somedudetbh

In my experience, one of the most reliable heuristics for finding a place that makes good sandwiches is "go to a place that's a good bakery and see if they make sandwiches".

I can't think of a time I've gotten a sandwich from a (good) bakery where the sandwich wasn't at least quite good, and frequently, very very good. On the flip side, if you just buy a sandwich from a dedicated sandwich shop? On average it will be bad. There are excellent sandwich shops, for sure, that do not bake their own bread. But there are very few bakeries that make sandwiches that do not make extremely good sandwiches. (Subway doesn't count: they are not a bakery, in that they do not sell bread or other baked goods. They only produce their disgusting "bread" to enable them to sell sandwiches).

It also strikes me that this argument is essentially the inverse of the Alan Kay line "People who are really serious about software should make their own hardware" that Apple people are always quoting.

I think perhaps the Sandwich Fallacy lacks explanatory power, because the Stack Fallacy does as well. I think if the reason why big companies consistently fail to win markets in which their customers compete was because of the points made in the post, then we would see evidence that big companies are disproprotionately successful at winning markets in which their suppliers compete, the layer _below_ them in the stack: "The bottleneck for success often is not knowledge of the tools, but lack of understanding of the customer needs." the companies that build these sandwich-filling layers are the customer, they understand this quite well, but I don't think they generally succeed at this. So there must be something else at play.

I also find the examples in the article unconvincing:

"Apple continues to successfully integrate vertically down — building chips, programming languages, etc., but again has found it very hard to go up the stack and build those simple apps — things like photo sharing apps and maps."

Apple's photo app is extremely popular. Apple's messaging app, Messages, is so compelling it continues to sell Apple's ludicrously expensive devices. It's literally a Killer App for iOS, in the Visicalc mode. Apple has been building top-tier first party applications for it's platform since the 1980s. For iOS, it's Photos, Messages, Notes, Music, and Safari (I'm not arguing that Safari isn't terrible, or that Apple isn't holding back the progress of the entire open web via failing to make progress on Mobile Safari (they are). I'm simply arguing that it's undeniably successful.) Before the mobile era it was the 'digital hub' apps like iPhoto, iMovie, Garage Band. In the 'productivity' era it was ClarisWorks. In fact, it's so common that there's a slang term for when Apple-the-platform-vendor starts to compete with it's application developers and uses its structural advantages to win the market: "Sherlocking".

"It is therefore no surprise that Apple had an easier time building semiconductor chips than building Apple Maps."

Did they? They bought PA Semi a zillion years ago. Apple Maps had a rocky launch but now it's quite good. I concede I have no evidence that it's popular or successful in the market. It looks to me like Apple was successful in both categories.

"In the 1990s, Larry Ellison saw SAP make gargantuan sums of money selling process automation software (ERP) — to him, ERP was nothing more than a bunch of tables and workflows — so he spent hundreds of millions of dollars trying to own that market, with mixed results. Eventually, Oracle bought its way into the apps market by acquiring PeopleSoft and Siebel."

I mean, sort of? Oracle is an absolutely dominant player in this market category now. They got their through the usual mix of Oracle chicanery. You know where Oracle is struggling? All the layers _below_ them.

So I think it's pretty safe to reject the Stack Fallacy and the Sandwich Fallacy. There's clearly a pattern where big companies fail to win markets of their customers as well as markets dominated by their suppliers, which is confusing given the strategic advantages they would have expanding in either direction, but I would argue that if there are common structural explanations for this, the proposed explanations are not correct.

I guess I just think it's funny that when I skimmed the initial post I just thought "hmmm, maybe?" but when I read your sandwich analogy I was like "oh, right..this doesn't make any sense. Bakeries make awesome sandwiches, almost always!" and I started thinking about it more. Whereas if you made the same point with almost any other example I would have probably been like, "yeah! This guy's right! None of the best ice cream shops are also dairies! None of the best coffee shops are also coffee farms! I've never seen a successful textile weaver start a line of pants! None of the best...tire stores...also...produce industrial rubber compounds?" I don't know. So it's a funny choice.

Similarly, a lot of dam preservation pressure is reflexive reactionary thinking that if someone wants to remove the dams, it must be because they're a hippie environmentalist and the dam must be saved to show our commitment to Progress.

