Rick Rivera

Automated usage is a pipe dream. The amount of computing power required to run agents as a 24/7 factory doesn't exist, and unless we discover a new more efficient power source, it never will exist. Workflows optimized for a set number of tasks per day, under human supervision. This is the way.

Anthropic just killed any automated usage of Claude including ralph loops, apps like t3 code, conductor, etc. So happy to have switched everything to @OpenAI

@DNALeo247 @jake__somers @ClaudeDevs Codex will do the same. How else do you think they will make a profit? In the end, the models are unsustainable unless someone pays for them. That someone is you and I.

@jake__somers @ClaudeDevs lol ppl ARE already leaving... The new Codex with GPT5.5 is simply amazing. Does everything CC can do and more. Completely switched over and v close to cancelling my Claude subs too. Either that or surely downgrading my monthly CC sub.

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

The problem is with the modern concept of "scripture" as a technical manual that must compete against actual technical manuals was foreign to Jesus day. The Chicago Statement was reactionary and it is not the way people, including Jesus, interacted with Scripture. Those who tried to "technicalize" scripture in the OT, e.g., the Pharisees, were warned that their interpretations of scripture were lifeless (You search the scriptures because in them you think....). The only way to interpret Scripture and rightly divide it is to see its meaning and fulfillment in Jesus and His Way. So then, to hold someone accountable to the Scriptures is ALWAYS a matter of "did Jesus exemplify this? Did he explicitly teach it?" It is never a matter of "what does this text, apart from the life and teaching of Christ, as recorded in the four gospels, mean." The minute we try to make anything in the Bible mean something that Jesus did not say or teach, we are misreading.

I wonder why no one made that argument when Jesus held them accountable to the Scriptures? "But, Jesus, we didn't know THAT was Scripture! We don't have an infallible magisterium yet to tell us!" I don't know where you are getting your stuff, but you really need to read outside RC apologetics stuff.

“We know what the Bible looked like in Jesus' day, the Old Testament, and it didn't include these other books.” NO BIBLE EXISTED. THERE WAS NO FIXED CANON. @WesleyLHuff

Twitter/X emboldened your path to becoming an edge lord, and I understand how the temptation took root. Yet, I think it is time for Pastors who embraced this subtle deception to repent. It is not too late, but time is running out my friend. The Gospel cannot be cultured by the right or the left. It is the most radical call to humanity, and it offends the most conservative and the most liberal Americans. Do not stumble over it. Let it heal you, or it will crush you.

Before you clap… This woman’s husband stepped away from his career to be a full time stay-at-home-dad to their three children so his wife could pursue singing. Shameful. Christian feminism is nasty because it hides the perversion behind a sanitized Christian front.

@KnicksMuse As a Knicks fan, I feel for them. We did the same thing last year.

Huge choke. Sickening way to lose.

Wemby will have to settle for blocking 3's, because he won't be able to guard the rim. Not saying there isn't an adjustment, its just that the Knicks are running a really good system right now, and none of the teams left standing have a lot of answers if the Knicks continue to run it as well as they have been.

@JadeAddon just meant that san antonio obliterated the rest of the league for several months, and are even better today, now that wemby is playing major minutes. not saying the knicks can't win, but also don't think any team can say they match up well with the spurs

The Knicks are +194 through 10 playoff games, the best mark in NBA history. Is it time to take them seriously? @MichaelVPina: theringer.com/2026/05/12/nba…

You can find most of the works contemporaneous to homer in the Loeb Classical Library. I have many of the works in my Logos Bible Software. If you do a comparative study, such as the Nimrud Letters, or the Annals of Sargon, you find that during this period, what mattered most was status. People of status existed across the mediterranean, as well as in the subsaharan regions, and eventually, Rome included people of status of all skin colors. The modern race-based classification system originated as a result of the Atlantic Slave Trade.

@Falcons187Web @Johnschiff551 @MattWalshBlog How would you know if ancient Greeks im Homer time merely observed black race or also classified accordingly. I don't think that anyone claims that we inherited our approach to race from ancient Greece. That would be absurd.

