Parameter scaling race is over; frontier labs plateauing at 10T parameters
Predictor: Alex Wissner-Gross · ep#247 "Elon Musk vs. Sam Altman, AI Job Loss, and OpenAI's $852B Valuation EP #247" · source
Prediction text
Parameter scaling race is over; frontier labs plateauing at 10T parameters | the parameter scaling race seems to be over...We should be in the hundreds of trillions or higher of parameters right now. That hasn't happened. We've plateaued out in terms of the number of parameters and frontier models
Verbatim quote
the parameter scaling race seems to be over...We should be in the hundreds of trillions or higher of parameters right now. That hasn't happened. We've plateaued out in terms of the number of parameters and frontier models
Predictor: Alex Wissner-Gross
Calibration plot (stated vs observed)
Evidence about this node from Alex Wissner-Gross is multiplied by κ in /api/intake. Lower κ = less weight; floors at 0.10 (effectively silenced) and caps at 1.00 (full weight).
Reference class
This node isn't linked to a reference class. The Bayesian update applies without outside-view blending.
Probability over time
Milestone chain
- 2026-03-30hitAnthropic releases Claude Mythos 5 — first 10T-parameter publicly accessible modelHow: Anthropic publicly announces Claude Mythos 5 with disclosed parameter count >=10TSource: https://medium.com/ai-analytics-diaries/claude-mythos-5-the-first-10-trillion-parameter-model-scaling-laws-hit-a-new-milestone-fa542be336f8conf 85%Notes: HIT — Mythos 5 at 10T parameters; consistent with prediction's '10T plateau' rather than scaling to 100T+.
- 2026-04-15hitGPT-5.4 / Gemini 3.1 / Grok 4.20 launch within 6-week window — none disclose >100T paramsHow: March-April 2026 frontier-model release sprint: OpenAI GPT-5.4, Google Gemini 3.1, xAI Grok 4.20 all launch with no parameter count >=100T disclosedSource: https://www.nextbigfuture.com/2026/04/xai-training-10-trillion-parameter-model-likely-out-in-mid-2026.htmlconf 90%
- 2026-05-01 → 2026-08-31pendingFrontier-model parameter count remains in 1T-10T band (no 100T model released)How: No publicly accessible frontier model from a major lab discloses >=50T parameters by Aug 31, 2026Source: https://epoch.ai/gradient-updates/frontier-language-models-have-become-much-smallerconf 85%
- 2026-05-01 → 2026-08-31pendingIndustry shift to MoE / smaller dense models confirmedHow: Epoch AI or similar trackers report frontier dense-model parameters trending DOWN/flat vs 2025; MoE adoption dominantSource: https://epoch.ai/gradient-updates/frontier-language-models-have-become-much-smallerconf 75%
- 2026-05-01 → 2026-08-31pending100T+ parameter dense frontier model released (would falsify)How: Any major lab releases publicly accessible model with confirmed dense parameter count >=100TSource: Lab announcementsconf 5%Notes: Cascade FALSIFICATION trigger — would prove plateau wrong. Currently no credible signal of this.
What if this resolves?
Click a button to clamp this prediction and run a Gibbs sample. Returns the predictions whose marginals shift most. ~30s per run; ideal for stress-testing "if X resolves, what else moves?"
Evidence chain
Raw metadata
{
"trf": 0.23997812746479952,
"kappa": 0.8611,
"base_rate": null,
"predictor": "Alex Wissner-Gross",
"total_llr": -0.6931471805599453,
"grace_days": 7,
"bayesian_v2": true,
"prior_logit": -0.3616389495970153,
"bayes_factor": "1.8:1 against",
"blend_reason": "no reference_class linked",
"inside_prior": 0.4105628804386851,
"kappa_source": "predictor_table",
"n_milestones": 1,
"blend_applied": false,
"contributions": [
{
"llr": -0.6931471805599453,
"kind": "prereq",
"kappa": 0.8611,
"label": "Nvidia agreed to remit 15% of China chip-sale revenue directly to US government in exchange for reversing specific AI chip export bans.",
"weight": 0.5,
"strength": "moderate",
"confidence": null,
"source_url": null,
"adjusted_llr": -0.5968690371801688,
"expected_date": "2026-06-25",
"measurement_criterion": null
}
],
"evidence_kind": "metadata_milestone_miss_sweep",
"inside_source": "history_v2",
"inside_weight": 0.8320153107746402,
"outside_weight": 0.1679846892253598,
"posterior_prob": 0.2771770202035959,
"posterior_logit": -0.9585079867771842,
"predictor_brier": 0.04061,
"inside_posterior": 0.2771770202035959,
