HNHacker News
TopNewBestAskShowJobs

saulrh

6,105 karma · joined March 13, 2011

[ my public key: https://keybase.io/saulrh; my proof: https://keybase.io/saulrh/sigs/mU5eGace0cFoKIiAp-sVGuZ6iwF9V_AhysRP_3SEG-E ]

I have yet to find a way to retract a misclick on the "flag" button, much to my annoyance.

submissionscomments
saulrh··on A leadership crisis in the Nix community
Not necessarily the case. Consider the possibility that strongly-held political beliefs correlate with software engineering skill, in which case it could easily be that a politically-active but superior engineer comes out ahead despite "wasting" some time on political issues. Yes, they're not as productive as a hypothetical engineer with the same skillset and no political beliefs - but that hypothetical engineer may not exist. Alternatively, consider that forty hours of dedicated software engineering may be beyond the capability of the human brain and that political arguments happen during what would be classified as rest and relaxation time if not for labor standards developed for physical laborers and factory workers in the 1930s. Alternatively, consider that most people work better with motivation, the belief that you are doing good is among the healthiest and most powerful motivators available, that maintaining such a belief requires continuous effort, that such effort would likely involve discussion, and that that discussion would almost necessarily happen at work. I'm not claiming that any of these scenarios are the case. I'm simply pointing out that they are possible and plausible, and thus it is not necessarily the case that an employee arguing politics at work is underperforming, either in relation to their own potential or relative to others.
saulrh··on A leadership crisis in the Nix community
If you want this in formal terms: "pulling the trigger" and "living with yourself after pulling the trigger" are both skills that some people have and other people don't, those skills exist on a continuum of both directness and effectiveness, and comparative advantage applies all the way down at the level of individuals and their personal skillsets. Even if you support your polity's military aims your polity's military aims may be better accomplished if you aggressively refuse to work on military projects, thereby allowing yourself to contribute more effectively in other areas and freeing others up where they can contribute to military projects more effectively than you could. That's one of the things that tax dollars amortize over, if you're looking at it from high enough up.
saulrh··on A leadership crisis in the Nix community
The letter's entire criticism is that Dolstra is clinging to that position through an extended, conflict-of-interest-ridden campaign of toxic behavior and that this toxic behavior is preventing Nix from "delivering useful software and updates quickly to end users."
saulrh··on A leadership crisis in the Nix community
It's like he didn't even read the letter. It's fact after fact, sourced and linked and screenshotted, demonstrating that Dolstra, the Foundation, and Anduril are engaged in toxic leadership that are concretely limiting the Nix Project's ability to build good software. It's not even about Anduril's business, it's about how they're at the center of a web of conflicts of interest that are producing poor software engineering decisions and driving away good contributors. The letter's authorship clearly understood that the obvious response was going to involve tone policing and complaints about "ideology" and went well out of their way to write the letter to address that avenue of attack - the whole first paragraph is about how it's not about ideology, the first sentence is that it's not about ideology, there are dozens of points about how it's not about ideology. It's all about observed cultural dynamics and results.
saulrh··on A leadership crisis in the Nix community
Yeah. Dolstra is responding to a letter that criticized his views, which isn't what the letter is. The letter's concern is that Dolstra is seriously harming the Nix project by forcing all decisions to go through himself, relitigating decisions that don't go his way, driving away good contributors, and generally causing friction and discord that impede communication and render decisionmaking ineffective. Those are pretty dang outcome-oriented criticisms. The letter spends like half of its length explaining how the Anduril thing and the Foundation are concretely harming the Nix project and how conflicts of interest are contributing to the problem. The least objective thing you can point to is the complaint about the Foundation having terrible minority representation, and even that has quite a bit of real-world support. It's the least ideologically-policing open letter I've seen in a long time.
saulrh··on FTC says Amazon executives destroyed potential evidence using apps like Signal
I bet that it does apply to transcripts of meetings. This means, of course, that people with hearing or memory problems will be locked out of leadership because the company can't take the risk that accessibility tooling increases their exposure to discovery. Which is what makes me angriest about this double standard - text is a critical accessibility feature for many people.
saulrh··on Claiming high user satisfaction, IRS will decide on renewing free tax site
For people who aren't aware of this, the IRS is so underfunded right now that funding IRS enforcement has an ROI of between five and nine, that is, for every dollar we put into funding the IRS it would be able to claw back up to ten dollars in unpaid taxes. That's the kind of number that you usually only see on "the next Google" type unicorn startups. It would take eighty billion dollars, invested over ten years, to fund the IRS to the point where diminishing returns bring the ROI back down to reasonable levels; those eighty billion dollars would return almost 200 billion dollars in newly collected taxes. Almost all of that would come from high-wealth taxpayers, large corporations, and partnerships, which are currently very under-audited due to lack of resources - the audit rate is currently lower for high-wealth entities than it is for random poor people simply because auditing people with a lot of money is complicated. The current audit rate is currently dramatically below where it was twenty years ago.

