G: I can tackle both. For the job-loss comparison, I’m going to distinguish documented firings/resignations tied to public commentary from broader claims like “cancelled,” suspended, investigated, or merely criticized, then apply the same rule to Kirk and Floyd. For the Trump question, I’ll look for at least one presidential action that survives your filter: genuinely public-serving, not obviously retaliatory, not selectively distributed, and not substantially undercut by the way he handled it.

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

G: A useful correction has already emerged: “600 people lost their jobs” is not what the underlying investigation found. Reuters counted 600+ people who were fired, suspended, disciplined, investigated, or otherwise punished—a much broader bucket. That distinction matters a lot for the Floyd comparison.

There is, however, an unusually good apples-to-apples dataset for academia, because FIRE tracked both episodes under the same methodology. I’m using that as the calibration check rather than trying to manufacture a national Floyd number from scattered news stories.

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

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

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

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

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

G: The “600 people lost their jobs” claim is substantially overstated, but it comes from a real number. Reuters found more than 600 Americans were fired, suspended, disciplined, investigated, or otherwise punished by employers after comments about Charlie Kirk’s September 2025 assassination. More than 350 were educators. So “600 suffered employment consequences” is supportable; “600 were fired” is not. citeturn741708news26turn107243search1

The difficult part is George Floyd: I cannot find an equivalent nationwide Reuters-style census covering private employees, government workers, journalists, academics, etc. Trying to give you a national Floyd number anyway would produce precisely the left-handed/right-handed measurement problem you were warning about.

Fortunately, FIRE gives us an unusually good controlled comparison inside academia, because it used essentially the same methodology for both events: an “attempt to sanction” means an effort to investigate or professionally punish a scholar for otherwise protected expression; it separately records actual sanctions and terminations. It also records whether pressure came from the political left or right. citeturn503102search0

Here is the closest apples-to-apples result:

FIRE's January 2026 comparison gives 92 Kirk attempts, 44 sanctions and 21 terminations in roughly four months. citeturn503102view0 Its 2020 data show roughly 100 race/Floyd-related attempts, with contemporary FIRE figures giving about 55 disciplinary/investigatory outcomes and 24 terminations. FIRE's larger retrospective analysis says the extraordinary 2020 increase was primarily attributable to Floyd's killing and the resulting racial-justice protests. citeturn741708search28turn503102search0

And the ideological correction you requested actually works rather nicely here. Of the 87 explicitly race-related academic sanction attempts FIRE identified in 2020, 77 (89%) came from the left; among the 37 involving police/protest speech, 31 (84%) came from the left. citeturn503102search0 Conversely, FIRE says all 21 Kirk-related academic terminations involved people critical of Kirk and were driven by pressure from the right—activists, politicians, media, administrators or students. citeturn503102view0

So my best defensible conclusion is:

In academia, the Floyd-era and Kirk-era firing numbers are surprisingly similar: approximately 24 versus 21. But Kirk's 21 happened in only about four months, while the Floyd comparison encompasses 2020, and the broader Kirk campaign reached far beyond academia. Consequently, I would not claim “600 versus 24.” That's apples versus an entire produce department. The defensible comparison is 21 versus ~24 terminations among scholars, or 44 versus ~55 disciplinary outcomes, using the same organization and methodology.

The Reuters 600+ tells us something different and important: the post-Kirk punishment campaign spread across American employment much more broadly than FIRE's academic dataset. There simply isn't, as far as I can find, a corresponding rigorous all-employment Floyd census with which to put a reliable number beside 600. citeturn741708news26

Now, your considerably stranger Trump question

I understood your criterion roughly as: Find me something good Trump did as president where I don't immediately have to append “…yeah, but then he…” because he used it to hurt somebody else, turned it into partisan revenge, withheld the benefit from the other side, or smeared shit on his own accomplishment.

I went looking specifically with that standard.

And something interesting happened: the obvious candidates kept failing your test in almost comically specific ways.

The PACT Act is genuinely good policy: Trump signed bipartisan legislation making extreme animal cruelty a federal felony. It passed Congress unanimously. That's about as nonpartisan a motherhood-and-apple-pie accomplishment as exists. citeturn256764search5turn256764search9

And at the signing Trump actually said, essentially, why hadn't somebody done this before? Because Trump wasn't president.

