HNHacker News
TopNewBestAskShowJobs

jl2718

3,462 karma · joined February 17, 2018

submissionscomments
jl2718··on The business of takehome assessments
It’s just astounding to me that people can pay $x00,000 to spend 4-10 years on a single discipline, and still not have the minimal required proof of competence to get a job.

And how are these big tech firms spending $billions on recruiting and interview time, as if the education credential is worthless, but still preferring to use it as a first-line filter?

jl2718··on Desire for loud car predicted by being a man: higher psychopathy and sadism
I am not into cars. Most of the men I know are not into cars. A few of them are very into cars. But an overwhelming majority of the women that I was exposed to at a young age were interested in what kind of car a man drives.

Sometimes the simplest explanation of things is that people are acting rationally for their preferences in their situation. And I think most people have similar preferences on different time scales. Most awful things in the world seem to be related to poor capacity for long-term hypothetical thinking. Unfortunately excellent capacity for long-term thinking has a strong correlation with depression, which is obvious if you seriously ask yourself “then what?” enough times. The loud car guy is having a great time in his own little world.

jl2718··on Waterfront home price down by 75% as erosion concern washes value away
I can explain. This bit of Melville exposes some uncomfortable truths about the age in which he was living, the "little ice age". First of all, he's writing from the perspective of a whaler, who may never have seen anything of the island but Coatue, where they would dissemble their ship for transport across the deep sands into the safety of the harbor for winter. The main inlet was too shallow for transport, and later had to be dug out to bring in ships for development of the town. Around then or soon thereafter the bulk of the island was used for farmland. The location of the house in that article was called "Ram's Pasture", but I had observed it to be impassably dense for such a use, so I was curious and sought to discover why it was once so barren. It turns out that the original natives were quite reckless with fire. They would use it as a hunting tool, and quickly depopulated the entire island of all large flora and fauna, leaving the Wampanoag to find subsistence only in the shores thereafter. When Europeans negotiated land rights, they started grazing ruminants, but preferred not to burn, so the forests grew back rather quickly, and the farmers eventually left. This had been happening all over North America, and the rewilding effect was so dramatic that it dropped the temperature of the earth. Property rights saved the world. You won't hear that from most urban hipster environmental activists, but it was the second time I had heard it, first from the eco-resort villages rewilding the East side of Dominica. There 1000 whiteys pay $1000 each to see giant leatherbacks drift in to lay their eggs, but a local with a hatchet got there first for the $100 he'll make selling turtle soup. I digress, but it's really the same thing. The wild western reaches of the island became irresistible to the ultra-wealthy, and those rude pastures grew into some of the most desirable real estate in the world. It's now going back in the other direction, and you can probably tell I don't think this is a good idea. Nature there has tremendous financial value, as single-family estates comprise 95% of tax revenues, but the tourist industry is 95% of the population, more organized, and more vocal by orders of magnitude, and they want golf courses and bars. The truth is that development destroys value. People will pay a lot to be left alone with nature. So that's what happened. Natives who discovered the land quickly destroyed it, then property rights brought nature back for a bit, and now populism is giving it back to the masses to do as they wish, but it seems the ocean will have the final say.
jl2718··on Waterfront home price down by 75% as erosion concern washes value away
“Erosion concern” is exactly correct for this particular location on this particular island, a place I knew very well. The rising sea levels are not particularly concerning in this situation. But why is it eroding so fast? Global warming? Maybe, but much more likely it is the fact that this once extremely verdant forest island has been shorn down to mere grasslands by environmental activists of all people. You see, most of that part of the island is owned by the Land Bank, which is a conservation trust that levies a 2% tax on every real estate transaction, for the stated purpose of purchasing and preserving land in its natural state. Well, so it was, but now they’ve gotten into the golf business, and their last major purchase was a $9M 8-bedroom house for seasonal staff at the golf course. Incidentally, they’ve also discovered a rare flower that they’ve never observed to grow in the thick brush and tree-cover that naturally grows there. So of course they’ve brought over a fleet of earth-shaving leviathan machines to fix nature’s mistake. My labyrinth of sandy-foot green running tunnels was cut down to awkward winding paths on a plain. The deer thinned out, the ground foul disappeared, the armies of migratory birds moved on. When the late October hurricanes swooped in on the island, you could see the aftermath of dead root systems strewn about the beaches, multi-millennial clays beaten back a few feet. It takes a lot to get those feet back. Root systems push out into the newly laid sand in Spring time to get access to partially-desalinated ocean minerals. If enough of them hold on to it together, they can keep it. Erosion is exactly the word to understand. It is not the ocean swallowing the land, but rather nibbling at it, defenseless, rootless, barren.
jl2718··on Steven Levitt and John Donohue defend the abortion-crime hypothesis
This is an uncomfortable and divisive topic, and as such, shrouds rationality. For instance, if you believe that the subject of this matter is morally wrong, then you should rationally also be able to fully accept evidence that your particular beliefs lead to a negative outcome. If you are both arguing about the outcome of your beliefs, and unwilling to change them with evidence, then you’re either holding a rationally delusional belief, or engaging in deception to trick rational thinkers into adopting your morals. Let’s just be honest about why you are arguing for or against an analytical result.

