Edit: I was wrong, Fog Creek did have a "CityDesk" product at one point, but Wasabi was for on-prem FogBugz[1].
2,817 karma · joined December 16, 2014
Edit: I was wrong, Fog Creek did have a "CityDesk" product at one point, but Wasabi was for on-prem FogBugz[1].
Most of my programming career has been in Python/JS and as I've gotten a bit older, I've found bare metal computing a very rewarding hobby. I'm curious if - having done this to pay the bills as opposed to noodling around - you miss that era?
I work at a company with a social media policy. My understanding of it (garnered by clicking through the slides as quickly as humanly possible once a year) is that they only want me to consider these sorts of disclosures if I'm speaking in a capacity where people MIGHT mistake what I say as coming from the company.
My employer is certainly discoverable from my HN account but there's no more value in me adding such boilerplate to my HN posts than there is me signing emails with my mom the same way.
The employee handbook lays out opportunity to take x% of your current hours for x% of your current pay. If you'd be happy working 4x8s and 80% salary + 80% bonus works for you? As long as your manager doesn't object, go for it.
I think this is a great benefit to offer. Abstractly, I imagine that the system means that once you hit an income threshold you're happy with, a 10% raise immediately gets converted into 10% less hours per week. You never retire, just get asymptotically close to doing so.
I look at residential ISPs offering Disney+ subscriptions, Apple Music subscriptions, Hulu subscriptions, etc, and - while I understand the draw for some - I consider "here, have a 'free' trial account for something you wouldn't have bought otherwise that helpfully auto-renews!" a definite negative. It's just another way companies can pay to get their foot through my door. In my ideal world, my ISP is a company that earns a healthy margin in exchange for making sure the dumb bit pipe that connects me to the world keeps working most of the time; continuing to earn that healthy margin is the height of their ambition, and a predictable monthly deduction from my bank account is the sum of our interaction.
I've seen advice in the past that buyers who want to avoid the "smart TV" tracking ubiquitous in consumer TV models should check out TVs that are marketed as commercial displays. This got me wondering: would similar logic hold up in picking a commercial ISP as opposed to a residential ISP?
There's a "small business ISP" in my area offering 1 Gbps symmetrical access for $250/mo for a shared line or $850/mo on a dedicated line + $750 installation.
Handwaving the building owner's permission here, it seems to me that $850/mo split eight ways between all the building tenants must yield a massively better load factor than residential ISPs give for a monthly price somewhere between the introductory price and the actual price for 200/100 Mbps Fios when I last had it in 2019.
I suspect we'd even come out ahead by splitting the $250/mo. Even one other tenant buying in would put the price in Fios territory. Doing $250/mo solo wouldn't necessarily be prohibitive either, but >$100/mo would be a pretty steep privacy premium.
My questions:
- Is anybody on HN using a commercial ISP at home today? If so, how has the experience been?
- Is my assumption - that independent commercial ISPs are less likely to bother with privacy invasive measures to wring out extra profit - a reasonable one?
- It seems like if one went through the trouble of doing this for one's own building on principle, one will have done most of legwork required to be able to repeat it N times. Does anyone have experience running a very small ISP? I could imagine this working quite well as a co-op model.
What if the shoe were on the other foot? What company would hire me if I negotiated a salary and then after we shook hands and everybody had the warm and fuzzies, I said, "Well now, you'll understand this is just the rate I charge, so it is exclusive of federal and state income tax. These are additional fees that I am responsible, by law, for collecting from you on behalf of the government. I've broken the numbers down this way in the interest of full transparency."
It would be cruel and unusual to deprive a prisoner of water because they have no other means of attaining it when you withhold it. Turning off city water service to a property is different. The property owner has other options they can take to get the water they need to stay alive.
This is not insider trading. When an insider tips off an associate, investigators have full information on the pool of insiders. "Who knew about this ahead of time?" is a strong filter. Assuming the attacker had perfect opsec and the attack itself doesn't leave evidence that exposes them, they live in a much larger and murkier pool.
Additionally, the attacker might have known of this vector months ahead of time. This gives them time to lay groundwork, find accomplices, and prepare.
Quick thought experiment:
There is a bar I used to go to pretty often. It was cash only. Aside from cell tower pings, possible Google Maps location history, and N days worth of security footage, there is nothing tying me to that bar. I've known one of the bartenders there for five years now. We grew up in the same town. We're not Facebook friends, we don't talk on the phone, but we've now known each other in this context for quite a while. If you had a perfect "back home" social graph, we're probably ~3-4 degrees of separation. Linking us to one another locally starting from his perspective would involve very invasive investigation (i.e. putting names to faces for everybody on the N days of security footage that bar has archived, a subpeona for bulk subscriber data of people who have been to that bar, etc).
