research taste
In plain wordsThe knack good researchers have for choosing worthwhile problems, designing the right experiments and drawing the right conclusions — a skill seen as key to whether AI can do research on its own.
The judgment that lets a researcher pick promising problems, design informative experiments and interpret results correctly, as distinct from the skill of implementing them. Forecasts of automated AI researchautomated AI research, n.: Using AI systems to do the work of AI researchers, such as running experiments and writing code for new models, so that AI helps build its own successors. treat it as a key input: scenario models such as AI 2027AI 2027, prop. n.: A detailed 2025 forecast, written as a month-by-month story, imagining AI becoming superhuman by the end of 2027, with two alternative endings depending on whether humanity slows down. scale the speed of AI-driven research by the quality of its taste, so that taste rather than coding ability may set the pace of recursive self-improvementrecursive self-improvement, n.: An AI improving its own design or training, so that each improved version is better at making the next improvement. It is the engine usually imagined behind an intelligence explosion.. Because it shows up only in the long-run quality of research choices, it is much harder to measure than coding skill.
The agents can run any experiment we want; what they still lack is research taste about which ones matter.
Origin
Long-standing jargon in research culture, of uncertain first use; it became an explicit variable in forecasts of automated AI research such as AI 2027 (April 2025).
Related
- automated AI researchautomated AI research, n.: Using AI systems to do the work of AI researchers, such as running experiments and writing code for new models, so that AI helps build its own successors.
- recursive self-improvementrecursive self-improvement, n.: An AI improving its own design or training, so that each improved version is better at making the next improvement. It is the engine usually imagined behind an intelligence explosion.
- AI scientistAI scientist, n.: An AI system meant to do research largely on its own: reading the literature, coming up with ideas, running experiments and writing up the results.
- superhuman codersuperhuman coder, n.: A milestone used by AI forecasters: an AI that can do the coding work of a company's best programmer, far faster and in many copies at once.
- AI 2027AI 2027, prop. n.: A detailed 2025 forecast, written as a month-by-month story, imagining AI becoming superhuman by the end of 2027, with two alternative endings depending on whether humanity slows down.
- intelligence explosionintelligence explosion, n.: The idea that once AI gets good enough at improving AI, each smarter version could build a smarter one faster still, so abilities would snowball far beyond human level in a short time.
Sources
- Kokotajlo, D., Alexander, S., Larsen, T., Lifland, E. & Dean, R. (2025). AI 2027. AI Futures Project, April 2025.