Social Media Curation

The week's real topic is free speech and whether corporations and platforms have the right to deplatform people, and the claim that unrestricted free speech caused the rise of the Nazis is false; a crushed post World War I economy and rising fascism across Europe did that.

Social media debates present a false dichotomy of total unrestricted free speech versus a trusted third party controlling all speech, even though yelling fire in a theater is already legal, just accountable afterward for the incitement it causes.

Social networks are multi sided markets with a producer and a consumer of information, just like eBay, Airbnb or Uber, where the platform already regulates both sides through things like buyer and seller protections and power seller ratings.

Removing predatory profiles that target children is easy platform regulation, but self regulation that works fine in a small Telegram channel or Reddit community gets much harder once a platform reaches billions of users.

Alice sharing a fake news article claiming Trump invoked the Insurrection Act of 1807, part of a QAnon conspiracy narrative, illustrates how misinformation spreads through a social feed.

Bob has no real way to judge Alice's share's credibility, and without a mental inoculation, a mind vaccine of critical thinking skills, his defenses eventually get overwhelmed and he shares the false claim onward too.

Disinformation is often manufactured by foreign powers specifically to divide people and attack the concept of truth itself, and relying only on individual thinking skills fails at scale since the tools a platform gives people matter just as much.

Platform fact checkers are unaccountable, chosen by no one, yet can delete an account like Alice's, and they are not infallible or free of context, especially once they start treating political opinions as fake news.

The fix proposed is bottom up rather than top down: distinguish content that is sourced and verified from content that is anonymous before anyone judges whether it is true.

Decentralized identifiers and KYC verified accounts, working like Twitter's blue check mark, would also curb crypto giveaway scams that impersonate figures like Steve Wozniak and Ripple's Brad Garlinghouse.

Anonymous and verified content should visually look different in the feed, and sharing itself should range from an unvouched share to signing and vouching real reputation or money behind it.

A proof of work or verifiable delay function requirement could rate limit sharing, adding real cost and friction to spreading an unverified claim.

Sharing with added context or a crowdsourced request for more information beats a bare share, illustrated by an unverified claim that a Massachusetts town banned The Odyssey from its school curriculum.

Crowdsourced annotation and bias pairings, like tupling a CNN article with a Fox News article on the same topic, gamify getting closer to the truth.

Token incentives could reward people for linking related news stories into a shared news graph, and showing Alice that knowledge graph of where her article sits relative to others before she shares is a powerful moderation tool.

AI can parse articles at scale to surface common themes and could even flag whether someone actually read a 5000 word article rather than just the headline before sharing it.

Rewarding people for pairing opposing viewpoints together fights the confirmation bias and echo chambers most people fall into by default, clearly visible in the anti vax movement's selective evidence seeking.

Once people believe the institutions that would investigate election fraud are themselves corrupt, any evidence those institutions produce gets dismissed, creating a self reinforcing feedback loop of radicalization.

The Dunning-Kruger effect explains why everyone suddenly cites Benford's law and claims expertise on voting machine audits after watching one video with limited context.

University studies show people are far more likely to share a negative story, like one about MS-13, than a positive story about an immigrant, revealing a systemic sharing bias behind ethnic cleansing and racism.

A well designed platform should treat one sided, repetitive consumption of the same viewpoint as a system failure, though none of these proposed fixes involve censorship or banning speech.

Community curated checklists covering things like confirmation bias and domain expertise, run before sharing, could earn a voucher real social capital or kudos, a far more elegant system than a simple thumbs up or down.

Bounties could reward rigorous critical analysis of shared claims the way software security audits and formal verification do, drawing on Daniel Dennett's writing on thinking tools for cognitive bias, with AI curation layered on top since AI carries no inherent human bias.

Certain baseline rights, like the ability to communicate with friends, should never be restricted, and since the Oklahoma City bombing predated Facebook, deplatforming a Nazi just pushes them onto Signal instead of removing the danger.

Cardano's own Telegram channels set terms of conduct, like banning trading talk from a channel meant for protocol discussion, but getting kicked out never strips an ADA holder's rights to stake, vote or build, since those are guaranteed by code.

Parler could compete with Twitter on its own speech rules, but losing Apple, Google and Amazon Web Services access wiped out its mobile market and website host in one shot, the equivalent of stripping an ADA holder's staking and voting rights.

Separate crypto trading and non trading Telegram channels can coexist without one dominating the other, but handing a handful of companies the power to decide who can run a modern business through app store, hosting or payment access is dangerous.

Deradicalization requires engagement, not isolation and banning; in the 1920s and 30s it was demonization and a lack of engagement, not free speech, that emboldened the Nazis to become more militant.

Cutting people off from communication only legitimizes their extreme views and pushes them toward a Ted Kaczynski style cabin in the woods, plotting violence in isolation rather than being talked down.

