Skip to content
LongtermWiki

Quantitative Claims

Quantitative claims are numbers, percentages, dollar amounts, and timeline predictions extracted from content. Claims marked as "Notable" have importance indicators like "catastrophic", "surprising", or "only" in their context.

15,414Total Claims
2463Notable
8935Percentages
2326Dollar Amounts
1415Large Numbers
1250Multipliers
What are quantitative claims? Numbers, percentages, dollar amounts, and timeline predictions extracted from content. Notable claims have importance indicators like "catastrophic", "surprising", "only", etc. These are candidates for extraction as standalone insights.
$1BNotable
Dollar Amounts
...** | Surprising finding | "More data sometimes hurts" | | **quantitative** | Number or statistic | "$1B+/year spent on RLHF" | | **disagreement** | Expert conflict | "Researchers disagree on X" | | **neg...
indexLine 87
$100KNotable
Dollar Amounts
...------| | **US dominance** | 57% of top researchers | US remains primary hub | | **Policy risk** | \$100K H-1B fee | May accelerate brain drain | | **Key uncertainty** | Immigration policy | Critical infle...
research-reportsLine 129
$150BNotable
Dollar Amounts
...028) | Gartner | 2024-2028 | | Potential revenue share by 2035 | ~30% of enterprise app software (~\$150B) | Gartner | 2035 | | Organizations with significant investment | 19% | Gartner poll (Jan 2025, n=3...
agentic-aiLine 43
$100 millionNotable
Dollar Amounts
...ce across diverse cognitive domains. With over 100 billion parameters and training costs exceeding \$100 million, these systems represent unprecedented computational achievements that have shifted AI safety from...
language-modelsLine 21
$152.41Notable
Dollar Amounts
...ds MCPMark performance at approximately \$127.46 per benchmark run, compared to Claude Sonnet 4 at \$152.41. Organizations recognize the potential for significant productivity gains through automated digital...
tool-useLine 172
$100,000Notable
Dollar Amounts
...from 500,000 files in 2023 to a projected 8 million by 2025, with businesses losing an average of \$100,000 per deepfake-related fraud incident. The stakes are substantial: if AI provides significant capabi...
misuse-risksLine 19
$2BNotable
Dollar Amounts
...multiple specialized domains | Scaling laws, algorithmic improvements | | **Defense Investment** | \$2B+ annual cybersecurity AI; 3-5x growth occurring | Major enterprise adoption | 50% of critical infra...
misuse-risksLine 666
$150 millionNotable
Dollar Amounts
...billions. Wikipedia, one of the most successful epistemic public goods, operates on approximately \$150 million annually while supporting hundreds of millions of users—a funding ratio of roughly \$0.09 per month...
solutionsLine 552
$100BNotable
Dollar Amounts
...ation value**: Potential to automate most cognitive labor **Implications**: - Massive investment (\$100B+ annually) - Fierce competition (OpenAI, Google, Anthropic, Meta, etc.) - International competition...
case-for-xriskLine 177
$10Notable
Dollar Amounts
...median AGI arrival between 2028-2032, with development concentrated among 3-4 major labs investing \$10-100B+ annually. This concentration creates significant coordination challenges and [racing dynamics...
agi-developmentLine 15
$78Notable
Dollar Amounts
...e significant safety challenges. Training costs have grown 2.4x/year since 2016, with GPT-4 costing $78-100M and Gemini Ultra $191M. Frontier models now demonstrate in-context scheming capabilities, with...
large-language-modelsLine 3
$1.6MNotable
Dollar Amounts
...s that training costs can be reduced by an order of magnitude—achieving competitive performance at \$1.6M compared to GPT-4's \$100M—by activating only a fraction of parameters per inference. OpenAI's o1 r...
large-language-modelsLine 44
$100MNotable
Dollar Amounts
...e reduced by an order of magnitude—achieving competitive performance at \$1.6M compared to GPT-4's \$100M—by activating only a fraction of parameters per inference. OpenAI's o1 represents a different parad...
large-language-modelsLine 44
$1BNotable
Dollar Amounts
...ency gains will dominate | Compute arms race continues | DeepSeek at \$1.6M vs. Grok-4 hardware at \$1B | ### Research Priorities Leading safety organizations identify these critical research areas: *...
large-language-modelsLine 163
$100MNotable
Dollar Amounts
...research professionalized significantly between 2020-2024, with funding growing from approximately \$100M to \$100M annually and dedicated researchers multiplying several-fold, the gap between capabilities...
mainstream-eraLine 19
$100MNotable
Dollar Amounts
...rofessionalized significantly between 2020-2024, with funding growing from approximately \$100M to \$100M annually and dedicated researchers multiplying several-fold, the gap between capabilities and safet...
mainstream-eraLine 19
$10Notable
Dollar Amounts
...r, though some analyses put "trustworthy AI research" funding at only \$10-130M annually. For context, philanthropic funding for climate risk mitigation was approximately \$1...