I’m not saying that their study conclusions are correct; I’m just saying that acceptance or rejection of the claim should not be related to moral beliefs. It is, in my understanding of most religions, a pillar of faith to accept negative outcomes in adherence to scripture. It is easier, I think, to accept the self-flagellation form of this edict, than to accept burden onto others, and in this case, a wide social effect that seems incongruent with the aspirations of the faith itself. If they are right, are you wrong? That’s your burden.

I want to be clear that I am not taking a universal stance on the topic nor the analysis, nor religion. My personal experiences related to these topics are moderately severe, and it would probably not be productive to bring that here.

jl2718··on Reddit faces new reality after cashing in on its IPO
It seems to me that there is a tight connection between the timing of this IPO and the windfall demand for their data from AI companies. The advertising and subscription revenue business models have not improved.

Edit: from the prospectus “ Data Licensing Reddit is one of the internet’s largest corpuses of authentic and constantly updated human-generated experience. In an increasingly data-driven world, we recognize that this information is increasingly important for a multitude of different uses and applications. We believe the internet should work for consumers and they should be able to find the information they need or experiences they want. And, organizations need to prioritize sources of real-time human perspectives—from a company looking for feedback on a new consumer product to investors trying to capture market sentiments or signals—and Reddit provides a differentiated solution. Reddit data and information constantly grows and regenerates as users come and interact with their communities and each other in genuine and authentic ways. We expect our growing data advantage and intellectual property to continue to be a key element in the training of future LLMs. We are in the early stages of allowing third parties to license access to search, analyze, and display historical and real-time data from our platform: •Data API Access. Customers can pay to access real-time data streams of anonymous, public discussion on Reddit via data APIs once they exceed certain rate limits. These data APIs are able to provide real-time access to evolving and dynamic topics, such as sports, movies, news, fashion, and the latest trends, unlocking customer use cases such as building third-party applications, behavioral analysis, and algorithmic trading. •Model Training. Reddit data is a foundational piece to the construction of current AI technology and many LLMs. We believe that Reddit’s massive corpus of conversational data and knowledge will continue to play a role in training and improving LLMs. As our content refreshes and grows daily, we expect models will want to reflect these new ideas and update their training using Reddit data. “

https://www.sec.gov/Archives/edgar/data/1713445/000162828024...

So it seems the stock value is speculation on this revenue significantly outgrowing the losses.

jl2718··on The Case Against Children
Ummm.., yes? I don’t know how to respond because I agree with all of this (I think). It’s not inconsistent to say that raising high-contribution children is a massive cost to parents and an even greater benefit to society.

And just enough of them have been ‘tricked’ by religion or culture or noblesse oblige, etc. to keep post-agrarian civilization going.

But also there is a huge number of people that have children that are poorly prepared for net-positive contribution, and they receive almost all of the structural incentives allotted to parents.

There are actually structural disincentives to intensive parenting, like triple-taxation levied on parents that hire professional childcare. It would be easy to make this fully-deductible, but, like I said, the people at the top don’t like competition. Also consider that Social Security is just generational transfer payment, so it might make sense to pay some portion of taxes up the family tree, as well as deducting for any social burden like welfare and imprisonment. You do well, you get rewarded, you screw up, you pay, and there’s a big element of chance, just like anything in real life.

But I really would like to know, do you want less total number of children in the world, or do you want more childless leisure-class adults, or are you genuinely trying to warn people that their lives will get a lot harder with children?

It seems to me that you’re not wrong in the small sense, but in the grand scheme, we still need those hero parents out there.

jl2718··on The Case Against Children
And yet every article written in this same magazine about immigration suggests that our civilization will collapse without a growing population.

Perhaps there is truth to both, but also some unspecified nuance. It’s not just a numbers game. Each person on this planet either makes things better for the rest, or worse. If it’s getting worse, it’s because the size and effect of the latter group grows faster than the former.