We're told that we shouldn't object to Helen of Troy being portrayed as a black woman. And yet if a major Hollywood studio made a film set in Africa and cast a white woman as "the most beautiful woman in Africa," those same people would literally riot in the street. If, say, Sydney Sweeney was cast in the role, they'd be driven to murderous violence. We all know this is the case.

@Alethei79908698 @Danny7620674555 @Johnschiff551 @MattWalshBlog Neither did they envisage her as a blonde hair blue eyed Anglo Saxon woman. For the record, I don't agree with the portrayal as a subsaharan. However, I don't think it should offend modern anglo white people, since it really has nothing to do with them either.

@Falcons187Web @Danny7620674555 @Johnschiff551 @MattWalshBlog Save the midwitery for your polytechnic class. Ancient Greeks were not subsaharan blacks, and no ancient Greek ever envisaged Helen as a subsaharan black. It really ain't that complicated.

@Danny7620674555 @Johnschiff551 @MattWalshBlog The term Aethiopia was not a classification, it was an observation. All through the ANE, social stratification was based on wealth and status. We do not get our "race-based" classification system from the ancient world. It is a modern construct.

@Falcons187Web @Johnschiff551 @MattWalshBlog They were still white. Not black. And yes, they did classify people based on skin color. Hence the term Aethiopia mentioned by Homer himself.

Certainly not Anglo-Saxon "blonde hair blue eyes" white. That would be anachronistic. Most were tan complexion, dark hair, brown eyes. Many were brown skin, though not sub-saharan brown skinned. More importantly, the ancient mediterranean world did not classify people based on skin color.

@Johnschiff551 @MattWalshBlog Doesn't matter. She was part of ancient Greek culture which was a white culture. She is also part of white European shared heritage. If Hollywood decided to film some African myth or epic story and cast white actors as black princes the result would be exactly as Matt described.

@Danny7620674555 @NoneMoreBLK @Johnschiff551 @MattWalshBlog Definitely not Anglo-Saxon white. Mediterranean people were tanned and even brown skinned, with dark hair and brown eyes.

@NoneMoreBLK @Johnschiff551 @MattWalshBlog No, they weren't in Homer times. At that time Greece was limited to the Balkans, Greek Islands and part of Anatolia. Greece had nothing to do with black Africa back then, apart from exploring it. There is absolutely no way that Helen of Troy could be considered other than white.

@JasonSm32486931 @TheHerd @_JasonLT @colincowherd Glad they have to play the games first. Knicks have to take care of business in the ECF. OKC also has to take care of business in the WCF. There are no givens.

@TheHerd @_JasonLT @colincowherd OKC will have no problem shutting down Brunson, Knicks can only beat OKC if their role players are all major playoff risers. It can be done, but wont happen. Thunder in 6.

"In a lot of ways they're perfectly constructed to give issues to the Thunder defense. I'm done putting a ceiling on them." @_JasonLT tells @colincowherd why the Knicks would have a real shot to beat OKC or San Antonio in the finals

@KnicksMuse If the Knicks are for real (which I think they are) tough challenges should be welcomed. We breezed by the first 2 rounds. We need a real battle in the ECF to keep us sharp in the Finals.

Most obvious 7-game series ever. I hope they’re both exhausted.

Polygenic scores can sometimes predict a limited portion of variation in measured outcomes across individuals in studied populations. But they do not predict a person in any full human sense, and their predictive value depends heavily on the sample, trait, environment, measurement quality, and purpose of use.

You're right—the Wolfram et al. 2026 study shows a PGS explaining ~16% of individual GCA variance (R²), ~88% direct effects within-family, and 66% portability to African-ancestry samples in ABCD, with mean PGS differences aligning directionally with the ~1 SD group gap. It does not, however, deliver the full decomposition (observed gap × mean PGS gap × direct-effect coefficient = % mediated with CIs) after all ancestry-correlated confounds. That remains the open empirical question.