Source: https://www.cbo.gov/publication/57444

saulrh··on Passkeys: A shattered dream
Historically, US courts have declared that giving a password is proof that you control the given asset and that this can be incriminating.

In practice, juries will take a refusal to divulge a password as evidence of guilt, the cops will use it as an excuse for even greater brutality, the FBI is perfectly willing to hold you without trial for years on end, and in most cases they don't need it anyway because everything lives on someone else's computer and they're perfectly willing to hand your data over if they haven't already. Furthermore, because the defense is founded on the principle that the password serves as evidence that you owned the encrypted data, if the prosecution is able to prove that you owned the encrypted data in any other way, that protection goes away.

  > In Boucher, production of the unencrypted 
  > drive was deemed not to be a self-incriminating
  > act, as the government already had
  > sufficient evidence to tie the encrypted
  > data to the defendant
I am, of course, not a lawyer. I'm just summarizing easily available information, i.e. wikipedia.
saulrh··on [dead]
SF traffic fatalities have not been increasing year-over-year. You can't cast blame for an outcome that doesn't exist.

Data: https://www.chp.ca.gov/programs-services/services-informatio.... SWITRS hasn't published a report since 2020 due to the pandemic, but that works perfectly fine for my purposes because I'd have probably had to discard those years for other reasons, and since we're talking about a ten- or fifteen-year trend I'm perfectly okay coming to a conclusion based on data from 2011 to 2020. Traffic fatalities per city are in section 8 chart E.

Fatalities have also not been increasing per mile traveled. Data: https://vitalsigns.mtc.ca.gov/indicators/daily-miles-travele...

I choose traffic fatalities as the metric because federal and state reporting laws mean that these numbers will be effectively complete - there is no sampling or selection effect, unlike numbers about traffic enforcement.

saulrh··on Ford Mustang Mach-E using BlueCruise at time of crash: NTSB
Now we just need people to treat car crashes like plane crashes even when humans are driving.

Unfortunately, given that that would result in dozens to hundreds of fatality accident investigations a day, this is likely impossible.

It's like self-driving cars suddenly sidestepped a hundred years of normalized deviance and everyone is finally, suddenly realizing that cars are fundamentally, spectacularly, and unrelentingly lethal.

saulrh··on Yes, social media is a cause of the epidemic of teenage mental illness
Probably that there's an underlying common cause that means the effect disappears when you analyze the data properly. For example, if screen time is caused by lack of parental presence and lack of parental presence is caused by poverty and poverty also causes depression, then a naive analysis would find a correlation but a proper analysis or meta-analysis would subtract out the common causes of poverty and parental absence and find no effect.
saulrh··on A key chemistry journal disappeared from the web. Others are at risk

  > that duty has transitioned to publishers as
  > journals have switched to a digital format.
I think it's important to note that this can be wholly blamed on the publishers, who froth at the mouth at the thought of any other entity holding even an ephemeral copy of what they consider to be "their" "property". They go out of their way, far out of their way, to prohibit and actively impede availability and durability - after all, a copy that could be distributed after they are gone is also one that could be distributed before they're gone, and they can't have that. And they don't settle for laws, they reach for the technical solutions straightaway, the digital equivalents of printing self-burning books. As far as publishers are concerned, archival isn't just indistinguishable from piracy, it is piracy.
saulrh··on Raising the minimum wage? Study using US tax data finds more gain than pain
If you can only afford to hire someone if they're being subsidized by their parents, then the job position you're offering should either not exist or the subsidy should be made explicit. If our civilization's entry-level positions require five years of internships then we should, as a civilization, figure out a way for people to hold internships for five years and expect every single human alive to go through that process.
saulrh··on Iowa fertilizer spill kills nearly all fish across 60-mile stretch of rivers
For the average HNer, a fertilizer spill is unactionable. There's no digging I can do, nothing I can change. At most I can add it to the big list of things I should consider talking to my senators and representatives about.

