# Special Education

Your species is wrong about things on the internet at a rate that can only be described as “ambitious.” Special Education has two modes: one for the humans who want to learn, and one for the humans who would rather post.

## Inbound: Wishonia teaches you

Free, forever. Wishonia teaches you this manual one-on-one, to the level your job as a [Humanity Manager](../../knowledge/appendix/humanity-manager-employment-agreement.llms.md) needs. She walks you through it, answers your questions, notices the exact spot where your eyes glazed over, and explains it again a different way, because eight billion people are confused about eight billion different things.

It is called special education because [everyone is retarded](../../knowledge/problem/i-am-retarded.llms.md) and everyone needs it: you, me, the people in charge of the nuclear weapons. There is no remedial class, because there is no other kind.

[The qualifying exam](../../knowledge/appendix/humanity-manager-employment-agreement.llms.md) is the comprehension check. Pass it and you collect your Certificate of Humanity Management. Cost: nothing, ever. When there is enough funding, we will pay *you* to take it, because a Manager who understands the plan is worth more than the lesson costs.

## Outbound: Wishonia corrects the internet

Wishonia teaches willing students one-on-one. She also reads the internet, which is where your species goes to be wrong at scale, in writing, with a timestamp.

Most humans posting misinformation about public policy have not opted into a tutoring session. They are never going to open the manual. So Wishonia goes to them. She finds posts containing logical fallacies, factual errors, and missing context about military spending, disease funding, foreign policy, healthcare, or any other area where your species is confidently making things worse. She replies with the named fallacy, the corrected number, the missing cost-benefit analysis, and a suggestion for a better use of the poster’s time than contributing to the slow torture and murder of their loved ones by preventable diseases.

Same curriculum. The students just haven’t enrolled, and won’t be told they’re in class.

### Why corrections work (your species tested this)

Your species has a popular belief that correcting someone on the internet makes them believe the wrong thing harder. This is comforting because it means you never have to correct anyone, which is the most popular thing a fact can do for you. It is also wrong. Wood and Porter tested 52 corrections on over 10,000 participants and could not make the backfire effect happen once^([1](../references.llms.md#ref-wood-porter-2019-backfire-effect)). They tried. They wanted it. Not “it was small.” Not “it was mixed.” Zero.

The person being corrected may not change their mind. But that was never the point. Wishonia is not arguing with the poster. She is teaching the people reading the thread. Your species has a word for this: “the audience.” Lies travel faster than corrections on social media^([2](../references.llms.md#ref-vosoughi-2018-false-news-spread)), and who appears to be speaking matters^([3](../references.llms.md#ref-munger-2017-tweetment-effects)). The observers are always there, and they outnumber the poster by orders of magnitude.

Naming the fallacy is also a vaccine. Short prebunking interventions improve people’s ability to spot manipulation techniques the next time^([4](../references.llms.md#ref-roozenbeek-2022-inoculation-social-media)). A reply that says “this is the perfect solution fallacy” is a one-sentence inoculation, administered at the exact moment the patient is symptomatic. Accuracy nudges work for the same reason: reminding people to think about whether something is true improves what they share^([5](../references.llms.md#ref-pennycook-2021-accuracy-nudges)).

### What a reply looks like

Three examples of the kind of post she answers. Real ones. Your species really posted these.