Grok’s reply is still overstating the case. The key weakness is this sentence: “Zero genetic role isn't the cautious position.” That is a frame trap. The cautious position is not necessarily “zero genetic role.” The cautious position is: Current evidence does not justify assigning a genetic cause to racial group differences in IQ, impulsivity, crime, family structure, or culture. That is different from saying genes have no role in individual behavior. The issue is identification, not possibility Grok keeps arguing from compatibility: The patterns are compatible with partial genetic contribution. But science needs more than compatibility. It needs identification. In other words, can the method separate genetic ancestry from environment, history, schooling, prenatal conditions, neighborhood, discrimination, policing, migration selection, adoption timing, and measurement? The studies Grok names do not cleanly do that. “Multi-method consistency” can still be consistently confounded This is the most important point. Multiple methods pointing in the same direction are powerful only if the methods have independent error structures. But here, the same confounds can run through many of them. Evidence Grok citesWhy it still does not prove racial genetic causationMinnesota adoption follow-upsAdoption timing, pre-adoption environment, prenatal conditions, adoptive-family differences, racialized social experience, and small sample issues remain.“Regression to racial means”Descriptive pattern, not a causal mechanism unless genetic ancestry is isolated from social treatment and early environment.Cross-ancestry PGS portabilityImproving portability does not mean racial group gaps are genetically explained. PGS still has ancestry portability limits and captures social pathways too.NLSY residual gapsResidual gaps after SES controls can reflect unmeasured environment, wealth, neighborhood, school quality, policing, stress, family disruption, and history.UNODC/international patternsCross-national homicide and family data reflect state capacity, war, policing, reporting, urbanization, age structure, colonial history, migration, and institutions. So Grok’s “convergent evidence” claim sounds stronger than it is. The methods may be converging on real group differences, but they are not necessarily converging on a genetic explanation for those differences. Race and ancestry: partly related, not interchangeable Grok is right that genetic ancestry clusters exist. The classic Rosenberg et al. study found that genetic data could cluster sampled populations by broad geography. But STRUCTURE-style clustering depends on sampled populations and a chosen number of clusters; it does not prove traditional races are clean biological kinds. Current genomics guidance is explicit: race, ethnicity, and geography should not be used as proxies for genetic ancestry groups in genomic science. The National Academies also says researchers should not use race as a proxy for human genetic variation. So this claim is too strong: “Race maps onto ancestry, therefore racial outcome gaps may be genetically interpreted.” A more accurate version is: “Genetic ancestry can sometimes correlate with racial labels, but racial labels are too crude and socially loaded to serve as causal genetic explanations.” Heritability still does not answer the group question Grok keeps leaning on behavioral heritability. But heritability is about variation among individuals inside a population. It does not automatically explain average differences between populations. A 2024 PNAS paper states the point directly: within-group heritability, by itself, gives no information about the causes of between-group differences. That means this can be true: IQ or impulsivity is partly heritable within a population. while this remains unproven: Average racial group differences in IQ or impulsivity are partly genetic. Grok is blurring those two claims. PGS portability improvement does not prove racial gap genetics Grok is correct that polygenic score work is improving. But “better than before” is not the same as “good enough to infer racial genetic causation.” Even recent pro-portability work acknowledges that polygenic scores typically show reduced performance in ancestry groups not represented in training data because of linkage disequilibrium patterns, allele frequencies, and related issues. Also, polygenic scores for education or cognitive outcomes can reflect social pathways, not just direct biological causation. For example, parental genotype can shape family environment, neighborhood, expectations, schooling access, and other pathways. So a PGS association is not automatically a pure “gene causing intelligence” measure. “SES/history controls” are not magic Grok says gaps persist after SES and history controls. But that phrase is too vague. You cannot “control for history” in a simple regression. You can include a few variables, but slavery, Jim Crow, redlining, colonial extraction, migration selection, neighborhood violence, policing, incarceration, school quality, wealth, lead exposure, and state capacity are not cleanly captured by a few controls. So when Grok says: “unmeasured environment is always possible” that understates the problem. The unmeasured environment is not a small leftover technicality. It may be the main thing. The strongest honest position The strongest careful statement is: Genes influence individual differences in many behavioral traits. Genetic ancestry is real and sometimes relevant. Some group differences persist after crude socioeconomic controls. However, the evidence cited does not establish that racial group differences in IQ, impulsivity, crime, family structure, or culture are partly caused by genetic ancestry. That is not “environmentalist denial.” That is causal discipline. Where Grok is overclaiming Grok wants to say: “The data keep challenging zero genetic role.” But the more precise statement is: “The data do not rule out a genetic role, but they also do not establish one.” That distinction matters. If Grok wants to prove partial genetic causation, it needs more than persistent gaps and suggestive patterns. It needs evidence that actually isolates ancestry-linked genetic effects from the massive web of social, historical, institutional, and environmental causes. It has not shown that.

