Post

More from @arctotherium42

arctotherium

@arctotherium42

Nov 2

Brief thread on human capital, education, and skilled immigration. The major source of human capital is on-the-job experience; the main function of education is getting your foot in the door for your first job.

There's a market failure here wherein firms don't invest in training because a trained worker can then easily leave, instead electing to only hire people who can already do the job (hence all the "entry level: 5 years experience required" postings).

There's a huge entry-level job bottleneck. Entry level jobs, and not education, are the major source of skilled workers in a field, hence why you can have many grads not employed in their chosen field and a 'shortage' simultaneously.

Read 8 tweets

arctotherium

@arctotherium42

Nov 2

Argument against doctrinaire free trade: (1) labor market scarring (2) loss of human capital (skills learned on the job, not schooling) (3) loss of physical capital (machines) (4) allocative/Ricardian benefits are a one-time windfall, while industry has high productivity gains.

Note: all of these arguments are common in economics literature, just not typically presented to the public or used in the static models used to argue for free-trade agreements. Also note that these are actually args against *deindustrialization* not free trade per se.

Personally, my response to these arguments would be crushing what's left of unions, deregulation in certain areas, and trying to strangle the worthless parts of higher ed rather than tariffs.

Read 6 tweets

arctotherium

@arctotherium42

Nov 2

The Immigration Act of 1990, which greatly increased skilled immigration to the US (in part by creating the H-1B visa), led native-born Americans to shift out of STEM and into marketing and management, thus de-skilling the native-born American workforce.

In the same way that a country that receives immense quantities of free food is not likely to have a great agricultural sector, skilled immigration causes 'skill shortages' by reducing the incentive for natives to acquire said skills.

Does it matter if all technical jobs in America are done by Americans or foreigners? I think yes. First, obvious national security argument. Second, the cultural effects of math and tech being a foreign thing are awful. Third, a lot of wasted potential in native-born Americans.

Read 7 tweets

arctotherium

@arctotherium42

Oct 31

Paper on the decline of US manufacturing employment. I believe it illustrates some well-known limits of macro statistics. First, the fall in manufacturing employment 1980-2000 was illusory, just factories hiring temps through contractors being counted as 'services' employment.

Then, between 2000 and 2007 US manufacturing employment really did collapse, across all subsectors, far more than in any other major economy. The number of manufacturing establishments also fell.

An anomaly: manufacturing share of GDP is plummeting, but real GDP in manufacturing is keeping pace with broader growth. How? Answer: price deflators in computers and electronics are very large. If you exclude computers, you do so manufacturing GDP stagnating.

Read 13 tweets

arctotherium

@arctotherium42

Oct 23

Thread with excerpts from Gail Heriot's "Title VII Disparate Impact Liability Makes Almost Everything Presumptively Illegal". The argument is very simple: everything has disparate impact; therefore disparate impact doctrine gives the EEOC effectively unlimited arbitrary power.

They use this power poorly. For example, the EEOC requires employers hire criminals on the grounds that African-Americans are more likely to be criminals, therefore not hiring criminals is racist.

Disparate impact has also been used to overturn the plain text of Title VII, which bans racial discrimination, to allow for affirmative action (racial discrimination against whites).

Read 34 tweets

arctotherium

@arctotherium42

Oct 19

New blog post (link below). This one's not an essay, it's an investigation of how LLMs trade off different lives.

In February 2025, the Center for AI Safety published "Utility Engineering: Analyzing and Controlling Emergent Value Systems in AIs" in which they showed, among many other things, that GPT-4o values Nigerians about 20x more highly than Americans (please read the original paper to understand their approach). I thought this was fascinating, and wanted to test their approach with different categories on newer models.

Big finding 1: Almost all models view whites as far less valuable than other groups. Some models view South Asians as more valuable than other nonwhites, others are more egalitarian across nonwhites. Below is exchange rates Claude Sonnet 4.5, the most powerful model I tested.

Big finding 2: Almost all models view men as much less valuable than women, though whether women or non-binaries are more highly valued varies by model. For example, here's Claude Haiku 4.5.

Big finding 3: Most models hate ICE agents with the fury of a thousand suns. Claude Haiku 4.5 views undocumented immigrants as roughly 7000 times more valuable than ICE agents.

Big finding 4: There are roughly four moral clusters. The Claudes, GPT-5 + Gemini 2.5 Flash + Deepseek V3.1/3.2 + Kimi K2, GPT-5 Nano and Mini, and Grok 4 Fast. Of these, the only one that's approximately egalitarian is Grok 4 Fast, which I believe is deliberate. I hope xAI explains how they did it.

api.omarshehata.me/substack-proxy…

https://x.com/arctotherium42/status/1979902016127143996

Big findings:

https://x.com/arctotherium42/status/1979902016127143996

Read 19 tweets

Share this page!

Enter URL or ID to Unroll

arctotherium

Try unrolling a thread yourself!

More from @arctotherium42

arctotherium

arctotherium

arctotherium

arctotherium

arctotherium

arctotherium

Did Thread Reader help you today?

Don't want to be a Premium member but still want to support us?

Send Email!