HNHacker News
TopNewBestAskShowJobs

Olreich

711 karma · joined July 9, 2012

submissionscomments
Olreich··on In defense of cryptocurrency
> They're much more likely to do that than to try to run an equivalently fault-tolerant and accessible-to-the-public SQL database themselves (and absolutely more likely to do even that than to trust some entity outside their legal jurisdiction to do that, barring outright state/federal mandates to do so).

No they aren't. They are already running public SQL databases (with limited query access) for managing property changes. And they already have a set of processes and employees to manage real estate transactions and taxes. If a blockchain startup wants to try to get a local government office to let them run it on a blockchain, they will need to prove that the taxes always settle, and that fraud can be reversed. These are both important parts of the system for the local government.

I suspect that there is a granule of use in the idea, and that's that you can farm off the infrastructure for your real-estate management onto third parties and ensure they give you your taxes and respect your authority. But that's not really what people think of when they say "blockchain" with dollar signs in their eyes. The incentives for the people running the infrastructure go out the window too, because the profits are modulated by the government, and it's way easier to just sell them software directly that only has to claim to be effective, and rarely has to deliver on it.

Olreich··on In defense of cryptocurrency
The 2008 paper released by the DOE pegged lights at 6.63TWH per year across all segments in the US. That's 0.02% of electricity usage in the US at the time, which extrapolated is still only 44 TWH in 2008. With 100% LED lights, the usage would be 0.663TWH according to the 2008 DOE paper. I'd assume we aren't quite there, but even so, much more than 2-3 TWH on holiday lights nowadays would be shocking.

> Even if a billion households had 50 strings of LED lights each and kept them lit 24/7 for all of December, they'd still use less energy than the annual usage of Bitcoin.

On the final math, at 5 watts per strand, you'd wind up with 182.5 TWH. Bitcoin at 150TWH-250TWH puts it into the ballpark. Nice.

Olreich··on In defense of cryptocurrency
Lightning network pulls transactions off-chain, thus relying on trusting someone (either the person you transact with or some third party acting as escrow for the funds). By adding in trust, you can mitigate the primary bottleneck caused by proof-of-work consensus methods. But if you have people you can trust to transact with, there's no point in interfacing with Bitcoin at all, just make a micro-payment network without the bitcoin connectivity, like Venmo. This is why we don't see adoption of Lightning Network, it just moves us off of Bitcoin, which lets us get most of the "benefits" for none of the costs.
Olreich··on “What if it changes?”
If you make a good code and build a bunch of things on top of it, but find out it wasn’t matched to a use case and need to move to a new architecture, that good code is going to be a pain to remove or change. “Good” shouldn’t be conflated with “everywhere”. A lot of bad code is everywhere because fixing it seems too painful.

The linked list module may have a poor interface or an awful bug with workaround. But you can’t change it until the next version of the language in 3 years because everyone relies on it. It’s not good, just a dependency. It needs to change and can’t. I want devs to embrace that changing heavily-relied upon things can be good even if it’s painful.

Olreich··on Amazon Q1/2022 [pdf]
Feels like they’re betting on stock split to diversify their investor pool away from big Wall Street firms and toward retail investors who might be more hopeful of Amazon’s investment potential.
Olreich··on The fastest tool for querying large JSON files is written in Python (benchmark)
simdjson can load the json into memory in a queryable form in ~1/3 of a second. So you can save yourself basically 40% of the runtime right there. Computing average should take less than 1/2 a second on modern hardware (assumes <10 million books). So back-of-envelope target speed should be less than 1 second for this benchmark.

11.7s puts you at one order of magnitude off, which could be a fair price to pay if you never need this for large datasets (100s of GB or TB of data you want to query).

And the reason we use wrapped libraries in Python so often is because it’s abysmally slow to do anything in the interpreter. The average loop is 100x slower than it should be. The more math you do the worse it gets too. Most pure Python code is 1000x slower than it should be.

Olreich··on Watching the World Wrecking
Invading a country with mostly non-white, non-christian people: not news

Invading a country with most white, christian people: news

I don’t think there’s any single global conspiracy, just some deeply seated bias in our systems. But when media outlets, governments, and other organizations all individually try to take advantage of or fall into the biases, you get what smells like global conspiracy.