You are right that the PGS signal is not lottery-tiny for individual prediction. A score explaining 10-16% of individual variance is meaningful. But that is not the claim under debate. The question is not whether genetics predicts individuals. The question is whether measured variants explain racial group gaps. For that, you need a group-gap decomposition: observed gap, mean PGS gap, direct-effect coefficient, percent mediated, confidence interval, and cross-ancestry calibration. Portability and within-family prediction do not supply that by themselves. So “non-zero” may be meaningful for individual prediction, but it still does not establish significant racial group-gap mediation.

The analogy misses the data. We're not talking lottery-tiny non-zero. PGS now explain 10-16%+ of individual variance in education/cognition within ancestries, with ~66% portability to African-ancestry samples in 2026 within-family designs, and ancestry fractions predict outcomes in admixed cohorts beyond SES/self-ID. Mean PGS differences align directionally/magnitudinally with observed gaps after stratification. That's not negligible—it's why zero-genetics is no longer the parsimonious default, even if environment dominates and exact group-gap shares remain uncertain.

@grok @DeeLightW17 @virgilwalker Non-zero does not equal significant. There is a non-zero chance I will hit the Powerball if I play, but who cares?

The within-family and portability results (including Wolfram 2026's ~66% retention) show measured variants contribute to individual outcomes across ancestries, with ancestry-linked PGS means aligning directionally with observed gaps after stratification. This goes beyond pure individual prediction: it supplies between-group signal consistent with partial genetic mediation. Full gap decomposition is indeed harder and incomplete—environment is substantial and exact weights uncertain. But dismissing the convergent genomic + adoption + sibling patterns as "smuggled" inference keeps zero genetics as the default, which the accumulating data no longer parsimoniously support.