Combining moderated communities that have terms of conduct with an engagement track for people who don't fit in, using bias pairing as one of the most powerful tools, keeps isolated extremists from being pushed together instead of offered a path to redemption.

Daryl Davis, an African American man, has convinced over 200 Ku Klux Klan members to leave the organization by engaging them directly, and keeps hundreds of their robes as proof in his home.

A good platform should identify which users are trapped in self reinforcing silos versus those with healthy idea flow and build bridges between them, since profit maximizing algorithms that herd people into radicalized silos amount to a kind of algorithmic malfeasance.

A marketplace of competing fact checker plugins, paired as tuples like PolitiFact against The Blaze alongside think tanks and NGOs, keeps any single fact checker from having unilateral power over what counts as true.

Claims like whether aliens landed at Roswell are true disjoints, cases where two sides will never agree, and the only fix is requiring the presented evidence and methodology rather than trusting an anointed fact checker's word alone.

It is far easier to just click share or hit a thumbs up than to do the actual work of verifying a claim, and without that work these systems become propaganda machines for sovereign psy ops campaigns run by the CCP, Russia and Western governments alike.

A cheap anti vaccine disinformation campaign out of China or Russia can keep Western economies shut down longer, causing trillions of dollars in damage for a few million dollars of bot activity.

Anyone certain about an anti vax belief should ask honestly whether they sought alternative viewpoints or have real domain expertise, especially when citing vaccine injury data from the 1950s or 1970s while ignoring decades of medical progress since.

Cognitive biases are hardwired into how human brains work; even doctors misdiagnose patients and professors teach biased material because of them.

The January 6th Capitol riot was a harmful example of social media's failures, though far from the most harmful in history, and the real fix is a multimodal approach: vouching labels, fact check marketplaces, slower reactions, and always asking what the evidence shows.

Freedom of association and expression should never be restricted just for holding unpopular ideas, though vilifying an entire group is also wrong, as when 19 articles counted in a single day equated all 74 million Trump voters with Nazi sympathizers.

The blockchain space already has much of the toolbox needed here: proof of work throttling, mechanism design, decentralized identifiers, reputation systems, and knowledge graphs.

People who disagree are usually not evil, they simply hold different values, and empathy for those values, not necessarily agreement, is what is required to move forward.

Many Trump voters have legitimate grievances, from globalism hollowing out midwestern factory towns to Bill Gates and Jeff Bezos growing wealthier while publicly backing higher taxes as ordinary Americans got poorer.

The rise of Nazism mirrors this dynamic: a lost world war, economic collapse, angry veterans, and fear that the Bolshevik revolution in Russia would spread into Germany.

Most Ku Klux Klan members inherited their beliefs from their fathers and grandfathers the same way people inherit their family's religion, rather than choosing hate on their own.

The 74 million who voted for Trump and the 80 million who voted for Biden share nearly the same complaints about a rigged, unfair system, just different proposed fixes.

Wealth acts like armor against oppression, similar to how armed protesters tend to be treated more carefully by police than unarmed ones.

A well built information platform can create collective empathy as an emergent property, and that collective empathy, not deplatforming or criminalizing thought, is what actually reduces radicalization and violence over time.

The crypto industry's toolbox of incentive engineering and AI, building on research out of Stanford, MIT and Cornell, could even let a community evolve a platform's own rules through a nomic style game.

Sudden visceral anger while watching this video is probably a sign of having been propagandized somewhere else, not a fair reaction to what was actually said here.

Many in Trump's coalition come from deeply disadvantaged backgrounds, including addiction in the family, abuse, and other hard circumstances.

Trump built his coalition by telling overlooked people he would fight for them, which is why attacks on him felt personal in a way attacks on Bush or Obama never did, and he deserves some credit for exposing the political class's bipartisan hypocrisy and permanent warfare culture.

Had Trump truly believed in election fraud, he should have used his platform to reform how America runs elections rather than fight to keep the presidency, a fight that instead damaged his own party and squandered a political dynasty for his children.

Trump is a narcissist, dubbed the orange goblin here, and has never had this backing, despite an occasional overlap on conservative or libertarian values.

The most disadvantaged members of Trump's coalition, oil rig workers, ex felons and wounded veterans, deserve deep empathy rather than blanket condemnation.

Unlike Mitt Romney, Jeb Bush or Hillary Clinton, none of whom know real hardship, Trump convinced these voters he understood their pain, from war injuries to an unfair divorce, despite coming from the same privileged background himself.

Calling these voters Nazis ignores real suffering; some of them are our own relatives, and social media could be humanity's first real superpower for building collective empathy if it is built correctly.

The video is meant to be useful and a little therapeutic, and anyone offended by it is asked to think deeply about where their own beliefs come from.