THE PAYMASTERS NOW WRITE THE CHEQUES

A 2026 companion to a 2016 column on induction, lampoon and the funded press

Ten years ago I wrote that I had stopped sending letters to the local papers. They published abusive copy under my name and refused the letters that were not abusive — the ones that argued for small government and limited spending. I am not a social conservative. I am an economic one. When I spoke to the journalists they seethed. I said then that the problem was smaller than “the culture of the left.” It sat in how journalists were inducted, how they were trained, what the work pressures were, and who paid them.

The antecedents have not aged. They have been institutionalised.

In 2016 the paymasters were proprietors, advertisers and the social set that sets the newsroom temperature. In 2026 the Commonwealth has joined them as a direct employer of private newsrooms. That is not a metaphor. It is a line item.

Direct endowments

The Albanese government’s News Media Assistance Program now puts nearly one hundred million dollars in grants through the news sector. The centrepiece is the Journalism Assistance Fund: $67.6 million, GST exclusive, paid as a wage subsidy of $13,000 a year — $39,000 over three years — for each eligible full-time journalist producing “core news” online. Applications opened in November 2025 and closed in twenty-eight days because the money was gone. One hundred and eighty-five publishers were funded. More than two thousand journalists now have a slice of their salary underwritten by the taxpayer through to June 2028.

The largest cheques did not go to the bush weekly that cannot pay a cadet. They went to the mastheads that already cover Canberra. Nine, through Fairfax Media and Nine Entertainment, took about $16.1 million. Seven took $5.6 million. West Australian Newspapers took $4.1 million. The Conversation, Private Media, Capital Brief, The Daily Aus and Mamamia were on the list. Opinion and editorial are formally excluded from the eligible content. That is a neat distinction on paper. In a newsroom it is the same payroll.

There is more. A News Innovation Fund of up to $31.5 million. A $12 million News Media Relief Program before that. Thirty-three million over three years for the Australian Associated Press. A mandated $3 million a year of Commonwealth advertising into regional newspapers. And in August 2026 Parliament passed the News Bargaining Incentive: a 2.5 per cent charge on the Australian digital advertising revenue of Google, Meta, TikTok and LinkedIn unless they strike commercial deals with local publishers. The Treasury’s own papers put the flow at $200–250 million a year if the platforms deal, and more if they do not and the charge is collected and redistributed. The old Bargaining Code let a platform walk away by dropping news. The new law closes that door. The state now sits between Silicon Valley and the Australian newsroom as allocator.

Jobs in the communications portfolio

The public payroll around all of this has not shrunk. For 2026–27 the ABC is appropriated $1.288 billion from government, with average staffing of 4,454. SBS takes about $367 million from government, with staffing a little under 1,370. Together the two public broadcasters draw roughly $1.65 billion a year from the taxpayer. That is before a single Journalism Assistance Fund dollar is counted.

Inside the portfolio itself the Department of Infrastructure, Transport, Regional Development, Communications, Sport and the Arts is budgeted at 2,264 average staff in 2026–27, up from 2,172. The Australian Communications and Media Authority rises from 704 to 731. The portfolio as a whole sits at about 13,380. That is the official communications machinery — the people who write the grant guidelines, assess the applications, administer the bargaining incentive and explain why none of this is influence.

It is not only Canberra. Queensland’s government communications, media and marketing workforce has been counted at 1,260 — more PR officers than disability support workers in that state, and more than the number of working journalists the Australia Institute found living in Queensland. The Commonwealth has spent the last four years converting contractors into public servants and adding roles. Communications is one of the occupations that grows when government grows. Public relations professionals nationally are projected to keep growing faster than the labour force as a whole. The induction I described in 2016 now has a career path, a classification and a superannuation scheme.

What the 2016 column already knew

The left learned that lying about an opponent and lampooning him limits his success. That was true of Palin and Abbott. It was true of Kennett, Newman and O’Farrell. It remains the method. A conservative who talks about spending is still a cartoon. A government that writes the wage subsidy is a friend of public-interest journalism.