If I try to involve him in my market manipulation scheme, there's the risk he turns me in outright or that he rips me off and keeps the money for himself. Basically, the criminal conspiracy version of counterparty risk. Set that risk aside for a moment. Assume that he's on board with the plan. Also assume that I, as the attacker, leave no digital evidence pointing back to me.
Think about how egregious his trading behavior would have to be to bring enough scrutiny upon himself that the SEC has people reviewing this bar's security footage, building profiles of the randos who have been to that bar, all that. I don't claim that "tipping off the bartender" is the world's most original securities crime, but unearthing that connection is a much more involved process than the cases of "spouse/sibling/college roommate/tennis partner of CEO bought OTM options a week before acquisition was announced."
X people make, say, 1-10 million dollars off of a zany bet on stocks. Imagine how obvious your trade would have to be such that Y years from now, one of those new millionaires moves back to East Bumblef and makes a money-losing real estate transaction with another East Bumblefian (say, moi), and the SEC jumps over a hedge like "ah-HA! We've been watching your accounts this whole time, that other East Bumblefian knows how computers work and lived in an apartment four blocks from your old workplace in 2015, nobody could possibly negotiate this poor of a land deal, checkmate!"
There are many things countries and people do which they will not accept having done to them in turn.
I have a social media account which I check approximately biweekly and definitely considered myself someone who wasn't falling into the attention trap in my pocket. I only read HN, a few news sites, send texts, and play chess on my phone.
I was amazed to see just how much time I spent doing these things. Even if in some way how I'm using the phone is more virtuous or less problematic ("this hour-long article about the architecture of the classic XBox gratifies my curiosity, is informative and gives me more context with which to understand my field!"), that time and attention sink still comes at the cost of everything else in life.
Do you think there would be meaningful changes if the police moved to using terms of enlistment? I can imagine a hypothetical society where signing up to serve as a police officer/constable/whatever would be mostly something people did for a single term because they felt it was a civic duty, or maybe because it let them access govt-provided education benefits.
Nobody on the right or left is really at ease with the militarization of police forces in terms of equipment, tactics, and outlook towards the policed. Even those who support harsh policing implicitly support it as something to be used against other people; no sane person would choose to be on the receiving end of a pre-dawn, no-knock warrant themselves.
Given that, it'd be ironic if it was a different sort of militarization which improved the state of policing.
Thanks for taking the time to look this up. $226 vs $472 is wild. 2.1x the price for the same service.
Playing around with other car companies at DSM for 7/11-7/18:
- Budget: $347 all-in
- Avis $410 all-in
- Hertz $484 all-in
- Enterprise $374 all-in
I was expecting all of these prices to be pretty similar given "economy class rental car" feels like it would be fairly commoditized.
> It's differential pricing as a result of yield management [1]. You can hate on it but it's part of the reason the economics even works for certain classes of goods and services.
Being familiar with the concept doesn't make it more enjoyable to be the one being fleeced!
Playing around with variations on a 7/13-7/17 M-F rental (also trying 7/11-7/18 Sat-Sat, checking in Manhattan, at LGA, and at DSM in Des Moine, IA for comparison) I'm seeing numbers 1.5-3x higher. I'm curious if the difference is entirely price discrimination or if there is something else like a negotiated bulk rate at play.
Numbers are for lowest tier car, same pickup and drop off location.
Manhattan M-F: $72/day base ($348 w/ taxes and fees)
Manhattan Sat-Sat: $59/day base ($501 all in)
LaGuardia Airport (luxury car, economy car is $22/day more) M-F: $94/day base ($512)
LGA Sat-Sat: $73/day base ($693 total)
Des Moines International Airport (standard car, economy car is $6/day more) M-F: $92/day base, $472 total
Des Moines International Airport Sat-Sat: $49/day base ($463 all-in)
----
So...
