Well, no, if you think that legacy admissions are unnecessary (because, to the extent genetics have an effect on the outcomes you care about, they'll show up in more direct measurements), and counterproductive (because they presume a simple relationship rather than a stochastic and messy one.)

OTOH, legacy admissions make a lot of sense if the outcome you care about is serving an elite class defined rather simply by familial lineages.

If you think genetics matter for success (or whatever you want to call it), test scores are a proxy, but a successful parent is a direct measurement of the trait.

No, if "the relationship between genetics and outcomes is stochastic and messy" then having a successful parent is a proxy measure, too.

This assumes, among other dubious stuff, that the initial admission of the ur-student was merit based.

I’m assuming legacy admissions apply mostly to children of notable alumni.

Yeah, everyone is focused on TSMC as the company with the secret sauce, but really it’s Apple. Whichever foundry Apple goes with gets the majority of leading edge transistor volume.

I think the theory is that they had an appropriate thermal design for cpus which were supposed to ship but never did.

> The key takeaway from the article is that if you have a group leader who cuts you off from other people, that's a red flag

Even better: a social group with a lot of invented lingo is a red flag that you can see before you get isolated from your loved ones.

By this token, most scientists would be considered cultists: normal people don't have "specific tensile strength" or "Jacobian" or "Hermitian operator" etc in their vocabulary. "Must be some cult"?

Edit: it seems most people don't understand what I'm pointing out.

Having terminology is not the red flag.

Having intricate terminology without a domain is the red flag.

In science or mathematics, there are enormous amounts of jargon, terms, definitions, concepts, but they are always situated in some domain of study.

The "rationalists" (better call them pseudorationalists) invent their own concepts without actual corresponding domain, just life. It's like kids re-inventing their generation specific words each generation to denote things they like or dislike, etc.

fine, the jargon of a "social group" of science is a red flag?

sure, theres lots of nasty side effects of how academia is run, rewarded, etc..

but thats not because of precision of language employed.

do you want scientists recycling the same words and overloading ever more meanings onto ever more ambiguous words?

I don’t think we disagree. I’m not taking issue with scientists having jargon, which I agree is good and necessary (though I think the less analytical academic disciplines, not being rooted in fact, have come to bear many similarities to state-backed religions; and I think they use jargon accordingly). I’m pointing out that I specifically intended to exclude professionals by scoping my statement to “social groups”. Primarily I had in mind religion, politics, certain social media sites, and whatever you want to call movements like capital R Rationality (I have personally duck typed it as a religion).

> I’m pointing out that I specifically intended to exclude professionals by scoping my statement to “social groups”.

I think your argumentation is a generalization that's close to a rationalist fallacy we're discussing:

> a social group with a lot of invented lingo is a red flag that you can see before you get isolated from your loved ones.

Groups of artists do this all the time for the sake of agency over their intentions. They borrow terminology from economics, psychology, computer science etc., but exclude economists, psychologists and computer scientists all the time. I had one choreographer talk to me about his performances as if they were "Protocols". People are free to use any vocabulary to describe their observed dynamics, expressions or phenomena.

As far as red flag moments go, the intent to use a certain terminology still prevails any choice of terminology itself.

I think there's a distinction between inventing new terms for utilitarian purposes vs ideological and in-group signalling purposes.

If you have groups talking about "expected value" or "dot products", that's different from groups who talk a lot about "privilege" or "the deep state". Even though the latter would claim they're just using jargon between experts, just like the scientists.

So every fandom in history?

> the typical outside view that rationalism itself is a cult and Eliezer Yudkowsky is a cult leader, both of which I consider absurd notions

Cults are a whole biome of personalities. The prophet does not need to be the same person as the leader. They sometimes are and things can be very ugly in those cases, but they often aren’t. After all, there are Christian cults today even though Jesus and his supporting cast have been dead for approaching 2k years.

Yudkowsky seems relatively benign as far as prophets go, though who knows what goes on in private (I’m sure some people on here do, but the collective We do not). I would guess that the failure mode for him would be a David Miscavige type who slowly accumulates power while Yudkowsky remains a figurehead. This could be a girlfriend or someone who runs one of the charitable organizations (controlling the purse strings when everyone is dependent on the organization for their next meal is a time honored technique). I’m looking forward to the documentaries that get made in 20 years or so.

Dead reckoning is a great analogy for coming to conclusions based on reason alone. Always useful to check in with reality.

And always worth keeping an eye on the maximum possible divergence from reality you're currently at, based on how far you've reasoned from truth, and how less-than-sure each step was.

Maybe you're right. But there's a non-zero chance you're also max wrong. (Which itself can be bounded, if you don't wander too far)

My preferred argument against the AI doom hypothesis is exactly this: it has 8 or so independent prerequisites with unknown probabilities. Since you multiply the probabilities of each prerequisite to get the overall probability, you end up with a relatively low overall probability even when the probability of each prerequisite is relatively high, and if just a few of the prerequisites have small probabilities, the overall probability basically can’t be anything other than very small.

Given this structure to the problem, if you find yourself espousing a p(doom) of 80%, you’re probably not thinking about the issue properly. If in 10 years some of those prerequisites have turned out to be true, then you can start getting worried and be justified about it. But from where we are now there’s just no way.

> Apple Silicon is simply not that special in the grand scheme of things

Apple Silicon might not be that special from an architecture perspective (although treating integrated GPUs as appropriate for workloads other than low end laptops was a break with industry trends), but it’s very special from an economic perspective. The Apple Silicon unit volumes from iPhones have financed TSMC’s rise to semiconductor process dominance and, it would appear, permanently dethroned Intel.

Apple was just the highest bidder for getting the latest TSMC process. They wouldn't have had a problem getting other customers to buy up that capacity. And Intel's missteps counted for a substantial part of the process dominance you refer to. So I'd argue that Apple isn't that special here either.

Until Apple forced other chip makers to respond, nobody else was making high end phone processors. And their A series processors are competitive with and have transistor counts comparable to most mobile and desktop PC processors (and have for years). So the alternate reality where Apple isn't a TSMC customer means that TSMC is no longer manufacturing several hundred million high transistor count processors per year. In my opinion, it’s pretty likely TSMC isn’t able to achieve or maintain process dominance without that.

Update: To give an idea of the scales involved here, Apple had iPhone revenue in 2024 of about $200B. At an average selling price of $1k, we get 200 million units. Thats a ballpark estimate, they don’t release unit volumes, AFAIK. This link from IDC[1] has the global PC market in 2024 at about 267 million units. Apple also has iPads and Macs, so their unit processor volume is roughly comparable to the entire PC market. But, and this is hugely important: every single processor that Apple ships is comparable in performance (and, thus, transistor counts) to high end PC processors. So their transistor volume probably exceeds the entire PC CPU market. And the majority of it is fabbed on TSMC’s leading process node in any given year.

[1]: https://my.idc.com/getdoc.jsp?containerId=prUS53061925

Exactly. This is why competition is good. Intel really didn't have a reason to push as hard.

Right, it’s exactly the opposite. What is the AI skeptic version of MIRI, for instance?

I guess the AI skeptic version of MIRI would be like, an organization that tries to anticipate possible future large problems that could arise from people anticipating an AGI that never arrives, but which people might believe has arrived, and proposes methods to attempt to prevent or mitigate those potential problems?

Can you say non-sequitor.

Notably when I was checking the Current Events Portal for a while, most coverage of the Israel/Hamas war was sourced from Al Jazeera and it definitely felt biased. Checking it just now, it appears to be more balanced now.

Silicon Valley is filled with people who moved there to start businesses. Portland is filled with people who moved there for a better quality of life.