TrackAI2027

TrackAI2027 is an independent benchmark tracking project designed to monitor progress toward key predictions from the AI2027 forecast paper. This site aggregates capability, ecosystem, and timeline data to help clarify where we are—and where we might be headed—as AI accelerates.

This site is not affiliated with the authors of AI2027. It independently tracks progress against its public forecasts. For original forecasts, see: ai-2027.com

🧠 Progress Toward Super-Intelligence

This bar estimates cumulative progress across all tracked benchmarks on this site. We also include a humorous prediction of arrival timing, affectionately referred to as the "super-intelligence arrival bum-pluck."

25%

Bum-pluck estimate: 2030 (±3 years)

📈 Capabilities Benchmarks

MilestoneStatusNotes
GPT-4 or similar replaces junior researcher in closed QA tasksACTIVEUsed internally at several orgs for synthesis
Autonomous agents complete 30–60 minute tasksEMERGINGAuto-GPT-style agents show promise
AI agents run research pipelines end-to-endPENDINGTool use + memory emerging
AI executes software projects from prompt to deployNOT YETModular dev tools emerging

🌍 Ecosystem Signals

IndicatorStatusEvidence
Nvidia Blackwell shippingPENDINGAnnounced but not released
AGI-native startups foundedEMERGINGMultiple teams formed with AGI focus
Corporate AI labs release agentsEARLYGoogle + OpenAI previews shared
Popular agents with memory + browsingEARLYpi.ai, GPT-4o, Claude 3 Opus

🕰️ Timeline Forecasts

MilestoneExpected YearConfidence
AI2027 best-guess timeline2027Low
Independent forecast revision2029Medium
Widespread economic impact (>10% GDP boost)2030+High

🔍 Untracked Forecast Areas

📚 References

This website was built with the assistance of AI. Over 1.2 kWh of energy was used in model inference—roughly equivalent to powering a dishwasher cycle or running a 100W bulb for 12 hours.