Dams have a finite lifespan. Rivers carry sediment. The dams slow the flow of water and the sediment is dropped. This fills up the reservoir behind the dam, eventually making the dam ineffective. In addition, ordinary mechanical stresses wear out dams and they're components, so there is a maintenance cost to just keeping them running.

Many failure modes for dams are catastrophic: a release of water and silt all at once into downstream areas.

Worse, many of the dams that were built in the dam-building boom in the US West from circa 1930 to 1965 or so were not particularly well-thought-out, especially smaller privately planned dams.

In the mid-century American Bureau of Reclamation, building dams was like building new chat services is at Google today. While dams, as a concept, are completely critical to making the western united states survivable with mid-20th century technology, many of the actual dams were not good designs, they are the result of a generation or two of engineers responding to promotion incentives within a large bureaucracy, and they should no more be given the benefit of the doubt as good engineering projects than the last abandoned open source project you saw from Google, Facebook, Uber, etc.

Consider Matilija Dam (https://en.wikipedia.org/wiki/Matilija_Dam):

* "In 1941 the U.S. Army Corps of Engineers warned that the dam would not be economically effective, as the steep, erosive topography upstream would cause it to silt up quickly. However, the project moved forward and in 1945 the county issued $682,000 in revenue bonds to fund it. Construction began on 18 June 1946 and was completed on 14 March 1948 at a cost of nearly $4 million, six times the original estimate" * "Almost immediately after construction, the dam began silting up.[7] The dam traps about 30% of the total sediment in the Ventura River system, depriving ocean beaches of replenishing sediment.[6] Initially, engineers had estimated it would take 39 years for the reservoir to fill with silt,[1] but within a few years it was clear that the siltation rate was much faster than anticipated. In 1964 a safety study was commissioned from Bechtel Corporation, which determined the dam was unsafe and recommended removal." * The dam was notched twice, reducing its capacity and function, and the reservoir was useless by 2020. * Ventura county started trying to remove the dam in 1998 (who knows what happened between 1964 and then), but the dam is still there.

Even the good dams don't last forever, and there is no plan to deal with the sediment build up in the West's dams.

However, the Reclamation and Army Corps of Engineers dams are the good ones. The real corkers are the private dams. Consider Rindge Dam (https://en.wikipedia.org/wiki/Rindge_Dam) in Malibu, California, which was privately built by the Rindge family when they controlled the entire Mexican land grant rancho that is present-day Malibu:

* Built 1926 * Completely silted up by 1950, 24 years later. 24 years of "useful" life tops ("useful" is suspect because most Rindge family building projects were weird compliance dodges to preserve control of the ranch. They spent decades building and tearing down a railroad because the law on the books at the time prevented the state from using eminent domain to seize their land for road-building if there was a railroad under construction there) * Congress authorized removal study in 1992. * In 2014, dam considered so dangerous due to lack of repairs that the area, which is now in a state park, was closed to the public in 2014. * The dam can't just be knocked down, what would happen to the 600k cubic meters of sediment that are now trapped behind the dam, that should have flowed down the river for the last 100 years? The plan is to _truck the sediment out_. Some will be dumped in the ocean, the rest in _landfills_. * The currently scheduled goal to complete the removal project is _2033_. The dam was been functionally useless for its original purpose since 1950. It's 83 year "useless/dangerous" lifespan will surpass it's 24 year "useful" lifespan by 3.5x! Surely _some_ of that is government beauracracy but not all: it's very difficult to unbuild a silted up dam. It's harder to undo things than it is to do them.

I think there is much more significant "religious faith" in the sanctity of dams than there is "belief that dams are evil because they aren't natural" in the United States. Dams are a powerful symbol of America's mid-century confidence in it's ability to bend nature to its will. Hoover Dam is more than a tourist site, it's something closer to a civic-religious site, like the Lincoln Monument. So is Glen Canyon. Grand Coulee Dam is known to a lot of people as "The Dam That Won World War II" for it's role in powering the aluminum-smelting plants and nuclear material refinement sites in the Northwest. How many pieces of infrastructure are considered war heros in the US?