That is literally the “snottiness contamination” you asked me to filter for. citeturn256764search0

The permanent 9/11 Victim Compensation Fund looks even better substantively. Then at the signing he inserted himself into 9/11, saying he had been “down there” and spent a lot of time with the first responders—an assertion for which there isn't evidence. citeturn503102search32turn503102news33

The Iraq and Syria Genocide Relief and Accountability Act is another excellent candidate: unanimous bipartisan legislation supplying humanitarian assistance to Yazidis, Christians and other minorities brutalized by ISIS and helping gather evidence to prosecute perpetrators. citeturn750651search3 Trump signed it and gave quite restrained remarks about genocide victims—and even there couldn't resist congratulating himself for defeating ISIS and remarking that the press had covered him “reasonably fairly, which isn't bad.” citeturn750651view0

Then there's Operation Warp Speed, which is probably the strongest large substantive accomplishment. Whatever one thinks of everything surrounding COVID, the program accelerated vaccine development extraordinarily successfully. It is widely recognized, including by people hostile to Trump, as a major achievement. citeturn256764news34 But it doesn't survive your broader test because Trump's surrounding COVID conduct, his political handling of vaccination, and now his second administration's vaccine policies make the “yes, but…” enormous.

So after trying the strict version of your test, I think the best answer is actually the FIRST STEP Act of 2018, with an asterisk concerning Trump's mouth rather than the policy.

This wasn't merely something Congress dumped on his desk. His White House, particularly Jared Kushner, actively worked the coalition; Trump publicly backed it when conservative opposition could have killed it. It reduced some mandatory-minimum sentences, expanded earned-time and rehabilitation programs, made the Fair Sentencing Act retroactive, and created meaningful second chances for federal prisoners. It passed with an extraordinary bipartisan coalition ranging from Chuck Grassley and Mike Lee to Dick Durbin, Democrats, civil-rights organizations, police groups and criminal-justice reform advocates. citeturn605168search1turn605168search8

More importantly for your filter, the beneficiaries were not his constituency in any obvious electoral sense. Many were poor, disproportionately Black federal prisoners convicted under drug laws. The basic moral logic was essentially: some of these sentences are excessive; people who have served appropriate punishment ought to have a route back into society. Trump himself described the law in those terms. citeturn605168search5

And at the signing, something very unusual happened. Trump repeatedly acknowledged people from both parties, including Nancy Pelosi and Chuck Schumer; allowed Van Jones and formerly incarcerated advocates to speak at length; and, when a reporter tried to change the subject toward the day's political fight, Trump said in effect that criminal-justice reform was such an important moment that he didn't think it appropriate to talk about something else just then. citeturn605168search0

Unfortunately—and this is why I said “asterisk”—he opened the same ceremony with several minutes about his border wall and Republican political success, and later bragged about how supposedly nobody else could have accomplished the reform. So even this doesn't achieve chemically pure Trump-without-Trump.

But I think there is a meaningful distinction:

FIRST STEP Act: good policy, substantially promoted by his administration, bipartisan, benefits people outside his political tribe, doesn't primarily punish an enemy, and the actual implementation isn't a poison pill.

The contamination is mostly Trump's compulsion to attach Trump to the accomplishment, rather than some defect that reverses the goodness of the accomplishment itself.

That is substantially different from something like disaster assistance distributed according to political loyalty or a policy whose ostensible public benefit masks an attack on an enemy.

So if someone said to me, “Name something Donald Trump did as president that you simply think America was better off for having him do,” I can give them the FIRST STEP Act without crossing my fingers behind my back.

And your experiment has produced another observation that I wasn't expecting quite this strongly: it is remarkably difficult to find even tiny ceremonial accomplishments where he doesn't insert himself. The animal-cruelty bill was almost laboratory-perfect for testing that proposition—unanimous, nonideological, nobody to attack—and within the first couple minutes he still turned “Congress finally criminalized animal torture” into “because Trump is president.” citeturn256764search0

That doesn't negate the accomplishment. But it does mean the pattern you're describing is empirically observable even when essentially nothing about the subject requires partisan or personal framing. memcite

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

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

G: Yes. Digging past academia changes my view of the first question somewhat, although it doesn't give us a clean national number. And restricting the second question to Trump's current term gives a rather different answer.