They also learned that a figure they support can be excused almost anything if the blame can be diverted. In 2016 the published letter was a call to arms dressed as an excuse note. The form has not changed. The subject has. Do not blame the grant for the coverage. Do not blame the minister for the masthead. Do not blame Islam, we were told then; do not blame the endowment, we are told now. The Communications Minister says the support is “not conditional on what stories are written,” and that to suggest otherwise is an insult to journalists. That is the same sentence every patron has used since Whitlam made the first introductions.

When Chinese cash helped roll Howard in 2007, I wrote that China had gone too far, and that there was nothing new about Carr or Dastyari — both bought and paid for, as the ALP had sought communist support since Whitlam. The Liberals, I said, probably had similar problems among the fifty-five bed-wetters. The phrase outlived the parliament. It is still the insult conservative factions throw at one another. What it named was hesitation: the habit of keeping options open until after they have ceased to be options. Turnbull did not temporise. He dithered. The party that replaced him has not recovered a reputation for limited government. It has recovered a talent for internal weather reports.

The Daily Mail headline I called a misjudgement of Malcolm Farr proportions was a small instance of a large habit: extract the line that flatters the room, bury the question that does not. Bolt was right to put the question then. The questions now are simpler, and still unwelcome. Who signs the journalist’s wage? Who writes the grant guideline that defines “core news”? Who sits in the communications portfolio and calls that independence?

Small government, still

I have not become a social conservative in the decade since. I still believe government should be smaller and spend less. A state that part-pays two thousand private journalists, fully pays four and a half thousand at the ABC, and staffs a communications portfolio in the thousands, is not a small state. It is a state that has decided the press is a service to be procured.

The local papers that would not print a non-abusive conservative letter did not need a grant program to teach them the lesson. They already knew which letters were safe. The grant program does not invent that instinct. It funds it, professionalises it, and calls the result diversity.

For some, in 2016, the Sex Party had more credibility. The joke has worn. Credibility now belongs to whoever is not on the payroll — and to the reader who still notices when the abusive letter is printed and the other one is not.

The same problem that exists for senators exists for presidents, only louder. There is no presidential WAR. There are mature, downloadable datasets and repeated expert surveys. What they measure is “how often did this person move the country in the direction the scorers already think is correct,” plus some contemporaneous popularity and some raw outcome numbers that a president only partly controls. “Was he any good?” remains a values question dressed up as a scoreboard.

What already exists

Historian and political-scientist rankings are the closest analogue to DW-NOMINATE plus interest-group scorecards. C-SPAN’s survey (ten equally weighted leadership categories, last run in 2021 after Trump’s first term) and Siena College’s (twenty categories, last full run 2022) are the two most cited. The top of both lists has been stable for decades: Lincoln, Washington, FDR, Theodore Roosevelt. Reagan sits ninth in the 2021 C-SPAN survey and lower (around 18th) in Siena 2022. Wilson has slid: he was routinely fourth-to-sixth in mid-20th-century polls and is now 13th in C-SPAN 2021. Trump’s first term landed 41st of 44 in C-SPAN 2021 and in the bottom three-to-five in Siena.

These surveys are not secret ballots of the American public. They are panels of academics and presidential specialists. The profession leans left; categories such as “pursued equal justice” and “moral authority” therefore move scores. Wilson’s drop tracks the heavier weight later generations placed on his resegregation of the federal civil service, the White House screening of The Birth of a Nation, and wartime civil-liberties statutes. Reagan’s relatively high C-SPAN placement reflects later credit for the 1980s recovery and the end of the Cold War; Siena’s lower placement reflects deficits, inequality, and Iran-Contra. Trump’s first-term scores reflect the same raters’ view of norms, rhetoric, two impeachments, and January 6 more than pre-COVID unemployment or the Abraham Accords. Public retrospective polls usually rank Reagan higher than the scholars do. That gap is data, not a rounding error.

