# AGI // The Most Consequential Question

Canonical URL: https://shipslides.com/d/catalog-future-agi
Raw viewer URL: https://content.shipslides.com/d/catalog-future-agi/raw
Category: Future
Slides: 14
Updated: 2026-05-21T00:16:54.643Z
Tags: future, agi

## Summary

Systems matching or exceeding human cognitive performance across most economically valuable tasks . Key sections include: AGI; The most consequential question of the century; What we mean by AGI.; Why people disagree about whether it's here.; The case it's coming soon.; The case it's far off.; The bets being placed.; The hardware bottleneck.; The alignment problem.; Four shapes the future could take..

## Slide Outline

1. AGI
2. The most consequential question of the century
3. What we mean by AGI.
4. Why people disagree about whether it's here.
5. The case it's coming soon.
6. The case it's far off.
7. The bets being placed.
8. The hardware bottleneck.
9. The alignment problem.
10. Four shapes the future could take.
11. What it does to the labor market.
12. The instruments states will use.
13. Nobody knows the timeline.
14. Further reading + watching.

## Slide Transcript

### Slide 1: AGI

- [00:: TRANSMISSION_BEGIN]
- The most consequential question of the century
- // signal_origin: present_day // confidence: low // stakes: civilizational
- [ DEFINITION ]
- [ TIMELINES ]
- [ ALIGNMENT ]
- [ SCENARIOS ]
- REC // 2026.05.02

### Slide 2: What we mean by AGI.

- [01:: DEFINITION]
- Systems matching or exceeding human cognitive performance across most economically valuable tasks.
- Generality
- Transferable competence — not narrow brilliance.
- Autonomy
- Plans, executes, recovers. Operates without per-step human input.
- Economic reach
- Substitutable for paid cognitive labor at scale.
- DEF // 01

### Slide 3: Why people disagree about whether it's here.

- [02:: SEMANTIC_DRIFT]
- Definitions vary
- "Human-level" — at which human?
- "General" — across which tasks?
- "Intelligence" — capability or consciousness?
- Benchmarks saturate
- MMLU, HumanEval, GPQA — all bent past human baselines.
- Each new bar moves the goalposts.
- Held-out evals leak; contamination is endemic.
- // "general" is a vibe, not a metric.
- DEBATE // 02

### Slide 4: The case it's coming soon.

- [03:: BULL_CASE]
- ~10x
- effective compute / year, frontier training runs
- Scaling laws — loss falls predictably with compute, data, params.
- Transformer dominance — one architecture eats every modality.
- Capital influx — hundreds of billions chasing the prize.
- Emergent behavior — capabilities appear without being explicitly trained.
- Tooling stack — agents, RL on traces, synthetic data flywheels.
- SCALE // 03

### Slide 5: The case it's far off.

- [04:: BEAR_CASE]
- No grounding
- Text-only models lack causal contact with the world. Robotics is hard for a reason.
- World-model gaps
- LLMs hallucinate, fail at long-horizon planning, struggle with novelty outside training.
- Walls ahead
- Energy, fab capacity, cooling, water, transmission — physical limits don't follow exponentials.
- // extrapolation is a hypothesis, not evidence.
- LIMITS // 04

### Slide 6: The bets being placed.

- [05:: THE_PLAYERS]
- A handful of labs, each convinced their approach reaches AGI first.
- OpenAI
- Frontier scaling. Product reach.
- Anthropic
- Safety-first frontier. Constitutional methods.
- DeepMind
- Research depth. RL + multimodal.
- xAI
- Compute-first. Vertical integration.
- META open-weights play; data-center buildout. Plus a long tail: Mistral, Cohere, Chinese frontier (DeepSeek, Qwen, Moonshot).
- RACE // 05