Some people can only see life as a zero-sum game, and all good people as competition for their not-so-good schemes. For some reason, the anti-natalists seem focused on pushing their message at elite universities and their high-literacy readership. This is just stupid. Our sustained quality of life depends on children from these people, and probably suffers from those that would never thoughtfully consider the message.

jl2718··on White House urges devs to switch to memory-safe programming languages
Or use a C++ linter that disallows raw pointers, non-const references etc. There’s more, but, it’s mostly a subset of the language, although I admit some common syntax can get bloated this way.

https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2023/p27...

jl2718··on Institutions try to preserve the problem to which they are the solution
This is actually the point, and is analogous to convex optimization, in that systems always evolve along the boundary of constraint, and in the linear case, it will always be an exchange of active parts of the problem.
jl2718··on Notes from an Interviewer
You should always understand that the reason they are interviewing you is because they are unable to solve their own problem.
jl2718··on Institutions try to preserve the problem to which they are the solution
True, but then how do problems actually get solved? Perhaps only by agents that would benefit if the problem were solved.

For instance, send UNICEF into nations with extreme poverty, and you will get back a bunch of sad stories, a bigger UNICEF, and a larger population to feed. Send a cigarette company, and magically the people will find a way to create an economy to fund their nicotine addiction. There are other solutions to the plagues of poverty, such as civil war, narco-states, natural resource extraction, and refugee migration.

Not your intent? Well, you probably didn’t actually want to solve the problem; you just wanted to feel good about your intent to solve the problem, and now I ruined it for you, so I’m the problem.

jl2718··on Why are amplitudes complex numbers? (2018)
This seems to me as backward reasoning. Maxwell’s equations were descriptive of an observational phenomena, which was electromagnetic oscillation, and Schroedinger’s equations used amplitudes to unify them as an energy conservation law. The author asks “why not quaternions”, and gives the reasons, for which he concludes that complex numbers must be the only fabric of the universe. This is backward reasoning, and also narrow to the author’s specialty. Quaternions are just one of many choices for a 4th-degree extension field of the reals, where complex are the only choice of 2nd degree extension. Yang-Mills theory uses Lie algebras of higher degree to form an energy conservation law that unifies the observations of electromagnetism with other observations made at high energy, of which the complex numbers of electromagnetism are a subset.

The article appears to suggest that QM phenomena exist because of the mathematical curiosities of the complex numbers applied to linear operators. Its not my place to say whether the universe “is” mathematics, but I do feel comfortable using math as an attempt to describe it, and there may be many choices, but our convention generally is to choose the description that eliminates possibilities that are rejected experimentally, and many such may be required, but our secondary preference is to choose the smallest set of descriptions that contain the observations.

jl2718··on How to make your anger work for you
I think emotions are over-classified. There are only a few major neurotransmitters that determine our limbic experience, and there is continuity between them. So, in the Kubler-Ross framework for the experience of extreme adverse events, it’s probably more accurate to describe as a transition from sympathetic to parasympathetic, and return to normal.

But ultimately, the cognitive component in all Kubler-Ross emotional stages is the same: a negative comparison of reality versus expectation, on some perceptual timeline. It takes small amounts of dopamine and adrenaline to make changes that affect reality, and serotonin, oxytocin, endorphins etc to accept and ignore reality. Those are really the only choices: do something, or stop caring.

But what happens when you mix them incorrectly?

And this is now more editorial, but I think the problems occur when a brain is constantly hitting the cocktail of accept and ignore to mask the effects of a desire for change. Eventually the inhibitory ligands fail to control the overwhelming buildup, and you get this flood of “do something” mixing in with the “don’t care” cocktail. So that cocktail says, “do something big, and don’t care about the result”.

That might explain a lot of risky human behavior: thrill seeking, promiscuity, gambling, substance abuse, etc. Do something big, and don’t care about the result. It seems completely illogical and stupid, but it’s exactly how we’ve been training our brains since birth.

The “don’t care” cocktail is how we get along and be nice to everybody. And yet we are never good enough, and under threat of losing everything, so we must “do something”, even when there’s not much logical connection between action and desired result.

For instance, the connection between the desire to have a nice family and the action of grinding leetcode is incredibly weak and uncertain, but you have to muster the dopamine for it anyway. And then after years of thankless toil, you finally have a job, but it’s layoff season and you have to accept all sorts of terrible deceit as everybody you’ve been so nice and giving for is stealing credit and shifting blame.