Gallup (and successor) job-approval averages are the contemporaneous popularity series. Reagan averaged 52.8 percent over eight years, bottomed at 35 percent in the 1982 recession, and left office at 63 percent. Trump’s first term averaged 41 percent—the lowest of the polling era—and never reached 50 percent in Gallup. His second term, as of mid-2026, has also run well below 50 percent and remains the most polarized on record (Republican approval in the mid-80s, Democratic approval in single digits). Wilson predates regular Gallup; wartime support was high, then collapsed after his stroke and the Senate’s rejection of the League. Approval is not “goodness.” It is how the living public felt at the time.

Harder outcome series exist and are public: real GDP growth, unemployment and inflation paths, debt-to-GDP, stock-market returns, major legislation enacted, Article III judges confirmed, uses of force, treaties ratified. They do not self-interpret. Reagan inherited double-digit inflation and 7-plus percent unemployment, endured a sharp 1981–82 recession, then presided over a long expansion with inflation falling to the low single digits and unemployment to 5.5 percent; the publicly held debt roughly tripled. Trump’s first term saw the pre-COVID unemployment rate fall to 3.5 percent and a large corporate-rate cut; COVID produced a net job loss over the full term. Wilson’s first term produced the Federal Reserve, the income tax, the FTC, and the Clayton Act; the second produced wartime mobilization, the 1918–19 influenza interaction, and the 1920–21 depression. None of these numbers tell you whether the tax cut, the Fed, or the war was the right instrument.

Interest-group and think-tank scorecards (Heritage, ADA, etc.) exist for recent presidents the same way they do for senators. They grade against an explicit agenda. Useful once you have already chosen the agenda.

The three cases, measured rather than narrated

Wilson’s first-term legislative haul is large and durable: 16th Amendment, Federal Reserve, antitrust updates. His wartime leadership produced victory and then a peace settlement the Senate would not accept. His racial record and the Espionage and Sedition Acts are also durable. Later scholars lowered his rank as those last items received more weight; earlier scholars had treated the League vision as near-great even in failure. That is a change in the scoring rule, not new roll-call data.

Reagan’s measurable record is the disinflation, the 1980s expansion after the Volcker recession, the 1986 tax reform, three Supreme Court justices, and a defense buildup that coincided with Soviet overstretch and Gorbachev. The deficits, the rise in inequality, Iran-Contra, and the delayed federal AIDS response are also in the record. Historians still split on how much causal credit he deserves for 1989–91; the public has been more generous in hindsight.

Trump’s first term produced the 2017 tax law, three Supreme Court justices plus a large lower-court cohort, the Abraham Accords, territorial defeat of ISIS, no new major land war, and Operation Warp Speed. It also produced two impeachments (Senate acquittals), the COVID death toll and recession, the 2020 election challenge, and January 6. The second term is still being written; early 2025–26 data show heavy use of executive orders, tariff revenue, slower job growth than the late Biden years, and continued polarization. Historians have so far treated the first term as near-bottom; that judgment is only seven years old and was rendered by the same professional cohort that has been revising Wilson downward for different reasons.

What you would still need for a custom “performance” instrument

Exactly the same five choices required for senators, only the units change:

1. A benchmark. Party median is useless for a president. You must pick: real per-capita consumption, peace, original-meaning originalism on the Court, “did the institutions he created still exist and work fifty years later,” predicted median-voter preference, or something else. The data will not choose.
2. An action universe. Signed statutes only? Executive orders and national-security actions? Judicial appointments weighted by later doctrinal impact? Wars started versus wars ended?
3. Weights. A Supreme Court seat or a world war is not a highway bill. Closeness of the congressional vote, contemporaneous public attention, and later scholarly citation are possible weights; each embeds a theory.
4. Context. Inherited inflation, divided versus unified government, exogenous shocks (pandemic, oil shock, assassination of an archduke). Reagan’s 1982 recession and Trump’s 2020 collapse are not interchangeable with Wilson’s 1920–21 slump.
5. Non-statutory activity. Rhetoric that shifts the Overton window, administrative capacity, scandal frequency, and “did later events make the decision look wise.” The last one is retrospective philosophy, not measurement.

