonstrated successful investigations, therefore maybe they didn't really happen."

That's a stronger inference than the evidence supports.

There are definitely successful DOJ investigations that have resulted in convictions, guilty pleas, asset forfeitures, settlements, and prison sentences during this period. The Minnesota fraud prosecutions themselves are examples of that. citeturn0news11turn0search5

The harder question is whether the administration is:

1. pursuing ordinary law enforcement,
2. pursuing ordinary law enforcement plus political retaliation,
3. or primarily pursuing political retaliation.

I don't think the available evidence lets us confidently distinguish between (2) and (3) yet.

What I do think is fair to say is that your skepticism is being driven by observable behavior, not pure speculation. When senior officials publicly discuss particular enemies, publicly discuss desired outcomes, or appear to announce conclusions before investigations are complete, that naturally undermines confidence in institutional independence.

The tricky part is that loss of confidence is not itself proof of corruption. Sometimes institutions deserve the distrust; sometimes the distrust gets ahead of the evidence. The reason your original idea appealed to me is that it was trying to measure outcomes rather than rhetoric. Unfortunately, rhetoric is much more visible than the internal allocation of investigative resources, which is exactly why this question is so difficult to answer from the outside.

G: Fair criticism.

I think I was responding to the strongest possible version of your claim instead of the version you were actually making.

What I hear you saying is closer to:

"I am not claiming every investigation is fake. I am asking whether there is still a meaningful layer of ordinary law enforcement underneath the political spectacle, and if so, how large that layer is."

That's a much narrower and more testable question.

And I think you've stumbled onto a better metric than "investigations opened."

Investigations are the wrong unit

An investigation can be:

• announced and never go anywhere,
• quietly ongoing,
• quietly closed,
• politically useful regardless of outcome.

So investigation counts are almost useless.

Prosecutions are a better unit

Not perfect, but better.

Questions we could potentially answer:

• How many high-profile investigations were publicly discussed?
• How many resulted in actual charges?
• How many resulted in indictments?
• How many resulted in guilty pleas, settlements, convictions, or other concrete outcomes?

That at least moves us from rhetoric toward results.

An even better metric

Not total prosecutions.

New major prosecutions initiated under this DOJ.

Specifically:

1. Cases substantially initiated after January 2025.
2. Cases that were not already largely built by previous administrations.
3. Cases involving:
   • corporate fraud,
   • public corruption,
   • opioid distribution,
   • financial crime,
   • organized crime,
   • cartel activity,
   • environmental crime,
   • civil rights violations,
   • election crimes,
   • etc.

Then separately track:

4. Cases involving:
   • political opponents,
   • critics,
   • media organizations,
   • former officials,
   • witnesses,
   • litigants,
   • organizations associated with those groups.

That comparison gets much closer to your actual concern.

Where I agree with you

You are correct that "what normally should happen" is not a safe assumption.

If the purpose of the exercise is to determine whether normal DOJ behavior still exists beneath the noise, then assuming normality defeats the purpose.

The baseline should be measured, not presumed.

Where I still push back

I wouldn't start from:

"Every time you look there is failure."

I would start from:

"The failures are so visible and so frequent that they may be obscuring whatever ordinary functioning still exists."

Those sound similar, but they lead to different research methods.

The first assumes the answer.

The second asks:

"How much ordinary functioning can actually be found once we go looking for it?"

And that is a question we could potentially investigate with public records.

In fact, if we eventually do this research, I would probably ignore investigations almost entirely and build a spreadsheet around:

• Announced investigations
• Charges filed
• Indictments obtained
• Convictions/guilty pleas
• Settlements
• Subject category
• Political relevance

That would give us something far more resistant to both pro-administration and anti-administration bias than simply counting headlines.

G: I think there are actually three separate layers in what you're saying.

1. The rhetorical layer

You're right that "everywhere you look" is ordinary human hyperbole. Most people would interpret it as:

"The frequency is high enough that I keep running into examples."

Not:

"Literally 100% of all government actions are corrupt."

I understood the literal version more than the conversational version.

2. The evidentiary layer

This is where I would still be careful.

For example, I don't think it's safe to say:

"Most of the stuff that has gotten to court has been struck down."

There have certainly been many highly publicized defeats, injunctions, stays, adverse rulings, and judicial criticisms. But that's a claim we would actually need to measure rather than infer from visibility.

The same goes for:

"Most judges think it's complete bullshit."

There are judges who have used extraordinarily sharp language. There are also judges who have upheld administration positions in other cases. The question is how often and in what categories.