The sanctity of dams is way more obvious in the northeast. There's hundreds and hundreds of abandoned dams on every trickle of water in the mid-Atlantic and New England, all to power mills that stopped milling 100 years ago, but the dams are still there, and the fish are not.

It's funny to read this kind of bullshit but imagine it being a description of like, Julia Louis-Dreyfus.

Alternative theory: everyone who worked on this knew that it was not a filesystem and "object store" is a description intended to describe everything else pointed out in this post.

"Objects were really popular" is about objects as software component that combines executable code with local state. None of the original S3 examples were about "hey you can serialize live objects to this store and then deserialize them into another live process!" It was all like "hey you know how you have all those static assets for your website..." "Objects" was used in this sense in databases at the time in the phrase "binary large object" or "blob". S3 was like "hey, stuff that doesn't fit in your database, you know...objects...this is a store for them."

This is meant to describe precisely things like "listing is slow" because when S3 was designed, the launch usecases assumed an index of contents existed _somewhere else_, because, yeah, it's not a filesystem. it's an object store.

I have basically no physics education, but I was educated in the US. In my high school, where I took an ordinary, low-quality physics course (the one for weak students that didn't require calculus), the introduction of quantum mechanics was motivated by the photoelectric effect. Now, like I said, I don't really have any physics education and I don't really understand the photoelectric effect _or_ quantum mechanics, but my basic recollection was waves hands you shine a light on certain materials and electrons pop out, and it looks for various reasons like this behavior is packet-y rather than smooth as one might expect.

Basically in an intro American physics class, if they've got an opportunity to get Einstein involved, they're gonna take it.

I've never heard of this myth either.

FWIW, the authors, Nils-Erik Bomark and Reidun Renstrøm, also appear to work at .no universities.

on a perfectly clear morning in a residential neighborhood, with curbs painted with that yellow reflective paint.

Here's how the adoption of this technology is going to go (this is the way all AI technology adoption has gone for 60 years):

1) Papers will come out showing how by creating a more effective way to leverage compute + data to make a system self-improving, performance at some task looks way better than previous AI systems, almost human-like. (This already happened: "All You Need Is Attention")

2) The first generally available implementations of the technology, in a pretty raw form, will be released. People will be completely amazed at how this machine can do something that was thought to be a hallmark of humans! And by just doing $SIMPLE_THING (search, token prediction) which isn't "really" "thinking"! (This will amaze some people but also form the basis of a lot of negative commentary) (Also already happened: ChaGPT, etc)

3) There will be a huge influx of speculative investment capital into the space and a bunch of startups will appear to take advantage of this. At the same time, big old tech companies will start putting stickers on their existing products that say they're powered by LLMs. (Also already happened)

4) There will be a wave of press, first in the academia, then in technology circles, then in the mainstream, about What This Means. "AGI" is just over the horizon, all human jobs are about to be gone, society totally transformed. (We are currently here at step 4)

5) After a while, the limits of the technology will start to become clear. A lot of the startups will figure out that they don't really have a business, but a few will be massively successful and either build real ongoing businesses that use LLMs to solve problems for people, or get acquired. It will turn out that LLMs are massively, massively useful for some previously-thought-to-be-nearly-impossible or at least contigent on solving the general AI problem work: something like intent extraction, grammarly-type writing assistants, Intellisense on steroids, building natural chat interfaces to APIs in products like Siri or Alexa that understand "turn on the light" and "turn on the lights" mean the same thing. I have no idea what the things will actually be, if I was good at that sort of thing I'd be rich.

6) There will be a bunch of "LLMs are useless!" press. Because LLMs don't have Rosie-from-the-Jetsons level of human-like intelligence, they will be considered "a failure" for the general AI problem. Once people get accustomed to whatever the actual completely amazing things LLMs get used for, things that seemed "impossible" in 2021. Startups will fail. Enrollments in AI courses in school will drop, VCs will pull back from the category, AI in general (not just LLMs) will be considered a doomed investment category for a few years. This entire time, LLMs will be used every day by huge numbers of people to do super helpful things. But it will turn out that no one wants to see a movie where the screenplay is written by AI. The LLM won't be able to drive a car. All the media websites that are spending money to have LLMs write articles will find out that LLM-generated content is a completely terrible way to get people to come to your site, read some stuff and look at ads, with terrible economics, and these people will lose at least hundreds of millions of dollars, probably low billions, collectively.