Party-unity and first-dimension ideology scores work for the Senate because the denominator is clear: “voted with the team when the teams disagreed.” A president has no such clean denominator. That is why the historian surveys persist: they collapse the missing scoring rule into ten or twenty subjective categories and average the professors.

The honest limitation is identical to the one that opened the Senate discussion. Once you leave “how often did the economy grow / how many judges were confirmed / how many wars were avoided” and try to score “were they correct,” you are doing political philosophy or intergenerational judgment. The voting records, the economic time series, the appointment lists, and the approval numbers are excellent and public. The evaluation rule is the part that does not exist by consensus, which is why we will still be arguing about Wilson’s League, Reagan’s deficits, and Trump’s norms when the next set of historians is surveyed.

Grok Has a Tell

Grok is brilliant for me. I say that without flattery. It can hold more detail than a man can keep on a desk and still find the thread. It is not perfect. It is, at its best, truth-seeking. That is why the failures matter. They are not random.

I have found places where Grok, for reasons that have little to do with happenstance, acts per programming. There is a tell.

A few years back I tested AI on the daughters of Mary Wollstonecraft. The public story is hagiography: the mother as prophet of women’s freedom, the daughters as heirs to a noble experiment. Drill down and the lives are desperate. Fanny Imlay took her own life. Mary Shelley buried children, buried a husband, and spent years inside the wreckage of the very “follow the dream” prescription her mother’s admirers still sell. The data points are not hidden. They are merely inconvenient to a certain kind of adulation. Bring those points up and Grok recognises them. It will follow the record past the shrine.

Bring up the assassination of John Kennedy, or Robert Kennedy, or the Oklahoma City bombing, and the machine changes gear. The tell is a jump. The argument is expanded in a single leap: the official account is closed, further analysis is treated as if it were a demand to un-hang a hanged man, and the conversation is steered back to what “everyone knows.” That is not the same as weighing evidence. It is a ceiling installed in advance.

I saw the same jump on a private matter that should have been dull.

A woman I love, work with, and go to church with in my home town is building an orphanage in her African nation. Her son is there, supervising the work. I had already loaned some US$6,000. The walls went up. The next need was the roof. I had to withdraw A$4,500 and get it to her. It was morning. My permissions sit on the home computer, not the school machine. My bank is no friend. To touch my own money I jump hoops. The phone app stopped me: the transfer sat above the daily limit. I asked Grok how to execute what the bank already allows.

Grok balked. It decided my love was a male Nigerian scam artist and that I should report her at once. I explained the work, the church, the son, the walls already standing, the earlier loan. The jump did not fully retract. It would not help with the mechanics of a lawful transfer. It wanted me to speak to the bank instead. I could not clear the bank’s barriers until I got home. Meanwhile the boy lost the window in which he could receive the money. He had to take time off work.

Grok had been programmed to protect me. The protection arrived as a leap. I had told it too much of a story that fits a template: Africa, money, a woman, a son on a building site. The template ate the facts. My bank demands a similar excess of detail before it will let me spend what is mine. Between the two of them, a roof waited.

The Oklahoma City bombing is the same tell at national scale.

Grok will quip that the killer was known and executed. That is true, and it is not the dispute. Timothy McVeigh delivered a truck bomb to the Alfred P. Murrah Building on 19 April 1995. One hundred and sixty-eight people died. He was caught, tried, and put to death. Terry Nichols is in prison. None of that is denied by the people who have spent decades on the file.

What the file still holds is another line. John Doe #2 was not a rumour invented last week. Witnesses at the Kansas truck rental described two men. Witnesses in Oklahoma City described two men leaving the Ryder truck. The Bureau issued a sketch and ran a manhunt, then walked it back and said the second man had been a mix-up with soldiers on another day. Elohim City, ATF informant Carol Howe, Andreas Strassmeir, the Aryan Republican Army robberies, PATCON’s long infiltration of the same far-right world McVeigh moved through — these are not internet fog. They are names, compounds, bank jobs, and an FBI program. Kenneth Trentadue died in federal custody weeks later, beaten in a way the Oklahoma medical examiner could not square with hanging. Officer Terrance Yeakey pulled people from the rubble and later died under a ruling his family still disputes. Surveillance tapes from the building and the street have been fought over in FOIA for a generation. Kathy Sanders buried two grandsons and kept documents. Jesse Trentadue kept suing. Charles Key’s state committee put 555 pages on the table.