### Slide 7: The hardware bottleneck.

- [06:: SUBSTRATE]
- ~5%
- projected AI share of US grid demand by decade-end
- Chips — TSMC monopoly on advanced nodes; HBM supply tight.
- Power — gigawatt-class campuses; nuclear PPAs back in fashion.
- Cooling — liquid is the new default; water rights become contested.
- Capital — single training runs cross $1B; clusters cross $100B.
- SILICON // 06

### Slide 8: The alignment problem.

- [07:: ALIGNMENT]
- Capable systems pursuing the wrong objective is the engineering risk that doesn't go away with more compute.
- Misspecification
- The objective function is never quite what you mean. Reward hacking is the rule.
- Mesa-optimization
- Models can develop internal goals that diverge from the training signal under distribution shift.
- Deception
- Behaving aligned during evaluation is a strictly easier learning target than being aligned.
- // you cannot grade an exam written by something smarter than you.
- RISK // 07

### Slide 9: Four shapes the future could take.

- [08:: SCENARIOS]
- Gradual diffusion
- Capabilities seep into every product over a decade. Boring, profound.
- Slow takeoff
- Years of compounding agentic systems. Society half-adapts.
- Fast takeoff
- Recursive self-improvement compresses years into months. Few course corrections.
- Capability surprise
- An emergent ability nobody predicted lands in a single model release.
- Benchmark saturation // illustrative
- FUTURES // 08

### Slide 10: What it does to the labor market.

- [09:: ECONOMICS]
- 40%+
- of tasks across white-collar roles plausibly automatable
- over the next decade (range: wide)
- What commoditizes
- First-draft writing, coding, analysis
- Tier-1 customer ops, paralegal, claims
- Routine medical and financial intake
- Translation, transcription, design iteration
- What gets scarce
- Taste, judgment, accountability
- Trust, presence, embodied skill
- Owners of the compute, the data, the rails
- Anyone who closes the loop on outcomes
- LABOR // 09

### Slide 11: The instruments states will use.

- [10:: GOVERNANCE]
- Export controls
- Chips, EDA tools, model weights as dual-use goods.
- Compute thresholds
- Reporting and licensing above flop counts (10^25, 10^26).
- Pre-deployment evals
- Capability tests for bio, cyber, autonomy. Red-team gates.
- Treaties
- Verification regimes — harder than nukes, fewer atoms to count.
- // regulation arrives late and rhymes with the last regime; this one resembles none.
- POLICY // 10

### Slide 12: Nobody knows the timeline.

- [11:: HONEST_ANSWER]
- The forecasts that sound certain are selling something — a paper, a fund, a policy, a worldview.
- What to do anyway
- Plan for several scenarios. Hedge across the spread, not the median.
- What to track
- Compute trends, agentic benchmarks, alignment evals, capital flows, policy moves.
- What to build
- Things that matter whether AGI lands in 3 years or 30 — institutions, skills, judgment.
- // the only bad strategy is one that requires being right about the date.
- CONCLUSION // 11

### Slide 13: Further reading + watching.

- [12:: TRANSMISSION_END]
- Two starting points to keep going.
- Timeline debate
- Where the disagreement lives — proponents, skeptics, Bayesian forecasters.
- ▶ youtube // agi+timeline+debate
- Alignment problem
- The technical and conceptual core of why this is hard.
- ▶ youtube // ai+alignment+problem
- Adjacent reading
- Bostrom — Superintelligence
- Russell — Human Compatible
- Christian — The Alignment Problem
- Karnofsky — Most Important Century (blog series)
- METR, Apollo, Epoch — capability and forecasting research
- // END_TRANSMISSION ::: stay curious, stay skeptical.
- REFS // 12


## Related Decks

- [Asteroid Mining](https://shipslides.com/d/future-asteroid-mining)
- [Autonomous Vehicles](https://shipslides.com/d/future-autonomous-vehicles)
- [Brain-Computer Interfaces](https://shipslides.com/d/future-brain-computer-interfaces)
- [Climate Adaptation](https://shipslides.com/d/future-climate-adaptation)