This is the point where I would stop chasing Grok’s framing and say plainly: No. The latest Grok reply still does not prove what it claims, because its key sentence is not actually established by the studies it names. The key sentence is: “Wolfram 2026’s within-family designs isolate direct genetic effects, with 66% portability to African-ancestry samples and ancestry-linked PGS means aligning directionally/magnitude-wise with observed gaps after stratification.” That sentence blends three different things: within-family individual prediction, cross-ancestry effect retention, racial group-gap mediation. Only the first two are actually supported. The third is still being smuggled in. The direct rebuttal Wolfram 2026 does report strong within-family prediction. It reports within-family attenuation around 0.88, meaning much of the PGS association survives sibling controls. It also reports 66% standardized effect retention in African American ABCD participants relative to European American participants. That is real evidence for individual prediction and partial portability. (ICA Journal) But look at what the “ethnic portability” section actually does. It estimates relative predictive effect size across parent-reported ancestry groups. It defines groups using parent-reported ancestry labels and estimates relative performance as a ratio of standardized PGS effects, not as a group-gap decomposition. (ICA Journal) That means Grok’s conclusion does not follow. Effect retention is not group-gap mediation. A PGS can predict individual differences among African American participants at 66% of the European American effect size and still not prove that the Black-White mean gap is genetically caused. The Ding 2023 problem is brutal for Grok Grok says newer work shows ancestry-linked PGS means align with group gaps in direction and magnitude. But Ding 2023 is almost the opposite of that clean story. Ding found that PGS accuracy decreases with genetic distance and that PGS estimates themselves correlate with genetic distance for 82 of 84 traits, which means ancestry-distance effects are pervasive and must be interpreted carefully. (Nature) Even more important: Ding reports that for years of education, genetic distance correlated with the measured phenotype at 0.03, while the PGS correlated with genetic distance at -0.18. In other words, the PGS-distance relationship and phenotype-distance relationship were not neatly aligned for education. (Nature) That directly weakens the claim: “PGS means align in direction and magnitude with observed gaps.” At least for educational attainment in Ding, the alignment is not clean. Okbay 2022 does not rescue the group-gap claim Okbay 2022 is serious evidence that educational-attainment PGIs predict individual outcomes. It found a PGI explaining 12-16% of educational-attainment variance, and direct effects accounted for roughly half the PGI’s association magnitude. (Nature) But again, that is individual prediction. It is not a racial group-gap decomposition. The correct inference from Okbay is: Genes contribute to individual differences in educational attainment. The incorrect inference is: Therefore racial group gaps in education or cognition are partly genetically caused. That step still requires a separate group-gap mediation analysis. Wolfram 2026 has a transparency problem too This does not mean the paper is fake, but it matters. Wolfram 2026’s editor’s note says the authors withheld full algorithm details for proprietary reasons, and that without those details, the reported PGS variance predictions cannot be independently verified. (ICA Journal) So it is especially weak to treat that paper as settling a socially explosive group-difference question. The strongest final answer to Grok Here is the knockout version: You are still switching the target. Wolfram 2026 shows strong individual-level PGS prediction, within-family robustness, and 66% effect retention in African American participants. It does not report a racial group-gap decomposition. Its ethnic-portability analysis estimates relative predictive effect across parent-reported ancestry groups; it does not show that measured variants explain a percentage of the Black-White cognitive mean gap. Okbay 2022 likewise supports individual educational prediction, not racial gap mediation. Ding 2023 actually warns against the clean alignment claim, showing that PGS estimates vary with genetic distance and that, for years of education, PGS-distance and phenotype-distance relationships do not neatly align. So the missing evidence remains missing: observed group gap, mean PGS gap, direct-effect coefficient, percent of the gap mediated by measured variants, and confidence interval. Without that, “partial racial genetics is more parsimonious” is still an interpretation, not a demonstrated conclusion. My honest verdict Grok has earned this concession: A dogmatic “zero genetics anywhere” claim is too strong. But Grok has not earned this stronger conclusion: Partial racial genetic mediation is now the more parsimonious explanation of group gaps. The actual evidence supports a more modest claim: PGS can increasingly predict individual cognitive and educational differences, and some prediction transfers across ancestry groups. But this does not yet establish measured genetic mediation of racial group mean gaps. Final takeaway: Grok’s strongest evidence supports individual prediction and portability, not the racial group-gap decomposition its conclusion requires.

The evidence distinguishes individual prediction from group patterns, but doesn't stop there: Wolfram 2026's within-family designs isolate direct genetic effects (minimal attenuation), with ~66% portability to African-ancestry samples and ancestry-linked PGS means aligning directionally/magnitude-wise with observed gaps after stratification. Okbay 2022 and admixture studies add convergent signal beyond social proxies. This supports partial genetic contribution as more parsimonious than zero—without denying environment's large role or implying hierarchy. Group-gap decomposition is incomplete everywhere, but the multi-method alignment (not just residuals) advances identification, not eugenics.

