HNHacker News
TopNewBestAskShowJobs

rented_mule

1,755 karma · joined July 4, 2021

submissionscomments
rented_mule··on Geneva makes public transport temporarily free to combat pollution spike
> And why is pointing out anti-Western agenda posts always met with multiple simultaneous downvotes

Maybe because acknowledging flaws in "y" is not necessarily "anti-y"? In fact, it is often "pro-y". I want to improve things I care about. A critical part of that is identifying flaws so they can be learned from and sometimes fixed.

rented_mule··on FCC abandons efforts to make U.S. broadband fast and affordable
> cell phones have basically filled that area in years ago

That doesn't work well in mountainous areas, which are often rural. There are three towers within a couple of miles of my house, but nothing close to line of sight to any of them. ~1/4 mile away I can get full service, but I have none at all at home.

Luckily Comcast / Xfinity rewired our little town a few years ago so we don't have to rely on satellite (not great during the heavy snowstorms and thunderstorms we get). I'd love to use cellular as a backup as we have 3~20 full days of blackouts per year (most of us have generators and/or batteries) and Xfinity goes down 80-90 minutes after a blackout starts.

rented_mule··on Show HN: I spent 6 years building a ridiculous wooden pixel display
Your idea is not too far from marble pixel art machines... https://youtu.be/w1ks0Vy98KI
rented_mule··on My bytecode optimizer beats Copilot by 2x
For simple string search (i.e., not regular expressions) ripgrep is quite fast. I just generated a simple 20 GB file with 10 random words per line (from /usr/share/dict/words). `rg --count-matches funny` takes about 6 seconds on my M2 Pro. Compressing it using `zstd -0` and then searching with `zstdcat lines_with_words.txt.zstd | rg --count-matches funny` takes about 25 seconds. Both timings start with the file not cached in memory.
rented_mule··on Do variable names matter for AI code completion? (2025)
Years later, when all memory of intent is long gone, I'd much rather work on a large code base that errs on the side of too much "throat clearing" than one that errs on the side too little. `sleep 1` tells what was written, which may or may not match intent.

Many bugs come from writing something that does not match intent. For example, someone writes most of their code in another language where `sleep` takes milliseconds, they meant to check the docs when they wrote it in this language, but the alarm for the annual fire drill went off just as they were about to check. So it went in as `sleep 1000` in a branch of the code that only runs occasionally. Years later, did they really mean 16 minutes and 40 seconds, or did they mean 1 second?

Leaving clues about intent helps detect such issues in review and helps debug the problems that slip through review. Comments are better than nothing, but they are easier to ignore than variable names.

rented_mule··on Do variable names matter for AI code completion? (2025)
It certainly can matter in any scope. `x` or even `delay` will lead to more bugs down the line than `delay_in_milliseconds`. It can be incredibly frustrating to debug why `delay = 1` does not appear to lead to a delay if your first impression is that `delay` (or `x`) is in seconds.
rented_mule··on A low power 1U Raspberry Pi cluster server for inexpensive colocation (2021)
100%!!!

I have an n305 with the CPU thermally bonded to its small aluminum case with a quiet 80 mm Noctua fan screwed into the fins of the case. The manufacturer on Ali Express said the fan is optional, but it can get to ~85°F in the room where the computer is, so I want to be careful. At idle, the CPU reports 5-10°F above the room's temperature.

It has 10 TB spread across 3 SSDs and 2 x 10 TB spinning drives attached. It's a Time Machine target for a handful of Macs and a Borg Backup target for several machines, including some across the internet using Tailscale. It's also running Home Assistant with AppDaemon (with dozens of little apps), Frigate (object detection for 3 Ethernet-connected cameras using a Google Coral TPU over USB), Paperless-ngx (15 GB of PDFs), LibrePhotos (1.2 TB of photos), Syncthing, Tiny Tiny RSS, a UniFi Controller, distinct PostgreSQL instances for various of those, and more. I count 21 Docker containers running right now, and not everything is containerized.

The spinning drives are powered down with a smart plug for all but 1-2 hours at night for backups. With those off, the thing sips power... 10-15 W most of the time with occasional spikes up to ~30 W when LibrePhotos is doing facial recognition or Paperless-ngx is updating its ML models. It never feels slow. I've been running one or more servers at home for 30+ years, and this single machine handles everything so much better than any combination of machines I've had.

rented_mule··on Uv: Running a script with dependencies
There is also marimo. For my use cases it has been a fantastic upgrade from Jupyter. I use it in the style of TFA (notebooks are just Python files). In the notebook UI, you can add packages to the script headers. It's so nice to have many notebooks in one directory with independent, but reproducible environments.

