# Chapter 05: MIRI: Research

> those who can't build, align

Canonical URL: https://foom.hyperplex.org/chapter/miri_research

Timeline: 2013–2021

After the early implementation timelines failed, MIRI turned toward decision theory, logical uncertainty, agent foundations, and alignment. The Institute increasingly treated the lack of any known safe path to advanced AI as its central problem.

Meanwhile, deep learning complicated the picture. Progress came through larger models, more data, better chips, and large research organizations, rather than through the small number of decisive architectural breakthroughs the foom model had anticipated. AlphaGo raised serious doubts about that picture; GPT-3 did far more damage.

## Predictions in this chapter

- [Doom by 2030, at 2:1 odds](https://foom.hyperplex.org/p/caplan/index.html.md): Resolves: 2030; predicted 2017.
- [No GDP boom before doom](https://foom.hyperplex.org/p/gdp/index.html.md): Falsified: 2025; predicted 2021.
- [Self-driving cars: approved only after the world ends](https://foom.hyperplex.org/p/selfdriving/index.html.md): Falsified: 2025; predicted 2021.
- [The trillion-scale AI boom: 'pretty darn prohibited'](https://foom.hyperplex.org/p/trillionboom/index.html.md): Falsified: 2025; predicted 2021.
- [Deepnets as ubiquitous as computers, maybe by 2035](https://foom.hyperplex.org/p/deepnets/index.html.md): Falsified: 2025; predicted 2021.
- [Codex: stagnation for 5 years or a massive leap in 2](https://foom.hyperplex.org/p/codexbet/index.html.md): Hedged: 2023; predicted 2021.
- [$1B/yr waifutech before AI freedom-of-the-road](https://foom.hyperplex.org/p/waifubet/index.html.md): Still open; predicted 2021.

## Chapter sources

- [MIRI Rebrand Announcement](https://intelligence.org/2013/01/30/singularity-institute-becomes-machine-intelligence-research-institute/)