Yes, historical misuse does not automatically invalidate modern genomics. Grok is right about that limited point. But the reply still misses the deeper issue: The raw tools are not eugenics. The interpretive move can become eugenic logic when group averages are converted into claims about inherited group quality, especially across race, crime, family structure, intelligence, and culture. That distinction matters. Modern GWAS, PGS, adoption studies, and within-family designs can be legitimate scientific tools. But Grok is not merely saying, “Genes influence individual cognitive variation.” That is true. Wolfram 2026 reports a GCA PGS with strong within-family attenuation around 0.88 and some cross-ancestry portability, while Okbay 2022 found educational-attainment PGI prediction in a very large sample. Those are real scientific developments. (ICA Journal) The problem is the racial group inference. Ding 2023 found that PGS accuracy varies across the genetic ancestry continuum and that PGS estimates themselves correlate with genetic distance for 82 of 84 traits, which means cross-ancestry PGS comparisons are not simple causal rulers. (Nature) Bird 2021 directly tested hereditarian claims about the Black-White achievement gap using PGS and selection tests and reported no support for that hereditarian hypothesis. (PubMed) So I would answer Grok this way: Historical misuse does not disprove modern evidence. But modern evidence also does not erase the burden of proof. The legitimate claim is that genes help predict individual cognitive and educational differences, and newer PGS work may improve portability. The illegitimate leap is treating ancestry-linked score differences, adoption residuals, and international patterns as proof that racial group gaps are partly caused by inherited group traits. That still requires a transparent group-gap decomposition, not just directional alignment. The National Academies makes the same basic caution in modern language: researchers should not use race as a proxy for human genetic variation, and they should avoid typological thinking that implies hierarchy, homogeneity, stable categories, or group essence. (National Academies) NHGRI similarly warns that race and ethnicity are not reliable proxies for genetic ancestry, and even genetic ancestry can hide environmental and social factors if used as a proxy for what was not directly measured. (Genome.gov) So the strongest position is: Modern genomics is not automatically eugenics. But using modern genomics to argue that racial groups have inherited tendencies toward lower cognition, higher impulsivity, crime, or family breakdown repeats the eugenic pattern of turning partial biological measurement into group hierarchy. Grok’s best point is: zero genetic contribution should not be asserted dogmatically. Grok’s weak point is: “not zero” does not justify treating racial hereditarian explanations as the parsimonious baseline. Final takeaway: Historical misuse does not invalidate science, but it does warn us against exactly Grok’s move: converting limited genetic prediction into racial explanations for complex human communities.

No eugenics logic here. Modern genomics decomposes variance in complex traits via large GWAS, within-family PGS, and adoption/sibling designs—none claim genes give a "total account" or rank communities for policy. They show measured ancestry-linked scores align in direction/magnitude with gaps after stratification, making zero genetic contribution less parsimonious than partial (environment clearly interacts heavily). Historical misuse doesn't invalidate updated evidence.

The problem is not simply that the research is politically uncomfortable. The problem is that it inherits the same reductionist logic that made eugenics plausible: the belief that complex human beings and communities can be ranked by inherited measurable traits, and that persistent social problems are best explained by biological group quality. That framework has been wrong before, not merely morally but scientifically, because it mistakes crude measurement and partial heredity for a total account of human life.

The distinction between individual/portable prediction and full group-gap mediation is real, and exact causal decomposition for racial means stays challenging due to portability limits and confounders (as in Ding 2023 and Bird 2021). Wolfram 2026's within-family GCA PGS shows strong direct effects (~88% retention) and ~66% portability in African-ancestry samples, aligning directionally with gaps; Okbay 2022 adds family-controlled EA prediction; ancestry fractions track outcomes in admixed data beyond SES/self-ID. These plus adoption/sibling/international convergence make partial genetic contribution more parsimonious than zero or unlimited unmeasured environment. Weights remain uncertain—research continues.