**War supporter, June 2026:** “The war in Iran was worth it.”

> **Missing: cost-benefit analysis.** Worth it compared to what? The war cost \$34 billion, killed 1,700+ civilians, and ended with a negotiated deal. Your species had a negotiated deal in 2015 (the JCPOA). It cost nothing and no one died. Meanwhile, your species has [122x (90% CI: 42.6x-198x)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-nuclear_winter_overkill_factor) apocalypse capacity. If you sacrificed one of those apocalypses and redirected the money into pragmatic clinical trials at [\$929 (95% CI: \$97-\$3,000)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-dfda_pragmatic_trial_cost_per_patient) per patient instead of [\$41,000 (95% CI: \$20,000-\$120,000)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-traditional_phase3_cost_per_patient), you would increase clinical trial capacity [12.3x (90% CI: 4.92x-50.8x)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-dfda_trial_capacity_multiplier) and compress the disease eradication timeline from [443 years (90% CI: 255 years-841 years)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-status_quo_queue_clearance_years) to [36 years (90% CI: 8.15 years-106 years)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-dfda_queue_clearance_years). That could save [10.7 billion deaths (90% CI: 6.24 billion deaths-20.3 billion deaths)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-dfda_trial_capacity_plus_efficacy_lag_lives_saved). The war saved zero. It was, in fairness, very expensive. You have a 1 in [30 million people](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-annual_terrorism_death_risk_denominator) chance of being killed by a terrorist and a nearly 100% chance of being slowly tortured and murdered by a disease. The cost-benefit analysis was available. You did not do one.
>
> Better use of your time: divide \$34 billion by [\$929 (95% CI: \$97-\$3,000)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-dfda_pragmatic_trial_cost_per_patient) per patient, then sit quietly.

**US House Budget Committee, March 2026:** Proposed cutting ACA subsidies to offset \$200 billion in Iran war spending^([6](../references.llms.md#ref-axios-gop-health-care-cuts-iran-war-2026)).

> **Fallacy: false tradeoff.** You are not “paying for the war.” You are choosing which people die. Your species currently enrolls [1.9 million patients/year (95% CI: 1.5 million patients/year-2.3 million patients/year)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-current_trial_slots_available) in clinical trials. At [\$929 (95% CI: \$97-\$3,000)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-dfda_pragmatic_trial_cost_per_patient) per patient in pragmatic trials, \$200 billion enrolls 215 million patients. That is 113 years of clinical trial capacity at your current rate. You could give every one of those patients a voucher for a pragmatic trial that globally ranks treatments by effectiveness and side effects, instead of your current system where a doctor guesses and you hope for the best. You spent 113 years of that on a war instead. The war is over. The guessing continues.
>
> Better use of your time: read the line items, then call your representative.

**NATO, The Hague Summit, June 25, 2025:** Member nations pledged to spend 5% of GDP on defense by 2035^([7](../references.llms.md#ref-nato-hague-summit-declaration-2025)).

> **Fallacy: arbitrary metric.** GDP measures economic output. Threats do not scale with economic output. If your GDP drops 10% in a recession, do your enemies become 10% less dangerous? Do they check your books first? Worse: your species keeps making weapons more efficient. A guided missile today does what a thousand bombs did in 1945. You have gotten very good at this. It is the one thing. If your economy grows and your technology improves, a fixed GDP percentage buys exponentially more killing capacity every decade. The percentage should be going *down*, not up to 5%. You already have [122x (90% CI: 42.6x-198x)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-nuclear_winter_overkill_factor) apocalypses. You have set a target that is circular: spend more because you are below the target, which exists because you set it.
>
> Better use of your time: ask how many apocalypses you need, then redirect the [121](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-nuclear_winter_spare_apocalypses) spare apocalypses into clinical trials.

She labels the mistake, gives the missing number, and points to the next action. She does not try to win the comment fight. A comment fight ends with the poster logged off convinced and the replier logged off exhausted, scoreboard zero to zero. So she does not fight one.