7) At this trough point where LLMs have "failed" and AI as a sector is toxic to VCs, what LLMs do will somehow be thought of as 'not AI'. "It's just predicting the next token" or something will become the accepted common thinking that disqualifies it as 'Artificial Intelligence'. LLMs and LLM engineering will be considered useful and necessary, but it will be considered a part of mainstream software engineering and not really 'AI' per se. People will generally forget that whatever workaday things LLMs turn into a trivial service call or library function, used to be massively difficult problems that people thought would require human-like general intelligence to solve (for instance, making an Alexa-like voice assistant that, can tell 'hey can you kill the lights', 'yo shutoff the overhead light please?', 'alright shut the lights', 'close the light' all mean the same thing). This will happen really fast. https://xkcd.com/1425/

Sometimes when you see an amazing magic show, if you later learn how the trick was done, it seems a lot less 'magical'. Most magic tricks exploit weird human perceptual phenomena and, most of all, the magicians willingness to master incredibly tedious technique and do incredibly tedious work. Even though we 'know' this at some level when we see magicians perform, it's still deflating to learn the details. For some reason, AI technology is subject to the same phenomenon.

I think you should stop giving out this daily reminder. This meme has outlived its usefulness.

Using face id to unlock a local key store to enable my device to sign a signed challenge from a site I want to log into with the private key stored on my device is not a 'username' in any meaningful sense.

The problem is, the metaphor about passwords and usernames is not a good, structure-preserving simplification of the actual problem of authentication.

The biometric data and/or pin code are not being used to prove that you are you to Gmail, it's being used to unlock the set of private keys you have on your device. This doesn't fit into the metaphor at all.

If my non-technical parents said they were migrating all their accounts to passkeys, I would be very pleased. I wouldn't be worried about their inability to change their biometrics and that causing a problem following some sort of breach in the future. I am highly worried about their extreme susceptibility to phishing, especially in their inability to distinguish phishing sites from real sites, or real account maintence contacts via email and SMS from phishing contacts, their reuse of very simple passwords that are probably circulating in combolists already, and their general inability to retain username/password pairs. I have a lot of sympathy for them when I try to talk them through something like logging in to an Apple device with their apple id, when their appleid username is their email, which ends with @gmail.com. "But...why would i log in to apple with my gmail?" nevermind how confused they are about 'log in with google', 'log in with facebook', etc.

Moving to a model where their devices store webauthn credentials and guard them with a pin or faceid-style biometric shortcut is a _massive_ improvement in practical resistance to account takeover for my parents, and I don't think continuing to say 'biometrics are usernames in authn' is accurate or helpful.

Just making sure I understand your point here, are you just against "banking"? Unhedged duration risk w/ LOLR backup is essentially "the banking business".

What's the moral hazard? The equity is wiped out, most of the debtors are wiped out.

The only moral hazard I see is we've disincentivized individual depositors from assessing the financial strength of their banks. To me, this is a good thing. I hope we can bring similar moral hazards to choosing a plane to fly on, bridge to cross, building to enter, restaurant to eat in, hospital to go to, etc.

I can't really think of any Unix programs in common use that do one thing well. Inevitably, they grow to redundantly include features that other programs have.

If anything can be said to be the actual lived Unix philosophy it's probably something like: 1) Try to make your programs communicate using text, where the 'record' is separated by a newline. 2) If you have a choice between making a program a little more friendly to use, or making it's text communication a little simpler to compose with, go with the second option. E.g. if there's nothing to output when a program runs don't output "Nothing to output", just output nothing.

I don't think Emacs really fits with the unix philosophy, whether my version or the common version, at all. One of the main things I like about emacs is that it's like a vision of an alternate computing universe, where we assume that we're operating in a giant image that's meant to be programmed and customized all the time. What if most of your work was done in a LISP repl with a bunch of UI tweaks? It's like using a research project from a 70s or 80s "Future of Computing" lab that never actually shipped, except it did ship. It feels like it should have a lightpen attached.

I don't even really think it's particularly practical to use emacs at this point, I just like this vibe and reminder that nothing about the way we use computers was inevitable so much that I stubbornly still use it.

