RAND
Article
RAND is a recurring organization in the Astral Codex Ten archive, appearing 6 times across 6 issues between July 13, 2022 and April 03, 2025. The archive places it in contexts such as “a RAND scientist later said”; “The INFER forecasting platform is now being led by RAND”; “In the 1970s, RAND gave thousands of people one of five types of insurance”. It most often appears alongside California, United States, CDC.
Metadata
- Category: Organizations
- Mention count: 6
- Issue count: 6
- First seen: July 13, 2022
- Last seen: April 03, 2025
Appears In
- Book Review: The Man From The Future
- 24
- Contra Hanson On Medical Effectiveness
- Highlights From The Comments On Hanson And Health Care
- The Case Against California Proposition 36
- Introducing AI 2027
Related Pages
-
- California (3 shared issues)
-
- United States (3 shared issues)
-
- CDC (2 shared issues)
-
- ChatGPT (2 shared issues)
-
- Goldin (2 shared issues)
-
- Lurie (2 shared issues)
-
- OpenAI (2 shared issues)
-
- Oregon (2 shared issues)
-
- Robin (2 shared issues)
-
- Ukraine (2 shared issues)
-
- “Most Drugs Are Bad For You” (1 shared issues)
-
- 1123581321 (1 shared issues)
External Links
Source Context
Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.
When the presentation was completed, he scribbled on a pad, stared so blankly that a RAND scientist later said he looked as if “his mind had slipped his face out of gear”, then said “Gentlemen, you do not need the computer. I have the answer.” While the scientists sat in stunned silence, Von Neumann reeled off the various steps which would provide the solution to the problem.
The dumbest possible way to do this is to ask GPT-4 to write a summary (“write the summary of a plot for a detective mystery story”), then ask it to convert the summary into a 100-point outline, then convert that into 100 minutes of a 100-minute movie, then ask Sora to generate each one-minute block. This wouldn’t work as written now (I don’t think Sora can do sound, it wouldn’t keep actors and style consistent unless you forced it), but it seems like something that requires incremental improvement rather than a grand breakthrough.
7: The INFER forecasting platform is now being led by RAND, suggesting that important people affiliated with the government are starting to take more interest.
Inline links: INFER, is now being led by RAND
We believe in medicine, and this faith has comforted us during the pandemic. But likewise the patients of the seventeenth century; they could probably also have named a relative cured by bloodletting. Yet health outcomes are typically too random for the experience of one family to justify medical confidence. How do we know our belief is justified?
But surely modern science must have some reliable way to study the aggregate value of medicine? Yes, we do. The key is to keep a study so simple, pre-announced, and well-examined that there isn’t much room for authors to “cheat” by data-dredging, p-hacking, etc. Large trials where we randomly induce some people to consume more medicine overall, and then track how their health differs from a control population–those are the key to reliable estimates. If trials are big and expensive enough, with lots of patients over many years, no one can possibly hide their results in a file drawer.
His argument: there have been three big experimental studies of what happens when people get free (or cut-price) health care: RAND, Oregon, and Karnataka. All three (according to him) find that people use more medicine, but don’t get any healthier. Therefore, medicine doesn’t work. If it looks like medicine works, it’s a combination of anecdotal reasoning, biased studies, and giving medicine credit for the positive effects of other good things (better nutrition, sanitation, etc).
3: Comparing OLS and IV results. I really didn't understand what point Hanson was trying to make here. In this context, OLS means comparing mortality among people who enroll in more months of health insurance to people who enroll in less. Differences in health insurance enrollment are non-random though, so we don't put much weight on the OLS estimate. Why would we be concerned that our 95% confidence intervals for the IV and OLS estimates don't overlap? Note also that the OLS standard errors are much smaller not because of a type-o in the table but because they are estimated from a different source of variation.
Here's one last way to understand the statistical significance of the results, which might be more intuitive. Suppose you were to take the individuals in our treatment and control groups and randomly re-shuffle them into (fake) treatment and control groups, and compare the difference in the mortality rates between the fake groups. You wouldn't expect to find an effect, but there might some differences just due to random noise. In Appendix Figure A.VII (below) we do this 1000 times, and compare the difference between the real treatment and control groups (our estimated effect from the study) to the distribution of the differences between these fake-groups. This tells us whether the difference between the treatment and control groups that we observe in the study (shown by the red line) is likely due to chance -- the figure below suggests that the answer is no, because it is more extreme than almost all of the fake comparisons.
What happens when patients suddenly stop their medications? We study the health consequences of drug interruptions caused by large, abrupt, and arbitrary changes in price. Medicare’s prescription drug benefit as-if-randomly assigns 65-year-olds a drug budget as a function of their birth month, beyond which out-of-pocket costs suddenly increase. Those facing smaller budgets consume fewer drugs and die more: mortality increases 0.0164 percentage points per month (13.9%) for each 100 per month budget decrease (24.4%). This estimate is robust to a range of falsification checks, and lies in the 97.8th percentile of 544 placebo estimates from similar populations that lack the same idiosyncratic budget policy.
Even the existing beds may not be accessible. A 2024 RAND study in five central California counties found that – in addition to severe shortages – "many facilities do not accept individuals with prior involvement in the criminal justice system, those with such physical health comorbidities as dementia, and those who either are enrolled in Medicaid or are uninsured." These policies are common in private treatment centers across the state. Of course, everyone forced into treatment by a felony conviction will have had contact with the criminal justice system, and the vast majority will be on Medicaid. In fact, Prop 36 requires that the court refer patients “to programs that provide services at no cost to the participant” – ruling out the most private treatment options.
Inline links: study
Eli Lifland, a superforecaster who is ranked first on RAND’s Forecasting initiative. You can read more about him and his forecasting team here. He cofounded and advises AI Digest and co-created TextAttack, an adversarial attack framework for language models.