mainstream-eraLine 234
$192MNotable
Dollar Amounts
...\$100M+ | Largest documented 2024 model | | Gemini 1.0 Ultra | 2024 | approximately 2×10²⁵ FLOP | \$192M | Most expensive confirmed training | | Llama 3.1 405B | 2024 | approximately 1×10²⁵ FLOP | approxi...
compute-hardwareLine 118
$100MNotable
Dollar Amounts
...on ### Current GPT-4 Training Costs **Initial Training (2023)**: - Official estimate: "More than \$100M" (Sam Altman) - Epoch AI hardware/energy only: \$40M - Full cost estimates: \$78-192M depending on...
compute-hardwareLine 382
$40MNotable
Dollar Amounts
...ng (2023)**: - Official estimate: "More than \$100M" (Sam Altman) - Epoch AI hardware/energy only: \$40M - Full cost estimates: \$78-192M depending on methodology **GPT-4-Equivalent Training Costs (Optim...
compute-hardwareLine 383
$78Notable
Dollar Amounts
...ate: "More than \$100M" (Sam Altman) - Epoch AI hardware/energy only: \$40M - Full cost estimates: \$78-192M depending on methodology **GPT-4-Equivalent Training Costs (Optimized)**: - Q3 2023: ~\$20M (...
compute-hardwareLine 384
$100 billionNotable
Dollar Amounts
...& Valuations ### AI Company Valuations (Private Market) **Top Valuations (2025)** - **OpenAI**: \$100 billion (most valuable private company ever) - Revenue: \$12 billion (2025), up from \$3.7 billion (2024)...
economic-laborLine 62
$12 billionNotable
Dollar Amounts
...Valuations (2025)** - **OpenAI**: \$100 billion (most valuable private company ever) - Revenue: \$12 billion (2025), up from \$3.7 billion (2024) - Annualized revenue: \$13 billion (July 2025) - **Anthropic...
economic-laborLine 63
$3.7 billionNotable
Dollar Amounts
...AI**: \$100 billion (most valuable private company ever) - Revenue: \$12 billion (2025), up from \$3.7 billion (2024) - Annualized revenue: \$13 billion (July 2025) - **Anthropic**: \$183 billion → \$350 bill...
economic-laborLine 63
$100MNotable
Dollar Amounts
...perations - **High performers**: 75% scaling or scaled AI vs. 33% of other organizations - **Under \$100M revenue**: Only 29% at scaling phase *Source: McKinsey State of AI*...
economic-laborLine 235
$98 billionNotable
Dollar Amounts
...umulative) | China leads in state funding | | **2025 AI Capital Spending** | Not specified | Up to \$98 billion projected | - | **Key Finding**: The US private sector invests nearly 12x more than China's privat...
geopoliticsLine 39
$5.6 millionNotable
Dollar Amounts
...chieved performance parity with OpenAI's o1 while operating at 15x reduced cost, training for just \$5.6 million. The open ecosystem "immensely accelerated in terms of capabilities, rivaling closed models on most...
lab-behaviorLine 427
$100 billionNotable
Dollar Amounts
...ent safety funding represents just 0.0004% of global GDP, while AI capabilities investment exceeds \$100 billion annually. This creates significant questions about whether AI safety research can develop adequate...
safety-researchLine 16
$131.5BNotable
Dollar Amounts
...*~\$10 million** public sector safety research - AI safety funding: **0.0004% of global GDP** vs **\$131.5B** in AI startup VC funding (2024) - Only **1-3%** of AI publications concern safety issues **For c...
safety-researchLine 87
$85KNotable
Dollar Amounts
...llowship | \$10K/year + tuition | 5 years | AI safety PhD research | | Google PhD Fellowship | \$85K/year | Variable | AI research including safety | | Global AI Safety Fellowship | Up to \$30K | 6 mo...
safety-researchLine 135
$30KNotable
Dollar Amounts
...hip | \$85K/year | Variable | AI research including safety | | Global AI Safety Fellowship | Up to \$30K | 6 months | Career transitions | | Anthropic Fellows Program | \$2,10...
safety-researchLine 136
$110Notable
Dollar Amounts
...AI existential risk, misuse 15-35%, and structural risks 10-25%. Based on 2024 funding analysis ($110-130M total), recommends rebalancing toward governance (currently underfunded by $15-20M) and agen...
ai-risk-portfolio-analysisLine 23
$100MNotable
Dollar Amounts
...n framework, addressing the critical question: How should the AI safety community allocate its \$100M+ annual resources across different risk categories? All estimates carry substantial uncertainty (±5...
ai-risk-portfolio-analysisLine 37
$10Notable
Dollar Amounts
...ernment (US/UK/EU) | \$32.4M | ~25% | Policy-aligned research | | Other (foundations, corporate) | \$10-20M | ~10% | Various | The breakdown by research area reveals significant concentration in interpr...