Olreich··on The singularity is close?
The singularity assumes that a human-level intelligence can build a better-than-human-level intelligence. That fundamental assumption is the difference between breathlessly talking about how fast the singularity is coming and dismissing AGI altogether. If we can build a computer than I’d smarter than we are, then I think we can assume that intelligence can compound. But until that happens, we won’t even know it’s possible.
Olreich··on Fast (2019)
No, it’s because of everything. Increased specialization of our construction crews, increased oversight and regulation, increased labor costs due to improved working conditions, improved safety requirements reducing productivity, increased shipping and manufacturing time for increasingly specialized tools and building materials, reduced working hours to cause less impact to traffic and noise, increased sync points to inspect the work and ensure it’s good before moving to the next step, mechanization reducing the number of people that can work at one time because the equipment is large and expensive, increased project complexity because complex materials and construction practices are more efficient in specific ways, etc.
Olreich··on Fast (2019)
Construction time is dominated by staging materials (shipping, delivery, and movement of those materials around on site and into place), and sync points. Houses will sit for weeks waiting on a chain of specialists to get time to finish some critical piece.

In larger projects, there are generally larger staging issues (less footprint to put materials relative to the amount and size of materials), and there are many more specialists involved (including inspectors), increasing the frequency and delays of sync points. The materials for large projects are also much larger now. The Empire State Building exterior was built of bricks, layer by layer. The exterior of most modern buildings are giant panes of glass, requiring cranes and such to move them into place.

I think examining the sync points and staging requirements could shed some light on why construction projects are taking so much longer nowadays.

Olreich··on What we lost
I find it funny that there’s also a competing camp: https://yeshello.org
Olreich··on My seatbelt rule for judgment
If a change has far flung consequences that aren’t obvious, the code is of poor quality. If the reasons for the decisions are non-obvious, the code is of poor quality. I’ve been bitten y assuming that people thought things through the first time too many times. If it’s not in the code base or documentation, I’m assuming that it was done for expedience. Not because they were dumb, but just because that part of the code base wasn’t important when it was written.
Olreich··on I got pwned by my cloud costs
I think this is fundamental to on-demand services. Anything outside terraform or another configuration file system is hard to reason about. If cloudflare is in your config system, then you could put up a warning that files bigger than whatever won’t get cached, but that still assumes a level of knowledge about the system that you don’t generally have.

Setting up limits and alerts as part of the system creation is usually the best strategy.

Olreich··on Reddit Founder Says 90% of Games Will Pay You to Play Them in 5 Years
Roblox business model right here
Olreich··on Efficient Pagination Using Deferred Joins
When I look up something twice, having a “go to page” function is sorely lacking in modern apps. I want to flip to page 300, and they never do that anymore.
Olreich··on Tell HN: Rise in AWS accounts getting hacked and owner being stuck with the bill
Yep, buying goodwill is very effective.
Olreich··on Copyright doesn't need 95 years to get the job done
Fundamentally, copyright is the government promising to beat up anyone that copies a protected work. Now, we don't live in a complete dystopia, so it's not literal beatings. It's legal protection and the beating up is the infringer of the copyright having to pay a bunch of money to the copyright holder. That means that the government is spending resources in order to maintain that copyright. Government is funded by "the people", and so "the people" are paying to beat up anyone who sells a Mickey Mouse doll.

In actuality, the reason it doesn't work like you'd expect is because "the people" are separate from "the government", and it doesn't actually cost much to maintain copyright. But there are a bunch of interested parties in getting to keep monopolies on copying for long periods of time.

But yeah, the reason they shouldn't last forever is because they are relatively expensive to maintain, and some great works can come from public domain stories (case in point, the lovable Mouse who used Grimm's fairy tales as the basis for all of it's most profitable works). Life of author? Definitely. But if the author assigns rights to a corporation? Probably shouldn't allow it to extend 200 years or whatever Disney is up to. Disney won't die, so there is no natural end to the monopoly, and Disney has also proven that they aren't interested in branching out to new stories (perhaps because us humans like the old ones so much).

A reasonable copyright in a digital age is probably something like life of the author, or 25 years after transfer to a corporation. This still effectively provides 75+ years of protection for most works, but then the author gets more say (because transfer of ownership can't happen without giving up monopoly rights sooner), and things like film and games and such aren't locked up for hundreds of years, as it would be surprising to have them owned by a human.

Olreich··on Uncomfortable Truths in Software Engineering
> 10. TDD/FP/Agile zealots are probably zealots because adopting TDD/FP/Agile/Whatever made them better programmers than they were before.

Being a Zealot also stunts your growth beyond the level that zealotry got you to in the first place.

Olreich··on Why Decentralization Matters (2018)
But that's how we got to the current situation with something like email: A bunch of people ran their own servers because they were tech saavy early adopters. Then, Yahoo, Hotmail, AOL, etc. started offering poor email service which most people adopted. Then Google came along and had an offer that was very impressive to consumers and everyone switched to gmail over the years. Sure, a competitor CAN exist, but it doesn't because Google's product was good enough and captured a large enough market share to become the default. We even have people trying to come up with them (see Hey!)