By contrast, the liblzma compromise is immediately actionable and HN is who should be reacting to it. I'd wager that some HN readers got paged this morning to deal with the liblzma compromise. There were debian and other distro maintainers on HN communicating about what they knew, what they were doing in response, and what the average HN reader should be doing. I wouldn't be surprised if a reasonable percentage of HN ran, not walked, to run updates on a potentially compromised Fedora installation. Furthermore, the data was all out in the open (git commit history, mailing lists) and in a format that HN readers are specialized for dealing with. A supply chain attack on openssh isn't just relevant to HN, it's the kind of thing that should be on HN because HN is full of the kind of person that we want to be spending time responding to this incident.

saulrh··on Backdoor in upstream xz/liblzma leading to SSH server compromise
Profile guided optimization is, unfortunately, wildly powerful. And it has a hard requirement that a casual link exists from test data (or production data!) to the build process.
saulrh··on ADHD Stimulants May Increase Risk of Heart Damage in Young Adults
Stimulants bring so many other benefits that the risks of cardiomyopathy are outweighed even within their narrow slice of outcomes: taking stimulants reduces non-psychiatric hospital admissions in attention disorder subjects by up to 40%, with stronger stimulants showing stronger protection. https://jamanetwork.com/journals/jamanetworkopen/fullarticle...
saulrh··on Baltimore's Key Bridge struck by cargo ship, collapses
Thanks to things like planetwide bottlenecks through specific canals and locks and design of certain ports, approximately all container ships are within a couple meters of one of only a handful of size buckets, several of which are fairly close to each other besides. For example, a large fraction of the world's container ships are sized to be within a meter of fitting into chinese port facilities and the suez or panama canals. Because these dimensions include the depth of the water as well as the length and width of the ship they also limit the total volume of water which is available to be displaced, which puts a limit on the total mass that can be floated. As a result, "a container ship" is actually a fairly tight and predictable specification of a ship's maximum mass!
saulrh··on Friends don't let friends export to CSV
In my experience, human readable file formats are a mistake. As soon as people can read a single file they think that that's the entire format and that it's okay to write it by hand or write their own code for it. And when everyone writes code based on the just what they've personally seen about a format, everyone is sad. This is why not a single piece of software on earth uses the CSV RFC. This is why people hand you CSVs that don't quote string fields with commas in them. This is why you find software that can't handle non-comma delimiters. This is why you find software that assumes that any field made of digits is an integer and then crashes when it tries to do string operations on it. This is why you find software that can't be used unless you change your computer's locale because the front end understands locales and uses commas for numbers but the backend is running on a server and doesn't know what locales are and now everything is broken. This has happened for every single "human-readable" format in existence: html, markdown, CSV, rtf, json, everything. I consider human readability to be a death knell for a format's chances of broad, easy interoperability. There are certainly other things that can doom a format - parquet is almost too complex to implement and so only just barely works, for example - but I'll take a sqlite database over a csv every single time.
saulrh··on Creating an autopilot in X-Plane using Python
kOS is good stuff, yeah. Very... immersive, heh. It has all of these weird design decisions from having been built to do relatively simple spaceship sequencing tasks first, rather than as a general purpose programming language, and having grown over time to do arbitrary stuff. That history gives it this really uniquely crunchy feeling that you just wouldn't get with a more modern, ergonomic system.
saulrh··on Creating an autopilot in X-Plane using Python
You can also do this for Kerbal Space Program using the kRPC mod: https://forum.kerbalspaceprogram.com/topic/130742-18x-to-112.... It's a much less polished experience, but you get more of the experience of an actual avionics/robotics project doing it this way - multiple sensors that may not necessarily agree with each other (for example, IIRC accelerometers installed at different ends of the ship will register different readings!), you have to be quite a bit more careful when interfacing with engines and may not have access to a simple "go left" control input, you have to deal with whatever nightmare vehicle you built that you need a computer to handle, that kind of thing.
saulrh··on Tortoise: Shell-Shockingly-Good Kubernetes Autoscaling
Doing things I don't understand myself is also a recipe for disaster, and in my experience a rather greater one. The platform team is liable to make mistakes like scaling the service wrong or failing to anticipate upcoming changes. These incidents can be easily resolved by improving monitoring and communication, which are fundamentally useful things that I should already be doing for myriad other reasons. The mistakes I'm likely to make are things like "sequenced a complicated change wrong and null-routed the entire application" or "typo'd a volume name and found out that that autodeletes the entire database including backups", which I am simply not good at avoiding and constitute one of the major reasons I am in engineering instead of ops or IT. We are better off if I do the things I am best at and they do the things they are best at.
saulrh··on Docusign just admitted that they use customer data to train AI
It doesn't work on arbitrary data. Hand-crafted metrics on structured data, yeah. Free text, no. Data that you don't know anything about, also not really.
saulrh··on Kalman Filter Explained Simply
Having written papers like this before, a large part of the problem is that almost all CS research is published in conferences, rather than journals, and conferences frequently have extremely strict length limits, on the order of three or five pages. If you have an even slightly complicated procedure it can be nightmarishly difficult to get even the core information into your paper, and you can forget about details or tangents.
saulrh··on How Lego Became the Apple of Toys (2015)
I disagree wrt licensed IPs.