And if you’ve gotten this far, I’ll give you a little story. Things can get crazy when layoffs are announced. One time, this guy was yelling at me about how they didn’t need a data scientist (me), and then he opened up his vest and pulled a gun on me. Real-deal loaded Beretta M9. Pointed straight at me and then slammed on the table. That one messed me up for a while. I know it sounds crazy, but I didn’t associate the things going on in my head with that incident and didn’t say anything, but I was really not feeling well. I was maniacally focused on proving my value, stopped sleeping, and did some relatively groundbreaking work. Eventually I requested two weeks of leave at the worst possible time. While away, I received notification of a prestigious award, and a layoff.

jl2718··on AI models just love escalating conflict to all-out nuclear war
> The researcher's note that the LLM is not really "reasoning,"

But are the humans?

jl2718··on Intel's Humbling
I think it's time to stop talking about node sizes. This reads as meaningless to me, not because I don't know what nanometers are, but because the nodes are all defined by a set of technological changes that have little to do with the measurements, and they are all very incremental and interchangeable at this point, so each node doesn't require an entirely new fab. Actually this has always been mostly true, but the capital investment of new equipment used to be less than the risk of variance from process modification. I don't think this is the case anymore. I expect to see gradually-evolving processes that are sold to customers as performance targets rather than fixed fabrication methods. Even the designs may need some kind of meta-representation that gets 'compiled' to a process, which may have traveller variances like metallization layers.

Of course we don't need to know everything, but in terms of news, maybe we could reference key technologies in use that have the largest effect on performance targets and downstream process modification.

jl2718··on Germany explores 4-day workweek amid labor shortage
Q: How do they define a 'labor shortage'? A: Companies hiring less.

Hmmm... real earnings are down 4-6% every year, so... no.

https://www.destatis.de/EN/Themes/Labour/Earnings/Real-Earni...

jl2718··on Dana-Farberications at Harvard University
Imagine a machine learning paper that released no code, and no data, and was funded by public money, and the IP was owned by a startup that was also funded by public money, and the PI was on the payroll. Laughable, right?

Generally there is more trust when the theory, the data collection, the data analysis, the model development, the product, and the application, are all completely separate and open. It also advances the science much faster. However, it definitely kills a lot of those monopoly profits.

I propose that there’s really no reason that pharmaceuticals research can’t progress in this manner, with great public benefit. And it would be so simple to enforce: all public grants require everything to be developed in the open, not just data release at the time of publication, or pre-registration, but full release at each stage, no patents, and a preference for funding independent dataset production rather than results. As it is, we pay for the research, then we pay monopoly prices for the drug. It’s like these companies don’t really do anything but collect money.

Also, If anybody knows of a way to donate money to cancer research in a way that doesn’t feed directly into somebody’s profits, please let me know. I have recently become loosely connected to a foundation that has been dumping a lot of money into Dana Farber, and this is highlighting a lot of my concerns as I learn more about the industry.

jl2718··on Thousands of small businesses are struggling because of R&D amortization
Who does this benefit in a zero-sum competition scenario, assuming that everybody is affected.

For instance, imagine a competition over the AI toaster market. You have Philburn, the broke PhD student with a great idea, Burnright LLC, the toaster algorithm and electronics designer, BurnCo, the manufacture, and B2N, the toast technology venture capitalist.

A few scenarios: 1) BurnCo hires Philburn for internal R&D. 2) Burnright develops the tech and licenses to BurnCo. 3) BurnCo contracts Burnright for R&D. 4) Philburn starts Smoke, gets B2N investment, and gets acquired by BurnCo.

Which of these, or others, become comparatively advantaged with this new tax treatment?

jl2718··on The Naz.API Credential Stuffing List
Let’s recap the problem:

Passwords don’t work because people can’t remember very much. Password generators don’t work because they have to store the key somewhere. Might as well just authenticate with the encryption key. Encryption keys often get lost, so they need to be stored in the cloud, accessible with a password, which brings us back to the first problem, but adds another point of exploit, or ‘lawful intercept’. Phone number second factor doesn’t work because that can be stolen or transferred, and the same device is probably used for password reset. Authenticator apps don’t work because devices can be stolen, and they require a password anyway to transfer to a new device. Biometrics don’t work because they can be cloned, or used on an incapacitated person, and in the case of phone biometrics, they can be overridden with a pin. FiDO devices don’t work because they can be lost or stolen and used by someone else. Social recovery doesn’t work because of moral hazard and probably the only people you trust that much are not particularly savvy with security around this kind of thing. Practically in an app, they would also have to authenticate in some technical manner, so this cascades.