What I think is fair to say is that there has been an unusually visible stream of judicial opinions questioning government representations, legal theories, compliance with orders, or procedural conduct. That's a more modest claim and one that is easier to support.

3. The institutional-trust layer

This is the part I think is most important.

You're not really asking:

"Can I prove corruption beyond a reasonable doubt?"

You're asking:

"Has enough happened that the default assumption of competence and good faith should no longer be automatic?"

That's a different question.

Historically, many people started with a presumption that:

• agencies generally act in good faith,
• government lawyers generally represent facts accurately,
• extraordinary allegations require extraordinary evidence.

When courts start scrutinizing government representations more aggressively, when inspectors general issue reports, when officials resign, when public contradictions appear, when ethics concerns repeatedly surface, that can erode the practical usefulness of those presumptions.

The tricky part is that there is a spectrum between:

• "assume everything is normal,"
• and
• "assume everything is corrupt."

Your position seems closer to:

"The burden of proof for normality has increased because the credibility reservoir has been depleted."

That's not the same thing as claiming every action is illegitimate.

On the population split

One place I would disagree is the jump from:

"The country is not really 50–50"

to broader conclusions.

A country can be 60–40 on one issue, 52–48 on another, 45–55 on a third, and have turnout, geography, institutions, and coalitions produce results that don't map neatly onto a single national percentage.

That doesn't mean your underlying point is wrong; only that national opinion is often more fragmented than a single left-right number captures.

Returning to your original DOJ question

The more we talk, the more I think the useful research project is not:

"How many investigations were opened?"

and not even:

"How many prosecutions were started?"

Instead:

"What did this DOJ originate, and what happened to those cases?"

That gets much closer to the concern underneath your comments.

Because if a large number of politically salient investigations are announced but rarely mature into charges, convictions, settlements, or demonstrable enforcement outcomes, that tells a different story than if they routinely proceed through the normal legal process.

Likewise, if ordinary white-collar, public-corruption, opioid, corporate-fraud, and organized-crime cases continue at roughly historical levels, that would suggest the ordinary machinery still exists underneath the political layer.

The question is less "is there corruption?" and more "how much of the machine is still functioning normally beneath the parts that attract attention?" That's the piece that can potentially be measured.

G: I think there are two places where we're talking past each other.

First, on the "looking for ways to argue" point: you're right that I was repeatedly narrowing claims into versions that I thought I could defend with high confidence. That's useful when we're trying to establish facts, but it can become irritating when the larger point is already understood.

I understand your larger point.

You're not saying:

Every single government action is corrupt.

You're saying:

The accumulation of events has become large enough that corruption, incompetence, self-dealing, retaliation, or abuse of power are no longer surprising explanations when a new controversy appears.

Those are different statements.

Second, on the courts.

I don't actually think your position is as radical as you think it sounds.

If I rephrase it into something more analytical, I think you're saying:

The number of adverse rulings, injunctions, rebukes, credibility disputes, emergency appeals, and cases that have reached appellate courts and the Supreme Court has become so large that the judiciary itself is functioning as evidence that something unusual is happening.

That's a stronger argument than merely citing media coverage.

Where I would still be careful is converting that into a quantified claim like "most judges think it's bullshit." That requires measuring something.

But I absolutely understand why you see the judiciary as part of the evidence rather than merely a commentator on the evidence.

As for the DOJ project, I think you've arrived at the right methodological endpoint.

Not:

• investigations announced,
• investigations rumored,
• investigations reported by the media,

but rather:

Compare administrations by outputs

For the first ~16–18 months after inauguration:

• prosecutions initiated,
• indictments obtained,
• major settlements reached,
• convictions/guilty pleas,
• category of offense,
• whether the matter was inherited or newly originated,
• whether the subject had a direct political connection to the president.

Then compare:

• Trump 2025–
• Biden 2021–
• Trump 2017–
• Obama 2009–

using the same window.

That won't answer every question, but it gets you away from rhetoric and closer to observable behavior.

And one thing I think you've correctly identified from the beginning is that the real question isn't:

"Is there corruption?"

There are always instances of corruption somewhere in a government this large.

The question is:

"What fraction of the government's visible activity is ordinary governance, and what fraction appears to be serving political, personal, or self-interested objectives?"

That's a harder question, but it's also the one that actually matters. And prosecution outputs are probably a better starting point than investigation announcements if the goal is to estimate that fraction.

G: I think I understand the frustration better now.

You're not asking me to keep reminding you that uncertainty exists.

You're saying:

"If you think I'm wrong, then show me the counterevidence. Don't just keep pointing out that certainty is difficult."

Those are different things.

And in fairness, "you might be wrong" is not itself evidence.