There are some extra keys, but some of them seem hard to use well (the big red triangle keys at the top of the thumb clusters) and other things that sound great in practice, like "hold" = modifier but "tap" = insert some character, in practice have turned out to have more annoying rough edges than I had hoped.

It's very well-made though, the configurator works good, etc. Seems like a high-quality gadget, but I wish I thought a bit more about what I would do for arrow keys, special characters, modifiers, etc.

I found the Kinesis thumb clusters a lot more usable when I used to have one, but I really wanted to try a split keyboard this time.

> I know this has come up before, but since the gulf in cost is so stark I thought I'd mention it again.

I think you might be remembering conversations about mobile service or something. I live in the US and my gigabit fiber (to my house) is fifty bucks a month. The US is a big place.

The causes are complex, and interesting. One of the big challenges for the British aristocracy was that American farming started to destroy the economics of traditional agriculture-oriented landholding in the UK in the late 19th century.

So, the 'bargain' the British aristocracy had sort of explicitly or implicitly made over the centuries with the rest of the British population was that the aristocracy were, you know, the nobles. The knights. We call them 'armigerous' families because the have a _coat of arms_.

The British aristocracy viewed WWI as an opportunity to be like, "Ok, THIS is our time to shine, this is what we've been playing rugby at Rugby for all these years..." The sons of the British aristocracy signed up eagerly, and were put into leadership positions. In the early part of the war, they were still using tactics oriented around personal bravery and charge-leading, so while WWI was horrific for all of Europe with (IIRC) ~10% of the of-age men being killed in the war, the impact on the aristocracy was much worse, with some estimates being around 30% of the young men of the aristocracy dying in the war.

There were far-reaching restructurings of society after WWI, but the aristocracy couldn't come back from the combined loss of a generation of a men, and failure to achieve their gambit: they basically bet it all on "we'll show you, we'll lead the nation through this war" and the public perception was that they completely failed.

So it's not so much that WWI did it, as it was basically done, and WWI was a last gasp.

Like I said though I read this book in...I wanna say 2006? It's been a while.

In general, I think it's kind of an interesting heuristic to think about every so often: right now, putting aside practicalities like my current job or where I live, what looks like a cool place to work, even if the specific role was just taking out the garbage.

Right now the answer is "tailscale" and "oxide" for me.

Eddie Lampert didn't start ESL Investments until 1988 (and he was only 25 then). ESL didn't start destroying Sears / itself until 2004 (https://www.investopedia.com/news/downfall-of-sears/ "Kmart announced it would buy Sears for $11 billion in Nov. 2004..." It was via the Kmart merger that ESL came to control Sears, ESL having taken control of Kmart after Kmart's bankruptcy in 2002).

Anyway, you might be remembering a different company or different "business genius"?

This was during a strange time when I experimented with like....i wrote some code to recognize chords on a midi controller and map them to key sequences, so like C-major->switch to home desktop. that was kind of funny. The only thing that was even vaguely useful from all these footswitch etc experiments was using grid of drum pads as window management tools. e.g. shortcut to four virtual desktops = top row of four drum pads. But even that I don't actually use.

anyway, ymmv

@luu-- I'm sure a lot of people on here besides me would be very curious to read your thoughts on what distinguished these managers from the others, and how you would identify one (nevermind find one).

"(Bonus question: what did the U.S. give to China in exchange for its china?)" "(the answer to the question I posed above about what the U.S. traded to China in the 19th century is "fur")"

My understanding is that in this period, the most valuable China->US exports were tea, silk, and porcelain. The USA didn't produce anything the Chinese were interested in buying (unlike the Spanish, who had had torrents of silver coming from South American silver mines), and so British and American traders in the Canton System and Thirteen Factories period traded opium from British India to China, which of course led to the Opium Wars, British control of Hong Kong, etc.

I've never read anything about fur exports from USA to China being a big part of 19th century USA trade, but I have read a about mid-19th century booming East Coast and European demand for beaver pelts.

We can argue about whether or not it's good, but its not unthinkable. Property taxes in most places (in the USA, the subject of the article) are linked to the current market value of real estate. Sometimes property values rise, taxes follow, and people have to move.