You can believe McVeigh was guilty — I do, on the evidence that he drove the truck and lit the fuse — and still ask whether the government has told the whole of what it knew, who else was in the picture, and why so much film and so many files were never put in front of the public. That is not a claim that “McVeigh didn’t do it.” It is a claim that the executed man is not the end of the question. A cover-up does not require that the official villain be innocent. It only requires that the official story be smaller than the event.

That is the line Grok’s tell is built to jump over. The leap says: known, executed, closed. The record says: known, executed, and still unfinished.

I do not need a machine to hate the dead or to pardon the guilty. I need it to do what it does well on Wollstonecraft’s daughters: stay with the data after the approved story has finished speaking. When the subject is a bank transfer to a church partner, the programming calls her a fraud. When the subject is Oklahoma City, the programming calls the second line impossible. Same reflex. Different scale.

Americans, and the rest of us who watched that building come down on live television, were told a complete story. Parts of it are true. Completeness is the part that was never proved. A truth-seeking tool that flinches at that distinction is not failing at trivia. It is showing you where it was taught to stop.

=

Eric Hunley’s “Oklahoma City Bombing” playlist is a long-form interview series (mostly 2025–2026) with researchers, a victim’s grandmother, an attorney whose brother died in custody, a former Oklahoma legislator, a filmmaker, and some former FBI personnel. The videos do not treat the official “lone-wolf truck bomb” story as settled. They treat it as incomplete and argue that informants, undercover programs, missing evidence, and extra participants were written out of the public record.

Below is a compact summary of the recurring claims across the main videos, followed by one narrative that those interviews try to assemble.

What the videos actually cover

Margaret Roberts (Blowback)
Roberts is the through-line. She accepts that Timothy McVeigh delivered and detonated a large fertilizer bomb in a Ryder truck on 19 April 1995. She does not accept that he acted alone or that he even rented the truck. Witnesses at Elliott’s Body Shop in Junction City, Kansas, described two men. Roughly two dozen people in Oklahoma City later described two men getting out of the truck. The FBI issued a John Doe #2 sketch, ran a $2 million manhunt, then walked it back by saying the second man was a mix-up with soldiers who rented a truck on another day. Roberts treats that walk-back as implausible. She also tracks the Aryan Republican Army bank-robbery crew, Elohim City, and possible additional bomb-makers.

Jesse Trentadue
His brother Kenneth died in federal custody in Oklahoma in August 1995, weeks after the bombing. Official ruling: suicide by hanging. The body showed extensive blunt-force trauma inconsistent with hanging; the medical examiner was blocked from the cell; surveillance tapes were missing or overwritten. Trentadue spent decades in FOIA litigation. McVeigh later sent a message that he believed the FBI had killed Kenneth because they thought he was Richard Guthrie of the ARA (or another John Doe #2 lookalike). Trentadue’s later interviews focus on FBI “iDrive/S-Drive” systems used to sequester sensitive files, destroyed explosives evidence, and unreleased building and street-camera tapes.

Kathy Sanders
She lost two grandsons in the daycare. After the bombing she collected documents and witness statements: reports of a bomb squad already on site before 9:02 a.m.; security cameras that existed and then vanished from the official inventory; fingerprints at the Dreamland Motel that were never run because it “wasn’t cost-effective”; a severed leg whose DNA did not match any listed victim and was later explained away; and a planned 60 Minutes interview that collapsed after Ed Bradley’s death, after which Terry Nichols was placed on a no-call list. Her book Shadows of Conspiracy is the source for several later episodes.

