A Taxonomy of R&D Orgs

18 hours ago

The essay proposes a taxonomy of R&D organizations with nine key parameters: project timelines, user input, discretionary funding, career incentives and talent models, revenue strategy, mission focus, organization size, governance and project selection, and openness and IP strategy.
Many historical R&D successes (e.g., Bell Labs, DARPA, HHMI) used novel organizational structures that combined specific parameter values to enable breakthroughs.
Current philanthropic efforts are creating new research organizations, but the full possibility space of organizational designs remains largely unexplored.
Each parameter interacts with others, and identifying gaps in the parameter space could lead to new effective structures for tackling unsolved scientific problems.
The authors encourage philanthropists and researchers to experiment with these parameters to address unmet research needs.

SETI@home was a project using home computer idle compute for distributed signal analysis in the Search for Extraterrestrial Intelligence.
Now, with AI subscriptions and local agent software, unused token allowances could be contributed to crowdsourced scientific or mathematical efforts.
Small teams and individuals combining domain expertise with AI systems have achieved research successes like solving Erdős problems.

Software design is best learned through hands-on experience, not just formal courses.
Conway's Law explains that software architecture mirrors the social structure of the producing organization.
Incentive structures (e.g., PhD deadlines) heavily influence the quality and nature of scientific software.

Researchers demonstrated that quantum computers can enhance the accuracy and scope of generative AI models for drug discovery.
A hybrid quantum-classical system accelerated AI to generate novel peptides, improving binding prediction to target proteins, especially with scarce training data.
The project was self-funded by the team using spare time and leftover money, as innovative science is often deemed risky by traditional funders.

CrucibleBench evaluates language models in a persistent MUD text world over 50 turns with hidden social objectives.
It uses 'lateral thinking with withered technology'—applying mature, inexpensive tech (MUDs) to measure AI behavior.
Key constraints include an enumerable action space, explicit social feedback from NPCs with trust/suspicion states, and within-run persistence.

The author believes a subtle 'GPT-4 moment' has occurred that current benchmarks fail to capture, with major implications for industries and the economy.
Chat-based evaluation of LLMs is insufficient; speed of response often matters more than perceived 'quality' for everyday use.
Gemini 3 Pro excels at design, generating visually appealing prototypes that other models cannot replicate, effectively acting as a midlevel designer.

SETI@home was a project using home computer idle compute for distributed signal analysis in the Search for Extraterrestrial Intelligence.
Now, with AI subscriptions and local agent software, unused token allowances could be contributed to crowdsourced scientific or mathematical efforts.
Small teams and individuals combining domain expertise with AI systems have achieved research successes like solving Erdős problems.

Software design is best learned through hands-on experience, not just formal courses.
Conway's Law explains that software architecture mirrors the social structure of the producing organization.
Incentive structures (e.g., PhD deadlines) heavily influence the quality and nature of scientific software.

Researchers demonstrated that quantum computers can enhance the accuracy and scope of generative AI models for drug discovery.
A hybrid quantum-classical system accelerated AI to generate novel peptides, improving binding prediction to target proteins, especially with scarce training data.
The project was self-funded by the team using spare time and leftover money, as innovative science is often deemed risky by traditional funders.

CrucibleBench evaluates language models in a persistent MUD text world over 50 turns with hidden social objectives.
It uses 'lateral thinking with withered technology'—applying mature, inexpensive tech (MUDs) to measure AI behavior.
Key constraints include an enumerable action space, explicit social feedback from NPCs with trust/suspicion states, and within-run persistence.

The author believes a subtle 'GPT-4 moment' has occurred that current benchmarks fail to capture, with major implications for industries and the economy.
Chat-based evaluation of LLMs is insufficient; speed of response often matters more than perceived 'quality' for everyday use.
Gemini 3 Pro excels at design, generating visually appealing prototypes that other models cannot replicate, effectively acting as a midlevel designer.

Hasty Briefsbeta