What have I missed?

jl2718··on What do premarital experiences have to do with marital quality? (2014)
I think it’s sad that common sense traditions that work for the vast majority of the population are only acceptable in a religious context, lest they be forced to fit the demands and ideology of every small and intolerant minority, or worse, the profit motive of the life dysfunction industry.

It’s like, people have had religious tradition for thousands of years, and the anti-religion cult thinks they’re the first ones to question the magic fairy tales. Like, maybe there’s a reason for giving everybody a common basis of values to relate to one another, and maybe inclusive secular humanism has repeatedly failed for lack of premise. Maybe wise people got tired of the debate. Nope, sorry, can’t do that; god says no.

Secularism worked for me as an independent person, but failed miserably in any instance of disagreement over values that affect more than one person. Turns out that my personal values weren’t much different than those of any major religion, but wow, I didn’t realize how far off it would be from other people without religion. My guess is that the marriage program would do very little for people not in the religion, aside from possibly identifying a lot of reasons not to get married, one of them being that a lack of strong cultural grounding in values, such as religion, is a starting disadvantage to an already difficult problem.

jl2718··on Cost of employer-sponsored health insurance is flattening worker wages
1. No negotiated rates. 2. No employer mandate. 3. No special tax treatment. 4. No individual mandate.

Crazy idea: Convert Medicare, Medicaid, and SSDI to universal medical loan insurance.

jl2718··on Will US companies hire fewer engineers due to Section 174?
I would have thought so too. Then I looked at the data.

In the first 11 months of 2023, the treasury ‘printed’ $22.7 Trillion. Tax revenue is about 10% of cash flow.

https://www.sifma.org/resources/research/us-treasury-securit...

jl2718··on Will US companies hire fewer engineers due to Section 174?
Little gems like this are just vestige of an obviously dysfunctional system of revenue. They’ll never get the details right because the premise is wrong to begin with.

It seems to me that the tax system is budget-negative for the US federal government. If we say that the budget is sustainable when the ratio of spending to GDP decreases year-over-year, then the only thing that matters is that economic growth is greater than inflation plus the cost of financing. Elimination of the tax system would imply an immediate economic expansion, and much better allocation of resources toward growth. We’re already quite far from ever thinking that we will fund the government with tax revenues, and the side effect of ongoing multi-trillion-dollar bond issuance is learning that tax revenues are probably irrelevant. So let’s just drop taxes and manage the budget in accordance with interest rate stability.

jl2718··on How Gyms Make Money (2015)
Pull-ups. All the pull-up bars have been removed from public spaces. They all used to have long rows of them from the 50s and 60s, some with iron gymnastics rings, sit-up anchors, parallel bars, etc.
jl2718··on Apple to halt Apple Watch Series 9 and Ultra 2 sales in the US this week
This seems like a problematic patent based on two of the doctrines, both of which the courts have largely abandoned. Firstly, it’s patently obvious. The pulse ox technology in a transmissive mode is very old, and so is HRM in a reflective mode, and the only difference is in the subtleties of the engineering, none of which is reflected in the patent nor the claims. The second reason, relatedly, is the requirement for “replicability by a person of ordinary skill”. The patent reveals nothing that would enable someone to do that, and certainly doesn’t add anything to the ability of a skilled person to do so. The whole thing is just an idea, not an original one, and ideas are not supposed to be patentable.
jl2718··on The Apollo Syndrome
Mark Burnett once warned me about this, saying that it’s been the cause of a lot of his success. In the first Eco Challenge, a team with 4 Navy Seals was beaten by a family that included a 72-year-old grandmother, and he bet everything that he would be able to document this in a real-world setting.

There is a wrinkle in this, however, something he doesn’t talk about. The fifth member, though much better than average within the competition, was not a Navy Seal. Mark was a commando himself, and knew what would happen.

jl2718··on US nuclear-fusion lab enters new era: achieving 'ignition' over and over
It’s a high-flux isotropic neutron source intended for things that have nothing at all to do with power generation, but, possibly great improvements over spallation sources for certain applications.

Or, from the more bureaucratic viewpoint, it’s a successful sale of ‘science’ to the congress for your money. You paid about 0.1% of your gross income for it (ROM).

jl2718··on The key to mother and child well-being may be many caregivers
> in indian communities you have to marry within the community

Is this not evidence for the original point?

jl2718··on Candela P-12 taking off – Electric hydrofoiling passenger vessel [video]
Idea: Voith Schneider Propeller with changing blade path and angle such that it has thrust phase on one side and a lift phase on the other.

Basically it would be two giant low-speed spinning hydrofoil discs. Heavy low-end torque for a direct-drive electric motor.

← PreviousPage 2 of 34Next →