1. Lowest I'm seeing is $49/day, which is why I ask in the first place.
2. Pricing shenanigans of a Saturday to Saturday rental coming out cheaper than a Monday to Friday rental in the case of a) 7 day Manhattan vs 5 day LGA and b) 7 day DSM vs 5 day DSM are so frustrating. Surely there's room in the economy for a competitor who offers transparent, predictable pricing? Dealing with industries where you know they are milking you based on reverse IP look up/browsing history/demographic profile is infuriating. How can someone be loyal to any company whose pricing model is "the absolute most we think we can get out of you, personally"?
gkoberger said:
∃m∈M ¬S(m)
then mmlgr said pardon me, but:
∃m∈M S(m)
and then you were like:
∃m∈M S(m)≡¬∃m∈M ¬S(m)
but I've got to tell you:
∃m∈M S(m)≡¬∀m∈M ¬S(m)
∃m∈M ¬S(m)≡¬∀m∈M S(m)
∃m∈M S(m)≢∀m∈M S(m)
I've read that companies like Google evaluate employees in an up-or-out manner up to a certain level; before level X, if you've been a satisfactory junior engineer for Y years but have not been promoted to the next tier on schedule, your performance reviews will nonetheless begin being judged against the next tier's performance criteria and the "meets expectations" junior level work you've been doing will become unsatisfactory (even though you're still a junior engineer). You either grow according to expectations as you gain experience or you are managed out. Happy to be corrected if I've misunderstood the system.
Given that, if companies require current employees to progress in their careers to stay with the company, why should prospective new hires from outside the company with umpteen years of experience (for which they will be compensated with higher-than-new-grad pay) get a pass on that same expectation of showing growth throughout their career?
They are referencing the sock puppet accounts the Reddit founders used to make Reddit look like less of a ghost town in the early days[1].
> Huffman reveals that, “In the beginning Alexis and I submitted all the content.” When they submitted content they created new fake user names so that it would look like the site was populated with a large user base, while, at least in the beginning, it was just the two of them submitting the content.
> Huffman explains that this did two things. For starters, it helped set the tone of the site. “Websites have this kind of inertia, and we submitted content that we would be interested in seeing. That meant the content on Reddit, at least for our peer group, was good, interesting stuff. We wanted a site with the most interesting content online, and so we did our best to find it and then we submitted it ourselves.”
> Creating fake users and submitting content through them also made the site feel alive. Huffman says, “Users like to feel a part of something. If they showed up to the website and the front page was blank, it just looks like a ghost town.” He says, “At the time I think we were just embarrassed to have an empty website so we submitted the content, and it worked.” After a few months they had grown their real user base to the point that they didn’t have to submit content anymore.
The original Red Alert came with a map editor way back when as well :-)
It's probably been 22 years (!!), but IIRC it was strictly for editing terrain only. You couldn't place buildings, place units, or script scenarios.
Even visiting the ISS is a form of "going down a gravity well." If every sentence must be accompanied with an asterisk and a "well, actually" explanation, nothing of value ever gets said.
Consider Halley's Comet, which has a volume of approx 960km^3 and a mass of 2.2×10^14 kg. An astronaut standing on the surface could reach escape velocity with a not-all-that-robust vertical jump. For similarly sized celestial objects, the real trick might be not escaping their gravity well.
Granted, it is technically correct to point out the difference is merely quantitative. However, the effects of that merely quantitative difference are so pronounced that in any practical sense there is a qualitative difference in how we would interact with such an object.
A very ambitious project which you will never finish might actually be three small projects which you will finish and seventeen small follow-up projects you'll decide to never take on. Better to get the three under your belt than try to sprint in 20 directions at once and bail with nothing to show for it.
Say that you decide "I'm going to write a personal ontology engine flexible enough to store everything I know, with a snazzy, super intuitive frontend written in <new-to-me framework X>, a plug-in system for easily creating visualizations for specific types of data (oh, like, what if my personal knowledge DB could take FEN+PGN and let me replay and annotate interesting chess games in-line, that'd be so sweet!), a bespoke query language, the whole thing will be backed by <new-to-me storage backend Y> and, hell, I've been meaning to start containerizing apps at work, I wonder if I should be using docker-compose as I do this or maybe like kubectl, ...also I need to dive into the literature around building ontologies, knowledge graphs, etc. Also, it should probably have emacs keybindings. Probably also support for rebinding keys and defining macros too. Also, it should have a pluggable module interface for scraping semi-structured data sources containing things I don't actually know right now but which I would like to learn at some point. Maybe I can use the hyperlink structure of Wikipedia documents as a scaffold to get the thing started. Oh geez, this is really going to need to support multimedia as well, not just text. Really, this kind of thing would be very valuable in many different contexts. It'd be great if it could be used to automatically generate questions about a topic. I could integrate a spaced repetition algorithm with the question generating bit and then I'd have an amazing study tool. Hmmm, I'd better keep the whole thing as flexible as possible so maybe at some point down the road I can productize it! Sure it's a long shot, but it'd be revolutionary if I could pull it off. I'll start tonight!