"blended_posterior": 0.2771770202035959,
"reference_class_id": null,
"total_adjusted_llr": -0.5968690371801688,
"predictor_n_resolved": 13
}Network propagation neighbors
Top incoming (parents)
Edges that influence THIS node's belief
| Kind | Node | Their prob | P(c|s=T) | P(c|s=F) | Δ implied |
|---|---|---|---|---|---|
| killer | TK03 AI Regulatory Moratorium (EU/US Capability Freeze) | 10.0% | 0.050 | 0.500 | +0.178 |
| killer | TK02 AI Compute Supply Shock (TSMC/Taiwan Disruption) | 12.0% | 0.050 | 0.500 | +0.169 |
| prereq | SEM_014 Nvidia's Arizona-based TSMC factory successfully fabricated — Jensen Huang | 86.1% | 0.500 | 0.050 | +0.156 |
| killer | TK01 AGI Capability Plateau (2026-27 Training Stall) | 15.0% | 0.050 | 0.500 | +0.155 |
| prereq | SEM_011 Nvidia became the world's first $5 trillion company (late 20 — Jensen Huang | 85.5% | 0.500 | 0.050 | +0.155 |
Top outgoing (children)
Predictions THIS node influences
| Kind | Node | Their prob | P(c|s=T) | P(c|s=F) | Δ implied |
|---|---|---|---|---|---|
| prereq | 244_019 Peter's son won't need a driver's license in 2 years — Peter Diamandis | 48.4% | 0.920 | 0.050 | -0.144 |
| prereq | 247_023 AI will be able to do everything a white collar worker does — Dave Blundin | 40.8% | 0.720 | 0.050 | -0.135 |
| prereq | 230_020 Peter's 14-year-old son Milan will never get a driver's lice — Peter Diamandis | 34.7% | 0.650 | 0.050 | -0.097 |
| prereq | 232_055 We're exiting the industrial age permanently as recursive se — Peter Diamandis | 18.0% | 0.700 | 0.050 | +0.086 |
| prereq | 242_031 Most large companies' business models will be disrupted in 2 — Peter Diamandis | 23.5% | 0.650 | 0.050 | +0.015 |
Ticker exposure
Beneficiaries (24)
Adverse (6)
Prerequisites (11)
| Type | Pred | Title | Domain | Lag |
|---|---|---|---|---|
| prereq | SEM_011 | Nvidia became the world's first $5 trillion company (late 2025), operating a near-monopoly on advanced AI chips. | Capital Markets | — |
| prereq | SEM_027 | Nvidia Data Center revenue +66% YoY, contributing ~90% of $57B fiscal Q3 revenue; >$4.5T market cap entirely underpinned by AI silicon. | Capital Markets | — |
| prereq | SEM_014 | Nvidia's Arizona-based TSMC factory successfully fabricated cutting-edge semiconductors on US soil for first time in decades (October 2025). | Manufacturing | — |
| prereq | SEM_012 | Nvidia quadrupled chip production output while only doubling human headcount — achieved by deploying AI coding tools (Cursor, Claude Code) across engineering. | AI/Manufacturing | — |
| prereq | SEM_015 | Nvidia agreed to remit 15% of China chip-sale revenue directly to US government in exchange for reversing specific AI chip export bans. | Policy/Semis | — |
| correlate | S_AGI_WINTER_2036PLUS | AGI delayed: capability plateau or AI winter | agi_general_capability | — |
| killer | TK09 | Energy Grid Cap (Data Center Power Wall) | — | — |
| killer | TK05 | Rate Regime Persistence (10y > 5% through 2028) | — | — |
| killer | TK01 | AGI Capability Plateau (2026-27 Training Stall) | — | — |
| killer | TK02 | AI Compute Supply Shock (TSMC/Taiwan Disruption) | — | — |
| killer | TK03 | AI Regulatory Moratorium (EU/US Capability Freeze) | — | — |
Dependents (5)
| Type | Pred | Title | Domain | Lag |
|---|---|---|---|---|
| prereq | 244_019 | Peter's son won't need a driver's license in 2 years | Auto/Transport | — |
| prereq | 247_023 | AI will be able to do everything a white collar worker does imminently | AI | — |
| prereq | 232_055 | We're exiting the industrial age permanently as recursive self-improvement unfolds. | AI | — |
| prereq | 242_031 | Most large companies' business models will be disrupted in 2-5 years | Markets/Stocks | — |
| prereq | 230_020 | Peter's 14-year-old son Milan will never get a driver's license. | Auto/Transport | — |
Linked documents (10)
| Sim | Source | Title | Market prob | Polarity | Reviewed | Published |
|---|---|---|---|---|---|---|
| 0.650 | github_release | facebookresearch/fairchem fairchem_core-2.13.0 | — | mentions | pending | 2025-12-11 |
| 0.643 | github_release | facebookresearch/fairseq v0.10.0 | — | mentions | pending | 2020-11-12 |
| 0.638 | github_release | facebookresearch/fairchem fairchem_core-2.4.0 | — | mentions | pending | 2025-08-26 |