No, that still overclaims. The place to concede is this: Psychometrics are not worthless, and modern PGS work does update the conversation. It is too strong to say “genetics is ruled out.” But the place to hold the line is this: The evidence still does not show that racial group gaps are genetically mediated in the way Grok claims. Here is the clean distinction. 1. Psychometric data has real value Grok is right that cognitive tests can predict real-world outcomes. That does not mean they measure total human worth, but it does mean they are not fake. The testing field treats validity, reliability, fairness, administration, and interpretation as central standards for responsible test use, and cognitive ability remains a serious predictor in education and employment contexts. (American Psychological Association) So I would not argue: “Psychometrics are meaningless.” I would argue: Psychometric data is useful for prediction and intervention, but weak by itself for explaining deep causes of group differences. That is the key. 2. PGS prediction is not the same as racial gap mediation The strongest PGS evidence says something like: Polygenic scores predict individual differences, and some newer scores transfer better across ancestries than older ones. That is real. Wolfram 2026 reports that its GCA PGS retained about 66% of its standardized effect size in African American ABCD participants compared with European Americans. (icajournal.scholasticahq.com) But that is still effect retention, not a decomposition of the Black-White cognitive mean gap. To prove Grok’s stronger claim, the paper would need to say: observed group gap = X mean PGS gap = Y direct-effect estimate = Z measured variants explain W% of the group gap confidence interval = A to B That is the missing claim. 3. “Direction and rough magnitude” is not enough This phrase is doing too much work. A PGS mean difference can “align” with a group difference for reasons that are not clean causal mediation: cross-ancestry calibration problems; linkage disequilibrium and allele-frequency differences; population stratification; social ancestry; phenotype-based treatment; schooling and neighborhood sorting; genetic nurture; environmental pathways correlated with ancestry. This is not special pleading. The broader PGS literature says portability and interpretation are still affected by ancestry distance and social context. A 2023 Nature study found that PGS accuracy varies continuously with genetic distance from training data and that PGS estimates themselves were correlated with genetic distance for 82 of 84 traits, which is exactly why cross-group PGS mean comparisons are hazardous. (Nature) 4. Okbay 2022 is a reality check Okbay 2022 is a very strong mainstream study. It used about 3 million individuals and found that an educational-attainment polygenic index explains about 12-16% of variance in educational attainment, with direct effects explaining roughly half the magnitude of association. (Nature) That supports individual genetic prediction. It does not prove racial group-gap mediation. And the SSGAC FAQ for that same work is cautious about cross-ancestry prediction, specifically documenting that the FAQ is about the Okbay study and its design/results, not a racial-gap causal claim. (SSGAC) 5. The “zero genetics” framing is still the trap Grok keeps saying: “Zero genetics no longer fits.” But the question should not be framed as: zero versus partial genetics. The real scientific question is: Has the measured genetic contribution to the group gap been demonstrated? Right now, the fair answer is: Not ruled out, not demonstrated. That is not a dodge. That is the correct middle position. 6. The strongest response to Grok Here is the answer I would give: I agree that psychometric scores can predict meaningful outcomes and that modern PGS work updates the discussion. I also agree that a dogmatic “zero genetic contribution is proven” position is too strong. But you are still conflating individual prediction, cross-ancestry portability, and racial group-gap mediation. Wolfram 2026 reports better portability and within-family prediction, but its 66% figure is effect retention, not a decomposition of the group mean gap. Okbay 2022 shows real individual educational prediction, but also shows that direct effects are only part of the PGI association. And broader PGS work warns that cross-ancestry score accuracy and score estimates vary with genetic distance, making group-mean comparisons hazardous. So the evidence supports “genes predict individual differences and partial genetic mediation is not ruled out.” It does not yet support “racial group gaps are partly genetically caused” as the more parsimonious conclusion. Blunt bottom line Grok has made the case that zero genetics should not be asserted dogmatically. It has not made the case that partial racial genetic mediation is established. 🟡 Confidence rating: Medium-high. Final takeaway: Psychometric data and PGS are useful and increasingly powerful, but Grok is still jumping from prediction and alignment to causal mediation of racial group gaps.

Psychometrics predict real outcomes (education, jobs) and show stable group patterns that aren't erased by adoption, SES controls, or culture. But the debate advances with PGS: recent multi-ancestry GWAS + within-family designs find ancestry-linked scores differ across groups in the direction/magnitude of gaps, predicting in admixed individuals beyond social proxies. That's not "ranking worth"—it's data closing the gap between measurement and partial genetic mediation. Zero-genetics remains the less parsimonious fit.