ai-risk-portfolio-analysisLine 110
$40MNotable
Dollar Amounts
...s://www.openphilanthropy.org/request-for-proposals-technical-ai-safety-research/) commits at least \$40M to technical AI safety, with potential for "substantially more depending on application quality." P...
ai-risk-portfolio-analysisLine 334
$15Notable
Dollar Amounts
...t | |--------|-------------|----------|----------|----------------| | **Capability Observatory** | \$15-35M | 12-18 months | Critical | 90% coverage of capability signals | | **Independent Behavioral Eva...
warning-signs-modelLine 267
$30Notable
Dollar Amounts
...months | Critical | 90% coverage of capability signals | | **Independent Behavioral Evaluation** | \$30-70M | 18-36 months | Critical | 70% coverage of behavioral risks | | **AI Incident Database** | \$8...
warning-signs-modelLine 268
$8Notable
Dollar Amounts
...30-70M | 18-36 months | Critical | 70% coverage of behavioral risks | | **AI Incident Database** | \$8-20M | 6-12 months | High | 95% coverage of incident signals | | **Social Impact Tracker** | \$10-25...
warning-signs-modelLine 269
$100MNotable
Dollar Amounts
...:10 | | **International Coordination** | \$50M | \$500M | 1:10 | | **Crisis Management Systems** | \$100M | \$1B | 1:10 | ## Research Gaps and Future Directions ### Critical Unknowns 1. **Multi-system i...
autonomous-weapons-escalationLine 281
$1BNotable
Dollar Amounts
...*International Coordination** | \$50M | \$500M | 1:10 | | **Crisis Management Systems** | \$100M | \$1B | 1:10 | ## Research Gaps and Future Directions ### Critical Unknowns 1. **Multi-system interact...
autonomous-weapons-escalationLine 281
$50MNotable
Dollar Amounts
...lue | |----------|-------------|----------|----------------| | Multi-system interaction modeling | \$50M over 3 years | High-fidelity simulation capabilities | Critical for risk assessment | | Circuit bre...
autonomous-weapons-escalationLine 299
$100MNotable
Dollar Amounts
...h-fidelity simulation capabilities | Critical for risk assessment | | Circuit breaker technology | \$100M over 2 years | Deployable pause mechanisms | High near-term impact | | Attribution forensics | \$75...
autonomous-weapons-escalationLine 300
$200Notable
Dollar Amounts
...---|-----------------|----------| | AI-enabled DNA synthesis screening | Evasion (-1.5 to -2.5x) | \$200-400M/year | High | **Critical** | | NIST/CAISI biosecurity standards | All domains (-0.3 to -0.5x)...
bioweapons-ai-upliftLine 264
$50Notable
Dollar Amounts
...0M/year | High | **Critical** | | NIST/CAISI biosecurity standards | All domains (-0.3 to -0.5x) | \$50-100M/year | Medium | High | | Frontier BDT evaluation requirements | Novel agent design (-0.5 to -1...
bioweapons-ai-upliftLine 265
$38Notable
Dollar Amounts
...ccine** | \$2B | 20-40% | \$50-100M | Very High | | **International monitoring** | \$300M | 3-8% | \$38-100M | Very High | DNA synthesis screening emerges as the most cost-e...
bioweapons-attack-chainLine 172
$1Notable
Dollar Amounts
...e | Timeline | |-------------|------------|----------|----------| | **Severity** | High-Critical | \$1-5T projected annual losses at Level 4 | 2029-2033 | | **Likelihood** | Very High | Level 3 already...
cyberweapons-attack-automationLine 43
$3Notable
Dollar Amounts
...- **Full Autonomy**: End-to-end campaign execution, strategic target selection - **Impact Scale**: \$3-5T annual losses projected, critical infrastructure vulnerable - **Response**: International cyber...
cyberweapons-attack-automationLine 116
$500BNotable
Dollar Amounts
...evel 3** | \$500B | \$1.5-2T | 3-4x | Persistent campaigns, evasion capabilities | | **Level 4** | \$500B | \$3-5T | 6-10x | Mass coordination, critical infrastructure targeting | ### Defense Investment G...
cyberweapons-attack-automationLine 151
$3Notable
Dollar Amounts
...| \$500B | \$1.5-2T | 3-4x | Persistent campaigns, evasion capabilities | | **Level 4** | \$500B | \$3-5T | 6-10x | Mass coordination, critical infrastructure targeting | ### Defense Investment Gap Ana...
cyberweapons-attack-automationLine 151
$20BNotable
Dollar Amounts
...id="50c35f1a31a6862b">FireEye Mandiant, government agencies | | **Infrastructure Hardening** | \$20B | \$50-100B | 2.5-5x | Critical infrastructure owners | **Key Finding**: Defense is currently unde...
cyberweapons-attack-automationLine 160
Data generated: 1/23/2026, 1:17:59 AM