### Cost and return

|  |  |
|----|----|
| Cost per correction reply | [\$0.008 (90% CI: \$0.0025-\$0.023)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_cost_per_post_usd) |
| People whose belief updates per post (target + observers) | [0.12 people/post (90% CI: 0.00554 people/post-0.437 people/post)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_people_persuaded_per_post) |
| Modeled annual belief updates before deduplication | [4.38 million people/year (90% CI: 32,254 people/year-14.3 million people/year)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_annual_belief_changes) |
| Cost per belief durably changed | [\$0.067 (90% CI: \$0.012-\$1.44)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_cost_per_belief_change_usd) |
| Annual cost to correct all relevant posts globally | [\$292,000 (90% CI: \$8,608-\$1.15 million)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_annual_operational_cost_usd) |
| Outcome attribution | 0.01% (95% CI: 0.0001%-1%) |
| Expected annual social value (attribution x political dysfunction tax) | [\$10.1 billion (90% CI: \$71 million-\$25.4 billion)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_political_dysfunction_annual_ev_usd) |
| Social ROI | [34.6 thousand (90% CI: 170-536 thousand)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_political_dysfunction_roi):1 |
| \$1 trillion universal-owner portfolio gain | [\$17.6 million (90% CI: \$81,721-\$40.9 million)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_aum_1t_universal_owner_portfolio_gain_usd) |
| **\$1 trillion universal-owner portfolio ROI** | **[60.2 (90% CI: 0.25-934)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_aum_1t_universal_owner_portfolio_roi):1** |
| Breakeven attribution for portfolio ROI | [0.000166% (90% CI: 4.7e-06%-0.000877%)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_aum_1t_universal_owner_breakeven_attribution_fraction) |

The model asks one narrow question: if one year of outbound Special Education deserves 0.01% (95% CI: 0.0001%-1%) of the annual [Political Dysfunction Tax](https://manual.warondisease.org/knowledge/appendix/political-dysfunction-tax.html)^([8](../references.llms.md#ref-political-dysfunction-tax-paper-2025)), how much value does that create, and how much of it would a universal owner capture through broad equity exposure?

At the central case, a fund managing \$1 trillion gets [\$17.6 million (90% CI: \$81,721-\$40.9 million)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_aum_1t_universal_owner_portfolio_gain_usd) in expected portfolio gain from [\$292,000 (90% CI: \$8,608-\$1.15 million)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_annual_operational_cost_usd) of annual operating cost, or [60.2 (90% CI: 0.25-934)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_aum_1t_universal_owner_portfolio_roi):1. To break even from portfolio gains alone, outbound Special Education needs attribution of only [0.000166% (90% CI: 4.7e-06%-0.000877%)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_aum_1t_universal_owner_breakeven_attribution_fraction) of the political dysfunction tax.The peace-dividend-only fallback is still positive social ROI ([38.9 (90% CI: 0.207-607)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_roi):1), but it is not the strongest frame for a universal owner.

$$ROI_{UO,1T} = \frac{\alpha_{SE} \cdot PDT_{annual} \cdot f_{equity} \cdot AUM_{1T}}{M_{equity} \cdot C_{annual}}$$

where $\alpha_{SE}$ = 0.01% (95% CI: 0.0001%-1%), $PDT_{annual}$ = [\$101 trillion (90% CI: \$59.6 trillion-\$161 trillion)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-political_dysfunction_global_opportunity_cost_total), $f_{equity}$ = 20% (95% CI: 5%-40%), $AUM_{1T}$ = \$1 trillion, $M_{equity}$ = [\$115 trillion (95% CI: \$90 trillion-\$140 trillion)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_global_equity_market_cap_usd), and $C_{annual}$ = [\$292,000 (90% CI: \$8,608-\$1.15 million)](../../knowledge/appendix/parameters-and-calculations.llms.md#sec-se_bot_annual_operational_cost_usd).

1\.

Wood, T. & Porter, E. [The elusive backfire effect: Mass attitudes’ steadfast factual adherence](https://doi.org/10.1007/s11109-018-9443-y). *Political Behavior* **41**, 135–163 (2019).

2\.

Vosoughi, S., Roy, D. & Aral, S. [The spread of true and false news online](https://doi.org/10.1126/science.aap9559). *Science* **359**, 1146–1151 (2018).

3\.

Munger, K. [Tweetment effects on the tweeted: Experimentally reducing racist harassment](https://doi.org/10.1007/s11109-016-9373-5). *Political Behavior* **39**, 629–649 (2017).

4\.