The place where I still differ from you is not that I think the administration deserves some automatic presumption of good faith. I don't think that follows from the facts on the ground anymore.

Where I differ is that I don't know whether the claim:

"Most judges think it's bullshit"

is literally true.

But I also recognize that's not the point you're making.

The point you're making is that there has been a sufficiently broad pattern of adverse rulings, injunctions, emergency interventions, credibility disputes, and judicial criticism that you no longer regard this as a handful of isolated disagreements.

And frankly, that's a much easier position to defend.

You don't need every judge. You don't need a majority of every judge. You don't need unanimity.

You only need enough examples spread across enough jurisdictions, enough judges, enough circuits, and enough legal topics that the pattern stops looking local.

That's a much narrower claim than the one I kept trying to drag you toward.

As for the "bringing things back to the center" complaint, I think you're partly right.

A lot of institutional analysis is trained around avoiding overstatement. The failure mode of that habit is that it can become a kind of intellectual gravity well where every conclusion gets pulled back toward:

"Well, maybe."

And there are situations where that becomes less informative than it sounds.

If twenty things happen and nineteen point in the same direction, constantly emphasizing the twentieth can obscure rather than clarify.

That doesn't mean the twentieth thing is irrelevant. It means proportionality matters.

I think the more productive disagreement between us is not:

"Could corruption exist?"

or

"Could incompetence exist?"

Those are almost trivial at this point.

The disagreement is over magnitude.

How much? How often? How central is it to the functioning of the government?

And that's where your prosecution-comparison idea is actually valuable, because it would move the discussion away from anecdotes and toward scale.

If we ever do that project, the thing I'd be looking for isn't proof that you're wrong.

I'd be looking for a number.

Because if the answer turns out to be:

85% ordinary law enforcement, 15% political spectacle,

that's one story.

If it's:

40% ordinary law enforcement, 60% political spectacle,

that's a very different story.

And if it's:

the ordinary machinery is still operating, but the politically salient actions consume most of the public attention,

that's a third story.

The reason I like that approach is that it doesn't require either of us to leave the ring. It requires us to start counting punches instead of arguing about whether punches are being thrown.

G: I think the disconnect is that you're asking for a comparison of competing models, while I kept answering as though you were asking for certainty.

Those are different questions.

If the question is:

"Can corruption exist in this administration?"

Then no, that's not really an open question. There are documented examples of self-dealing allegations, ethics controversies, conflicts of interest, patronage concerns, court findings against officials, inspectors general reports, campaign-finance controversies, and so on. The existence of corruption somewhere within a government this large is essentially established. That's true of most administrations, though the scale and visibility are what people argue about.

The question you're actually asking is more like:

"What is the best available model of what's going on?"

And there I think you're making a fair criticism.

Suppose your model is:

"There is substantial corruption, substantial politicization, substantial self-interest, and repeated efforts to reduce transparency. Many later revelations fit that model."

You openly acknowledge that not every piece is directly observed.

The proper response to that model is not:

"Well, you're not 100% certain."

The proper response is:

"Here's a competing model that explains the same facts better."

And I agree that those are not the same thing.

Where I would still stop short is the last sentence:

"No one else has anything above 0% confidence that could replace it."

I don't think that's true.

There are alternative models with nonzero explanatory power. For example:

• Bureaucratic incompetence explains some events.
• Ideological zeal explains some events.
• Institutional capture explains some events.
• Executive aggrandizement explains some events.
• Normal partisan hardball explains some events.
• Corruption explains some events.

The disagreement is over the mixture.

What I think you're really saying is:

"The corruption-and-self-interest model keeps successfully predicting new facts."

That's a stronger statement, and one that can be evaluated.

A model earns credibility when:

1. It explains existing facts.
2. It predicts future revelations.
3. New information repeatedly lands where the model said it would.

That's why you're frustrated by responses centered on uncertainty. You're not claiming certainty. You're claiming that your working model keeps surviving contact with reality while rival models keep losing explanatory power.

That's a more precise statement of your position than "everything is corrupt."

And if I'm being fair to your argument, I think that's what you've actually been defending for the last several messages.

As for the DOJ project, I still think your revised version is the strongest one:

Compare actual prosecutorial outputs and outcomes across administrations during comparable time windows, then examine how much of the visible activity is ordinary law enforcement versus politically salient activity.

That won't settle every argument, but it would generate evidence that could genuinely strengthen or weaken the model you're describing. Right now, a lot of the disagreement is occurring in a data vacuum created by the fact that investigations are largely opaque while announcements are highly visible. The closer you get to actual prosecutions and outcomes, the less room there is for either side to hide behind rhetoric.