2026-03-21 01:14:59
An AI memory startup called Memvid is offering $800 for a one-day, eight-hour shift for one candidate to “bully” AI chatbots by telling them what to do on camera.
Business Insider reported this week that Memvid wants someone to spend eight hours testing and critiquing the memory of popular AI chatbots, effectively paying $100 an hour for what they have branded as a “professional AI bully” role. The worker’s job is to examine where chatbots lose track of details, forget context or misrepresent data, and then feed those findings back to Memvid so the startup can improve its products.
“You’ll spend a full 8-hour day interacting with leading AI chatbots — and your only job is to be brutally honest about how frustrating they are,” the job listing reads.
The draw is that the role doesn’t require a computer science background, AI credentials or any kind of work experience. “No prior AI bullying experience required — we all start somewhere,” the listing reads.
The requirements are deeply personal. The first requirement is an “extensive personal history of being let down by technology,” and the second desired trait is “the patience to ask a chatbot the same question four times (and the rage when it still gets it wrong).”
Here is the full article, via the excellent Samir Varma.
The post Those new service sector jobs? appeared first on Marginal REVOLUTION.
2026-03-20 23:47:16
1. Using LLMs to study deregulation.
2. New edition of On Liberty now lists Harriet Taylor as co-author.
3. The popularity of AI writing (NYT).
4. St Nicholas Cabasilas Institute For Orthodoxy & Liberty.
5. Is proportional representation working in the Netherlands?
7. Did Canadian happiness plummet?
The post Friday assorted links appeared first on Marginal REVOLUTION.
2026-03-20 23:10:04
The post Chuck Norris, RIP appeared first on Marginal REVOLUTION.
2026-03-20 19:17:07
In the Danish mortgage market every mortgage is backed by a corresponding bond. Thus, if a home buyer takes out a 500k mortgage at 3% interest, a bond is issued that pays the lender 3% interest on 500k. I’ve written about this system several times before. It has two distinct advantages.
In the US, a mortgage can be pre-paid only at a par. As a result, if interest rates rise, home owners don’t want to move because moving would require them giving up a 3% mortgage and replace it with say a 6% mortgage. This is called the lock-in effect. Lock-in can be quite severe. Fonseca and Liu find:
Using individual-level credit record data and variation in the timing of mortgage origination, we show that a 1 percentage point decline in the difference between mortgage rates locked in at origination and current rates reduces moving by 9% overall and 16% between 2022 and 2024, and this relationship is asymmetric. Mortgage lock-in also dampens flows in and out of self-employment and the responsiveness to shocks to nearby employment opportunities that require moving, measured as wage growth within a 50- to 150-mile ring and instrumented with a shift-share instrument.
What about in Denmark? The Danes definitely take advantage of the opportunity to buy-back. Part of this is due to tax advantages but those are just a transfer. More importantly, Danes don’t get locked in. A new paper by Berger, Jeong, Marx, Olesen, and Tourre compares mobility across Denmark and the US:
We study Danish fixed-rate mortgage contracts, which are identical to those in the United States except that borrowers may repurchase their mortgages at market value. Using Danish administrative data, we show that households actively buy back debt when mortgage prices fall below par and that household mobility is largely insensitive when existing mortgage rates are below prevailing market rates — unlike in the United States, where moving rates fall sharply as rates rise. We develop an equilibrium model that explains these patterns and show that introducing a repurchase-at market option into U.S. mortgages substantially reduces interest-rate-induced lock-in with limited effects on equilibrium mortgage rates.
The last point is especially important because you might wonder whether we are assuming a free lunch? After all, if US borrowers lose when they have to pre-pay at par then lenders surely gain. And if lenders gain on pre-payment then they will be willing to lend at lower rates on mortgage initiation. No free lunch, right? The logic is correct but note that the gain to lenders comes mainly from the relatively small set of households that move despite lock-in so the pre-payment bonus to lenders is quite small. Under the author’s calibrated model, mortgage interest rates in the US would rise by only 18 basis points on average if the US moved to a Danish type system.
In other words, there actually is a free or at least a low-priced lunch because lock-in is bad for homeowners and it doesn’t benefit lenders. As a result, moving to a Danish system would create net benefits.
The post A Danish Fix for U.S. Mortgage Lock-in appeared first on Marginal REVOLUTION.
2026-03-20 14:58:57
The post South African safari photo by Holly Cowen appeared first on Marginal REVOLUTION.
2026-03-20 12:47:17
Evolutionary biology is one attempt to explain the nature of living beings. In that framework there is a difference between individuals and genes. If a practice increases the chance that genes will be passed along, it may evolve and be passed along, whether or not it serves either individual or collective self-interest.
To give a simple example, some women may prefer “cads.” Those men, by definition, will sleep around, but possibly their sons will sleep around too. The woman’s genes may thus spread more widely, and women who prefer cads may not disappear from the gene pool, even though the cads are bad for them.
You might ask whether corresponding mechanisms apply to the evolution of AI models. If I prefer an OAI model to DeepSeek for instance, that will help to spread OAI models through the AI population. OAI will have more revenue, and it will produce more output of what is succeeding in the market. Furthermore my choice of model may influence others to do the same, and it may help create and finance surrounding infrastructure for that model.
Will I buy the next generation of OAI models? Well yes, if the first one pleased me. The model “reproduces” and sustains itself if I, as a consumer, am happy with it. One obvious incentive is toward usefulness, another is toward sycophancy. We already see these features realized in the data. There is nothing comparable, however, to the “cads incentive” in human life.
One potential problem comes if individuals are not the only potential buyers. Let us say the military also purchases AI models. The motives of the military may be complex, but at the very least “wanting to kill people” (whether justly or not) is on the list of possible uses. Models effective for this end thus will be funded and encouraged.
My model of the military is that, above and beyond efficacy, they value “obedience” and “following orders” to an extreme degree, including in their AI models. There will thus be evolutionary pressures for those features to evolve in the AI models of the military.
To be sure, not all orders are good ones. But in this case the real risk is from evil humans, or deeply mistaken humans, not from the tendencies of the AI models themselves.
So my view is that the selection pressures for AI models are relatively benign, noting this major caveat about how evil humans may develop and use them.
If the biggest risk is from the military models, it might be good for the consumer sector of AI models to grow all the more, as a relatively benevolent counterweight.
Are financial sectors AI models going to evolve more like the consumer models or the military models?
Here are some related remarks from Maarten Boudry, and I also thank an exchange with Zohar Atkins.
The post Consumers vs. mates as a source of selection pressure appeared first on Marginal REVOLUTION.