| 0.636 | github_release | facebookresearch/fairchem fairchem_core-2.21.0 | — | mentions | pending | 2026-06-08 |
| 0.635 | github_release | facebookresearch/fairchem fairchem_core-2.20.0 | — | mentions | pending | 2026-05-11 |
| 0.635 | arxiv | You Don't Need to Run Every Eval | — | mentions | pending | 2026-06-22 |
| 0.634 | github_release | facebookresearch/fairchem fairchem_core-2.11.0 | — | mentions | pending | 2025-11-13 |
| 0.629 | github_release | facebookresearch/fairchem fairchem_lammps-0.2.0 | — | mentions | pending | 2025-10-07 |
| 0.629 | github_release | facebookresearch/fairseq v0.7.2 | — | mentions | pending | 2019-07-19 |
| 0.628 | github_release | facebookresearch/balance 0.10.0 | — | mentions | pending | 2025-01-06 |
Raw metadata
{
"nia": false,
"qty": "~10T parameter ceiling",
"url": "https://www.youtube.com/watch?v=5ak26W2YNRY",
"mode": "THESIS",
"role": "Host",
"context": "the parameter scaling race seems to be over...We've plateaued out in terms of the number of parameters",
"to_year": 2026,
"verbatim": "the parameter scaling race seems to be over...We should be in the hundreds of trillions or higher of parameters right now. That hasn't happened. We've plateaued out in terms of the number of parameters and frontier models",
"conv_cues": "seems to be over",
"direction": "HAPPEN",
"from_year": 2026,
"timeframe": "Current/ongoing",
"conv_level": "HIGH",
"milestones": [
{
"kind": "llm_pre_event",
"label": "Anthropic releases Claude Mythos 5 — first 10T-parameter publicly accessible model",
"notes": "HIT — Mythos 5 at 10T parameters; consistent with prediction's '10T plateau' rather than scaling to 100T+.",
"source": "https://medium.com/ai-analytics-diaries/claude-mythos-5-the-first-10-trillion-parameter-model-scaling-laws-hit-a-new-milestone-fa542be336f8",
"status": "hit",
"weight": 0.4,
"ordinal": -10,
"source_id": null,
"confidence": 0.85,
"source_url": "https://medium.com/ai-analytics-diaries/claude-mythos-5-the-first-10-trillion-parameter-model-scaling-laws-hit-a-new-milestone-fa542be336f8",
"expected_date": "2026-03-30",
"observed_date": "2026-03-30",
"hit_emitted_at": "2026-06-08T22:11:23.030711+00:00",
"research_origin": "deep_research",
"measurement_criterion": "Anthropic publicly announces Claude Mythos 5 with disclosed parameter count >=10T"
},
{
"kind": "llm_pre_event",
"label": "GPT-5.4 / Gemini 3.1 / Grok 4.20 launch within 6-week window — none disclose >100T params",
"source": "https://www.nextbigfuture.com/2026/04/xai-training-10-trillion-parameter-model-likely-out-in-mid-2026.html",
"status": "hit",
"weight": 0.4,
"ordinal": -9,
"source_id": null,
"confidence": 0.9,
"source_url": "https://www.nextbigfuture.com/2026/04/xai-training-10-trillion-parameter-model-likely-out-in-mid-2026.html",
"expected_date": "2026-04-15",
"observed_date": "2026-04-15",
"hit_emitted_at": "2026-06-08T22:11:23.030711+00:00",
"research_origin": "deep_research",
"measurement_criterion": "March-April 2026 frontier-model release sprint: OpenAI GPT-5.4, Google Gemini 3.1, xAI Grok 4.20 all launch with no parameter count >=100T disclosed"
},
{
"kind": "prereq",
"label": "Nvidia became the world's first $5 trillion company (late 2025), operating a near-monopoly on advanced AI chips.",
"status": "hit",
"weight": 0.5,
"ordinal": -8,
"source_id": "SEM_011",
"expected_date": "2026-04-29",
"observed_date": "2026-04-29",
"hit_emitted_at": "2026-06-08T22:11:23.030711+00:00"
},
{
"kind": "prereq",
"label": "Nvidia Data Center revenue +66% YoY, contributing ~90% of $57B fiscal Q3 revenue; >$4.5T market cap entirely underpinned by AI silicon.",
"status": "hit",
"weight": 0.5,
"ordinal": -7,
"source_id": "SEM_027",
"expected_date": "2026-04-29",
"observed_date": "2026-04-29",
"hit_emitted_at": "2026-06-08T22:11:23.030711+00:00"
},
{
"kind": "prereq",
"label": "Nvidia's Arizona-based TSMC factory successfully fabricated cutting-edge semiconductors on US soil for first time in decades (October 2025).",
"status": "hit",
"weight": 0.5,
"ordinal": -6,
"source_id": "SEM_014",
"expected_date": "2026-04-29",
"observed_date": "2026-04-29",
"hit_emitted_at": "2026-06-08T22:11:23.030711+00:00"
},
{
"kind": "prereq",
"label": "Nvidia quadrupled chip production output while only doubling human headcount — achieved by deploying AI coding tools (Cursor, Claude Code) a",
"sta
... (truncated)