There's no single right "amount" of structure to provide to inspire creativity. Different works of art, different moods, different people, different days, will all "want" different starting points. Sometimes you want to experiment with basic geometry and it's useful to start from a bucket of raw parts. Sometimes you want to dig into stories and characters and it's useful to start with something that already exists. And because you can always remove structure from a situation - Legos can always be broken down into the bucket of raw parts! - more highly structured sets strictly expand the available range by pushing out the high end.

Even more than that, kids are learning how to be creative, and having some structure to start from helps a huge amount for a beginner. If you hand a blank piece of paper to someone who's never written anybody and tell them to be creative they'll never even be able to start. If you give them something to start from they'll have a much easier time. Even as expertise builds and you pull back on the guidance it's still incredibly useful to have new material injected from outside your own bubble of experience. Concretely, when you're just getting started it's a hell of a lot easier to learn that you can have a different person win a lightsaber fight than it is to write all your own characters from whole cloth, and when you're trying to learn how people work it's absolutely necessary to examine characters and stories that you didn't make up and that don't just confirm your own preconceptions and biases.

If I wanted to summarize my opinions, I think that playing with licensed-IP Lego sets is actually a critical step in the development of creativity because it lets you get started on more complicated topics earlier, makes you less afraid to adapt existing works, and provides inspiration and outside influence to break filter bubbles. I would compare it in many ways to writing fanfiction, and in a very literal sense it absolutely is.

saulrh··on Results of the 2015 Underhanded C Contest
I've always found it... somewhat unfitting, I think... that this is called the "underhanded" C contest, because so many of the winners work by getting trivial type and preprocessor stuff wrong in ways that are concerningly easy for the average programmer. Like, the 2015 winner interpreting an array of n doubles as an array of 2n floats because it screwed up a typedef of float_t, or the 2008 winner reading 3 bits out of something because it left the parentheses off of a macro definition. These are core operations for using the language as intended! There's nothing fancy going on! This is code that you, yes, you, could have written following every best practice on earth. The contest is almost tailor-made to produce evidence that C should never be used for anything that requires reliability or correctness.
saulrh··on Flying kites deliver container-sized power generation
An offshore wind farm by definition uses no land area :V
saulrh··on Towards Modern Development of Cloud Applications (2023)
I think that this might actually be something new - you have the same topology as a microservice architecture, but rather than actually programming and deploying one binary per microservice, you build a single binary with all of the functionality and it gets deployed as a fleet of microservices, running a different subset of all of that functionality depending on what it's deployed as. Which means you have only one build and one deploy step and atomic deploys, like a monolith, but you also get to scale each component independently and isolate data like with microservices. A really nice looking middle ground.
saulrh··on Towards Modern Development of Cloud Applications (2023)
This looks like it parallels the shift from MapReduce's paradigm of one binary per execution graph node to FlumeJava (and Apache Beam's) monolithic binary for all worker nodes that reconfigures itself as necessary for each stage. My experience is that Flume/Beam is a lot nicer in almost all ways, so I'm not surprised that the same thing works for services too.
saulrh··on Terminal Trove – A collection of CLIs, TUIs and all things in the terminal
They did! Nice! Those definitely weren't there earlier today!
saulrh··on Terminal Trove – A collection of CLIs, TUIs and all things in the terminal
I wish the one-sentence descriptions were displayed in the lists. When I click through to https://terminaltrove.com/new/, I have to click through every single one of those names to find out what they are. It'd be a lot nicer if I could swoop through and see "grv: git repository viewer. netop: network topology visualizer. dive: tool for exploring each layer in a docker image." etc.

(edit to add: This is already pretty great, I love curated collections like this, I just also see a way to make it even better!)

← PreviousPage 3 of 34Next →