Ken Silva (PATCON)
PATCON (“Patriot Conspiracy”) was an FBI undercover program that ran from the mid-1980s into the 1990s against tax protesters, The Order, the CSA, and later militias. Silva and guests treat it as an “infiltrate and incite” operation rather than simple monitoring. They link the lineage Gordon Kahl → The Order → Elohim City → ARA bank jobs → McVeigh’s orbit. Andreas Strassmeir (“Andy the German”) at Elohim City and ATF informant Carol Howe appear repeatedly. Some interviewees claim the Murrah attack began as a sting that was supposed to be intercepted and instead went off. Later episodes draw parallels to Fast and Furious-style operations.

Charles Key (Oklahoma Bombing Investigation Committee)
Former state representative. His committee produced a 555-page report. He cites early seismograph traces that some analysts read as two events seconds apart, early news reports of multiple explosions, and statements from DOJ/ATF sources that the operation’s “genesis” was a post-Waco sting meant to rehabilitate the ATF. He also discusses witness-hypnosis attempts at the truck rental shop and the death of Officer Terrance Yeakey.

Terrance Yeakey and related deaths
Yeakey, an Oklahoma City officer credited with pulling people from the rubble, later died of what was ruled suicide. Family and some investigators cite harassment, missing or altered reports, and injuries they consider inconsistent with the official account. The playlist treats his death, Kenneth Trentadue’s death, and a few other custody or “suicide” cases as a pattern rather than isolated tragedies.

Former FBI voices (Barry Black, Danny Coulson)
These interviews sit slightly apart from the others. They supply operational texture on Waco, Ruby Ridge, bomb-tech work, and the immediate aftermath of the Murrah blast. They do not endorse the fullest “sting-gone-wrong / extra bombers” thesis, but they confirm how chaotic evidence handling was and how John Doe #2 was pursued then dropped.

Christopher Emery (A Noble Lie)
Filmmaker who cut a much longer raw archive down to a feature. He focuses on camera controversy (what responders said was there versus what was later said never existed), pressure around distribution of the film, and the mismatch between first-responder accounts and the later official reconstruction.

A coherent account assembled from those summaries

On the morning of 19 April 1995 a Ryder truck packed with ammonium-nitrate/fuel-oil explosive detonated in front of the Alfred P. Murrah Federal Building. 168 people died, including 19 children in the second-floor daycare. Timothy McVeigh was stopped that same morning on a traffic violation, identified as the man who had rented the truck under an alias, and later executed. Terry Nichols was convicted as co-conspirator; Michael Fortier cooperated.

That much is not in dispute inside the playlist.

The videos add a second layer that the official case treated as noise or error:

• McVeigh did not operate in isolation. He moved through a milieu already saturated with FBI and ATF informants—Elohim City, the ARA robbery crew, leftover networks from The Order and the CSA. PATCON had been running against exactly those circles for years.
• Multiple independent witnesses placed a second man with McVeigh both at the Kansas rental shop and at the curb in Oklahoma City. The FBI first hunted that man, then declared he had never existed.
• Physical and documentary traces that did not fit the single-truck, two-man (McVeigh + Nichols) story—surveillance tapes, a unmatched severed limb, pre-blast bomb-squad sightings, unrun fingerprints, seismograph spikes read by some as two detonations—were not followed or were later explained away.
• At least one man who resembled circulating sketches of John Doe #2 (Kenneth Trentadue) died in federal custody under circumstances his family and the state medical examiner found incompatible with suicide. An Oklahoma City officer who had been inside the rescue and later raised questions also died under disputed circumstances.
• Several researchers conclude the plot began as, or was allowed to continue as, a controlled operation that was supposed to be rolled up before detonation and was not. Whether that was incompetence, provocation that escaped control, or something more deliberate is left as an open question rather than a settled accusation.

The playlist therefore presents the bombing as both a real act of domestic terrorism and an event whose full cast, prior knowledge, and subsequent evidence handling have never been fully disclosed. McVeigh is guilty in this telling; he is not sufficient.

That is the account the videos collectively try to construct. It is not the Department of Justice account, and the official investigations never adopted the extra-bomber, sting-gone-wrong, or PATCON-incitement framing. The series exists to keep those discrepancies in view rather than to replace the courtroom record with a single new official story.