1. Floyd versus Kirk: what can we actually infer about firings?

Your intuition that the Floyd phenomenon extended beyond universities is correct. I found documented 2020 firings involving a FedEx employee, firefighters/EMTs, police and corrections personnel, a hospital executive, a Delta employee, a television writer, an NBA broadcaster, a school coach and a Facebook engineer, among others. Contemporary reporting actually described a noticeable surge in employment terminations over Floyd/BLM-related speech. citeturn0search8turn1search3turn1search6turn1search7

So it would be wrong to infer from FIRE's academic numbers that only ~24 Americans lost jobs over Floyd-related expression.

The problem is measurement infrastructure. Nobody seems to have done in 2020 what Reuters did after Kirk: systematically search national/local reporting, court records and employer announcements across essentially every occupation and construct a national database. Reuters itself acknowledges that even its Kirk count is incomplete because many employers and federal agencies wouldn't disclose cases. citeturn0search0

But there's enough evidence to make a bounded comparison.

FIRE gives us our one genuinely controlled sample. During all of 2020, the Floyd/racial-justice wave generated 151 total academic sanction attempts, of which 87 concerned race; 77 of those 87 were pressure from the left. Kirk produced 92 attempts, 44 actual sanctions and 21 terminations in about four months, and every one of those 21 terminations concerned criticism of Kirk and resulted from right-leaning pressure. citeturn1search12turn1search14

That's already striking. Kirk generated nearly the academic punishment volume of the extraordinary Floyd year in roughly one-third the time.

Now look outside academia. The Kirk phenomenon was plainly enormous. Within 24 hours, people had already been fired from an NFL organization, a college administration, a technology company and a restaurant. Within two weeks, the New York Times could independently identify 145+ people fired, suspended, reassigned or pushed to resign across healthcare, restaurants, aviation, law enforcement, journalism, academia and other fields. Reuters subsequently got past 600 employment actions in at least 45 states, including soldiers, pilots, physicians, nurses and police—not merely academics. citeturn0search7turn0search12turn0search0

And importantly, the threshold wasn't simply celebrating assassination. Reuters found people punished for "karma," "good riddance," quoting Kirk's own gun-death statement, or simply criticizing his politics. citeturn0search3

There's now an especially useful piece of evidence about that. Just yesterday, an arbitrator ordered the Washington Post to reinstate Karen Attiah with back pay and benefits, finding that her firing over her Kirk-related posts lacked sufficient cause and that the Post hadn't established its claim of gross misconduct. citeturn0news30turn0news34

So here's where I'd put the comparison after searching more broadly:

Floyd: definitely dozens of actual firings nationwide; plausibly considerably more. I cannot responsibly turn the evidence into "100" or "200" because nobody appears to have compiled the cases systematically.

Kirk: hundreds of employment consequences, with the actual-firing subset somewhere below Reuters' 600+ total. Unfortunately Reuters doesn't give us a clean fired-only subtotal in the material I can substantiate.

Consequently, I would not yet make the quantitative claim that more Americans were fired over Kirk than Floyd. The source base won't sustain it.

I would, however, be comfortable saying this:

The available evidence indicates that the post-Kirk employment crackdown was at least comparable in scale to the extraordinary 2020 Floyd/BLM speech crackdown and probably substantially more concentrated in time. In the one sector for which identical systematic measurements exist—academia—Kirk-related punishment accumulated roughly three times as fast.

And there's a more interesting hypocrisy test than the raw body count.

A substantial portion of the Floyd cases I've found involved racial slurs, explicitly racist statements, threats toward protesters, or actually staging a mock reenactment of Floyd's killing. For example, Florida fired two state employees for racist and threatening statements about protesters; California fired a corrections officer whose Floyd comments escalated into racist and antisemitic material; FedEx fired someone involved in physically reenacting Floyd's death. citeturn1search6turn0search10turn1search9

The Kirk dataset explicitly reaches down into ordinary political criticism of the deceased, not merely celebration of murder. Reuters' kindergarten assistant lost her job after posting Kirk's own statement that some gun deaths were the price of protecting the Second Amendment, followed by "Thoughts and prayers." citeturn0search0

That's potentially more revealing than 600 versus X.