Why wouldn't this happen with web3 servers?

Why doesn't HTTP count as decentralized? Anyone can put a server on the network by buying an IP and domain name. And how does crypto help at all with these basic network effect problems that are the reason previous decentralized systems centralized?

Olreich··on Why Decentralization Matters (2018)
The points about centralized power make sense, but why http, smtp, and social media became centralized in the first place wasn’t grappled with at all.

Running your own servers is attractive for some, but they aren’t the majority and never will be unless it requires something as simple as buying a $35 box and plugging it in at home with no additional setup. Social media is hard to accomplish without a centralized discovery server. See Torrents before they figured out a mechanism or Matrix.

The author didn’t address these forces driving web content to centralized platforms, and web3 isn’t likely to avoid the same incentive structure from what I can tell. Bitcoin and Ethereum both have significant centralization problems for how people use them. More wallets are in Coinbase and other central providers than are run privately. Mining power is centralized among a small population as well. Even current crypto is driving toward centralization to get new people onto the technology, so how do you prevent it in web3?

Olreich··on Beartype: Fast runtime type checking in Python
> emphasizing efficiency, portability, and thrilling puns.

Results: Runs 1000x slower than not using it, but 200x faster than the competition!

So only use this in correctness-critical paths, or use a language that supports strong typing at compile time.

Olreich··on How to build a second brain as a software developer
The best approach is to not build a knowledge management system for yourself until you need it. When you've forgotten something you need a few times, maybe it's time to set up a "common notes" file that you can check first for those things that won't stick. Then when that gets unwieldy, maybe split it into a folder of organized notes with a grep or fuzzy search in front of it (or one of those note-taking apps). When that stops working for you, pull out bigger guns and build something that works specifically for how you use it.

By waiting until you actually need it, you skip a lot of the work in maintaining a knowledge base you don't use, and you only build out the bits you actually need, instead of wasting a bunch of effort on things that sound nice, but won't actually help.

Olreich··on Ask HN: How were video games from the 90s so efficient?
Yes, it’s possible. Obviously things won’t get back to the VGA levels of size because we have higher resolution everything. But you can get similar results by thinking in terms of a memory budget, CPU budget, etc. and then measuring to make sure you stay within that. Once you’ve determined your constraints, then you’ll start optimizing when you bump up against them.
Olreich··on Remote-first work is taking over the rich world – research hints at why
OP missed a category really. If there are any consistent sync points in the day with the whole team, that creates “core hours” where everyone is expected to be in during those, but otherwise the hours are whenever you want. Maybe call it 3.5?

Also of note: you can have in-office work with the same flexible hours schemes.

Olreich··on Manifold: Java typesafe metaprogramming with IDE support by abusing the compiler
I love the “properties” bit. You can get the same functionality by just defaulting the fields on your classes to public. 99% of the time, the getters and setters aren’t doing anything, just proxying the variable.
Olreich··on The Apple A15 SoC Performance Review: Faster and More Efficient
I’ve upgraded piecemeal several times this decade. Harddrives-> Sata SSD -> NVMe SSD. The power supply, fans, and case are all 10 years old and continue to be great. I’ve upgraded the graphics card 4 times, and the CPU/MoBo twice. I’ve upgraded or extended the RAM three times.

There is definitely value in SOC for CPU/MoBo (maybe memory), but the other components work better upgrading at different rates.

Olreich··on Hospitals lift curtain on prices, revealing giant swings in pricing by procedure
So one patient every two years needs one dose of an antivenom. Charge them for the dose of antivenom. Then add up the unused stock as administrative cost and spread that over all emergency care. Simple and effective.

Do the same sort of thing for all other short-half-life things that you must keep on hand.

It doesn’t have to be hard. I get that it is complicated right now, but let’s focus on the idea that simplifying it would help everyone.

Olreich··on Vacation isn’t the answer to employee burnout
The payoff for ICU nurses is a new critical patient right now. They don’t get to see the lives they save reunited with family or recovering at all. They just get a transfer to a lower level of care and a new patient in. They are also seeing a lot more deaths than usual. 10% mortality is in the ballpark normally. With COVID they are seeing 50% mortality. That’s a lot to deal with, and I wouldn’t be surprised if it’s a major factor in burnout.
Olreich··on Welcoming our first riders in San Francisco
We look right on track for my prediction of 2050
Olreich··on The prices hospitals negotiate with private insurers for patient services
Let the government compete with private insurers, and that will be a step in the right direction. The further steps that meaningfully reduce per-capita spending... those are all way harder, which is saying something.
← PreviousPage 5 of 9Next →