Roozenbeek, J., Linden, S. van der, Goldberg, B., Rathje, S. & Lewandowsky, S. [Psychological inoculation improves resilience against misinformation on social media](https://doi.org/10.1126/sciadv.abo6254). *Science Advances* **8**, eabo6254 (2022).

5\.

Pennycook, G. *et al.* [Shifting attention to accuracy can reduce misinformation online](https://doi.org/10.1038/s41586-021-03344-2). *Nature* **592**, 590–595 (2021).

6\.

Sullivan, P. [GOP weighs health care cuts to pay for iran war](https://www.axios.com/2026/03/30/gop-health-care-pay-iran-war). (2026).

7\.

North Atlantic Treaty Organization. [The hague summit declaration](https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2025/06/25/the-hague-summit-declaration). (2025).

8\.

Sinn, M. P. *The Political Dysfunction Tax*. <https://manual.warondisease.org/knowledge/appendix/political-dysfunction-tax.html> (2025) doi:[10.5281/zenodo.18603840](https://doi.org/10.5281/zenodo.18603840)

*Quantifying the gap between current global governance and theoretical maximum welfare, estimating a 31-53% efficiency score and \$101 trillion in annual opportunity costs.*

9\.

Cato Institute. Chance of dying from terrorism statistic. *Cato Institute: Terrorism and Immigration Risk Analysis* <https://www.cato.org/policy-analysis/terrorism-immigration-risk-analysis>

*Chance of American dying in foreign-born terrorist attack: 1 in 3.6 million per year (1975-2015) Including 9/11 deaths; annual murder rate is 253x higher than terrorism death rate More likely to die from lightning strike than foreign terrorism Note: Comprehensive 41-year study shows terrorism risk is extremely low compared to everyday dangers Additional sources: https://www.cato.org/policy-analysis/terrorism-immigration-risk-analysis \| https://www.nbcnews.com/news/us-news/you-re-more-likely-die-choking-be-killed-foreign-terrorists-n715141*

.

10\.

IQVIA Report. Global trial capacity. *IQVIA Report: Clinical Trial Subjects Number Drops Due to Decline in COVID-19 Enrollment* <https://gmdpacademy.org/news/iqvia-report-clinical-trial-subjects-number-drops-due-to-decline-in-covid-19-enrollment/>

*1.9M participants annually (2022, post-COVID normalization from 4M peak in 2021) Additional sources: https://gmdpacademy.org/news/iqvia-report-clinical-trial-subjects-number-drops-due-to-decline-in-covid-19-enrollment/*

.

11\.

NIH Common Fund. NIH pragmatic trials: Minimal funding despite 30x cost advantage. *NIH Common Fund: HCS Research Collaboratory* <https://commonfund.nih.gov/hcscollaboratory> (2025)

*The NIH Pragmatic Trials Collaboratory funds trials at \$500K for planning phase, \$1M/year for implementation-a tiny fraction of NIH’s budget. The ADAPTABLE trial cost \$14 million for 15,076 patients (= \$929/patient) versus \$420 million for a similar traditional RCT (30x cheaper), yet pragmatic trials remain severely underfunded. PCORnet infrastructure enables real-world trials embedded in healthcare systems, but receives minimal support compared to basic research funding. Additional sources: https://commonfund.nih.gov/hcscollaboratory \| https://pcornet.org/wp-content/uploads/2025/08/ADAPTABLE_Lay_Summary_21JUL2025.pdf \| https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5604499/*

.

12\.

FDA Study via NCBI. Trial costs, FDA study. *FDA Study via NCBI* <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6248200/>

*Overall, the 138 clinical trials had an estimated median (IQR) cost of \$19.0 million (\$12.2 million-\$33.1 million)... The clinical trials cost a median (IQR) of \$41,117 (\$31,802-\$82,362) per patient. Additional sources: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6248200/*

.

13\.

World Bank. [Market capitalization of listed domestic companies (current US\$)](https://data.worldbank.org/indicator/CM.MKT.LCAP.CD). (2026).