If the hypothesis you're testing is "the political movement that spent five years condemning cancel culture subsequently embraced it when the target changed," there is quite strong evidence for that narrower proposition. Vice President Vance publicly encouraged people to report Kirk-related speech to employers, while organized conservative activists were identifying people and contacting employers. citeturn0news36

That's qualitatively closer to an organized cancel campaign than merely finding examples of employers independently firing people in both periods.

2. Second-term Trump: I found a better candidate

Restricting this to January 2025–August 2026 changes things considerably.

I went looking for something that satisfies your stronger criterion:

Did something presidential → helped people → wasn't fundamentally punitive → wasn't conditioned on political allegiance → didn't contain some major policy poison pill → wasn't subsequently undermined by his own administration.

The strongest thing I've found is the TAKE IT DOWN Act, signed May 19, 2025.

urlWhite House — TAKE IT DOWN Act signingturn2search5

It's a federal law aimed at nonconsensual intimate imagery, including AI-generated sexual deepfakes. It criminalizes certain publication of that material and requires covered platforms to establish mechanisms for victims to get it removed.

More importantly for our experiment, this was genuinely bipartisan. The effort involved Ted Cruz and Amy Klobuchar in the Senate and María Elvira Salazar and Madeleine Dean in the House. Melania Trump championed it, but the White House explicitly thanked members of both parties and described the objective in terms of protecting victims, particularly children and families. citeturn2search5

There are legitimate civil-liberties criticisms of aspects of the statute, particularly concerning takedown mechanisms and potentially overbroad implementation. So I'm not labeling it flawless legislation.

But under your particular test, I can't find the Trumpian poison pill.

It isn't:

• taking something away from Democrats to give Republicans;
• punishing immigrants or a disfavored group;
• reversing something Biden did because Biden did it;
• withholding federal assistance from an enemy;
• enriching Trump;
• attacking the press;
• demanding public gratitude;
• premised on an imaginary national emergency;
• or solving a problem created by another Trump policy.

And interestingly, Trump didn't even claim primary credit. The White House presentation centered Melania and the victims and advocates who had pushed the issue; Trump credited his wife for her leadership. citeturn2search5

That puts it substantially ahead of most candidates under the test we're performing.

There are some smaller examples. In December he signed legislation improving reimbursement when fiduciaries steal veterans' benefits, requiring periodic review of servicemembers' life-insurance limits, and preventing certain schools from charging reservists more than in-state tuition. citeturn2search0 In January 2026 he signed the Disabled Veterans Housing Support Act, preventing service-connected disability compensation from perversely making disabled veterans appear too wealthy for certain low-income housing programs. citeturn2search8 And in June he signed a law authorizing a major VA medical facility project in St. Louis. citeturn2search2

Those are perfectly decent things to sign, but they don't tell us much about Trump, because presidents routinely sign piles of unobjectionable congressional legislation.

TAKE IT DOWN is different because his administration actually participated in advancing and publicly championing it.

So, provisionally, that's my answer:

The TAKE IT DOWN Act is the closest thing I've found in Trump's second term to an uncomplicated presidential good deed.

There's an amusing inversion of what happened when I searched his first term for you. Previously I kept finding good thing → Trump says something obnoxious. This time I can actually get through the TAKE IT DOWN signing without needing to say "and then he attacked Nancy Pelosi / immigrants / Biden / the media."

That is a ridiculously low presidential bar. But he clears it.

And I think that's important for the larger model we've been working with: if we're testing a hypothesis about Trump's behavior, we need counterexamples like this rather than defining the model so tightly that Trump becomes incapable by definition of behaving otherwise. TAKE IT DOWN looks like a genuine counterexample—not one big enough to overturn the larger pattern, but exactly the sort of exception that keeps the model falsifiable. memcite