To steal from Alan Perlis, "When someone says 'I want to build a personal knowledge DB as a side project and I want it to have perfect ergonomics, top grade discoverability, unlimited extensibility, and unparalleled ease of use,' give him a lollipop."
Even if you don't get stuck in feature daydreams or analysis paralysis and manage to start the thing, your "gee it'd be swell if" side project has a five or ten plus year "labor of love" todo list attached to it. The mismatch will sink you unless against all odds this actually is your labor of love.
As a personal example of the scoping into smaller projects which were standalone (though not a good example of enforcing a novelty budget) I wanted to make a personal activity tracker in 2018 or so. Initial conception was the all-singing, all-dancing, kitchen sink Swiss Army knife of personal data tracking and visualization, implemented in all the technologies I'd been meaning to try out. The backend would be a gleaming beacon of modern infrastructure with everything containerized from the outset, code written in Rust (in which I was a neophyte), phone app written in Kotlin (in which I'd only done the Koans, and with no other mobile dev background, but hey, "Anko" seems to be The Thing to use now, hmm)...
Having been bitten by the bug you ask about in the past, I set that aside. Instead, I did a bunch of smaller projects with each on their own bringing a reasonable sense of completion:
1. Two Python scripts - one to configure desktop OS notifications to fire on a cron schedule ("Nag! Update your data for <x>") and one for CRUD operations on time series metadata and data points themselves. Just backed by local sqlite. Finished in <1 day. Hey look, MVP! Time to start tracking data!
2. Docker setup for building a Rust binary and then trimming away build tooling to get a deployable artifact. 1-2 days? Containerization itch: scratched. Could be a 101 level blog post.
3. PostgreSQL docker image with CREATE/DROP TABLE scripts. Half a day? I guess I learned about Docker data volumes. I can refer back to this Dockerfile in the future, which is like 9/10ths of how I make progress quickly on new projects.
4. docker-compose setup for a Hello World level Rust binary that reads a record from a postgres db running in a separate container and writes it to stdout. Half a day? Hooray, orchestration.
5. "Hello World"-as-a-Service web app that uses Rocket to send a Hello World application/text response. Not very long to finish. Novelty framework itch, scratched.
6. Learn about Diesel. Use Diesel to make a single SELECT COUNT(*) query against the DB. Send query result as application/text response on a dummy endpoint. Done. Could probably be broken off into a "Intro to X with Y and Z" type blog post.
7. Update SELECT-as-a-Service web app to make a meatier query and serialize response to JSON. Learn about Rocket request guards for managing DB connections. Another day or two? If step six wasn't enough for a 101 level blog post on the topic, seven definitely is.
8. Port the Python CRUD functionality to a Rust REST API atop Postgres. Deliverable: approximately all the backend we'll need.
9. Update the Python CLI to use the REST endpoints. Another day? Done.
10. Start doing Android Hello World apps. Using unfamiliar APIs in an unfamiliar dev environment in an unfamiliar (but nice!) language. Eww, this sucks. I'm not having fun any more. Maybe I should first learn vanilla Java Android app development, then Kotlin, then build on that basis with Anko, then...
And then, at that point, I realized that aside from enjoying daydreaming about the idea of a personal activity tracker, I didn't actually give a crap about having an all-singing, all-dancing personal activity tracker. While I'd been happily chugging along building the thing, I didn't care enough about the underlying functionality to bother taking two minutes a day to, y'know, actually track the data when the MVP nag script ran on its cron. When I bailed though, instead of "ugh, another incomplete useless waste of time," I ended up with several completed small projects and more knowledge about what I actually wanted (to futz around building a personal activity tracker, not to have one).
Reply to one of the GitHub issues discusses it some[1].
>> What was the motivation to create fe if you already made aria?
> The same motivations I had writing aria: for the fun of programming it; trying to make something terse but still practical. Every line of code has a weight to it with consideration that would be unreasonable outside this kind of project: a project that exists just to exist. Each time you write something it becomes — or at least should become — more clear and more concise than the previous time, and if you repeat this, the original version is by comparison an over-engineered bloated mess. fe is a sequel to aria, and as such ended up smaller, simpler and faster
If there is a buck to be made, who am I to not make it??? "Common sense"? I got bills to pay!