Marimo notebooks are easy to diff when committing to git. Plus you get AI coding assistants in the notebook, a reactive UI framework, the ability to run in the browser with Pyodide on WASM, and more. It will also translate your old Jupyter notebooks for you.

For me, what uv is to package managers, marimo is to notebooks.

https://docs.astral.sh/uv/guides/integration/marimo/

rented_mule··on We Turned Claude AI into an L5 Data Scientist [video]
This is a 6-7 minute video created by a friend about success his startup has had using Claude Opus 4 and ContextHub.ai to automate analysis of data from Google Analytics / Google BigQuery that they previously would have done manually (if they had time).
rented_mule··on The AI bubble today is bigger than the IT bubble in the 1990s
I have inside knowledge (that I'm not trading on) of multiple publicly listed companies with "every feature is AI powered or it doesn't ship" policies that are being applied to their entire product or major parts of their product. Those policies come from execs who are primarily informed by science fiction and FOMO, not by an understanding of the available tools.

AI is great when it's the right tool. But the rate at which features are shipping at these companies is plummeting because so many features would be cheaper, easier, and better without AI. It feels like self-gaslighting on a scale that I haven't seen since the dot com days. And I see engineers and their managers struggling to maintain their sanity as they try to improve their products within the constraints of these policies.

rented_mule··on Oakland cops gave ICE license plate data; SFPD also illegally shared with feds
This might be too pedantic, but a dataset is not a tool in and of itself. It's something that can be processed by a tool. And it's not simple for anyone to reproduce without significant access, either to the original observational opportunities or to the dataset itself. Information about individuals is often in datasets and those people too seldom have a say in the security practices used to safeguard it.

Tools (or pick another word that illustrates this distinction) like encryption, hammers, etc. do not contain our information. They are fairly straightforward to reproduce. And therefore nearly impossible to contain. Bad actors will have encryption and hammers, whether we want them to or not. The only question is whether good actors will also have them, or if they will be restricted by laws. This, for example, can make it easier for datasets to fall into the wrong hands, because they are less likely to be encrypted.

rented_mule··on Oakland cops gave ICE license plate data; SFPD also illegally shared with feds
Something that seems inherently different between GP's comment and encryption is that encryption is an algorithm / tool, not a dataset. Not creating literal tools because they might have bad use cases is clearly a bad idea (e.g., fire, knives, hammers, etc.).

I'd say that one thing inherently different about datasets is that they are continually used badly, including by well-meaning actors. Data is frequently misinterpreted, with good intent, to draw bad conclusions.

You might hit your thumb with a hammer. That hurts! People would be a lot more careful if misinterpreting data had such clear, immediate effects on them.

Also, there are many different groups with different passionate opinions in any community as large as this one.

rented_mule··on An Algorithm for a Better Bookshelf
I think there's a distinction between worst-case and adversarial behavior.

For some types of problems, identifying worst-case behavior is straightforward. For example, in a hash table lookup the worst-case is when all keys hash to the same value. To me, it seems like overkill to think in terms of an intelligent adversary in that case.

But in the problem described here, the worst-case is harder to construct. Especially while exploring the solution space given that slight tweaks to the solution can significantly change the nature of the worst-case. Thinking of it as adversarial implies thinking in terms of algorithms that dynamically produce the worst-case rather than trying to just identify a static worst-case that is specific to one solution. I can imagine that approach significantly speeding up the search for more optimal solutions.

rented_mule··on An Algorithm for a Better Bookshelf
Yeah, this seems applicable to algorithmic management of fill factor in B+ tree based databases.
rented_mule··on What's stopping America from going all-in on heat pumps?
We're not replacing our central, forced-air propane furnace because of unreliable electricity. We have stretches where our electricity is out for 1-10 days at a time. That tends to correlate with times when our daily high temperatures are below freezing. The outages are typically caused by trees falling after several feet of snow fall on them before a windstorm finishes them off. We're in rural mountains in Northern California.

We have whole-house backup batteries, which we can charge with a generator for longer outages (we also have solar in summer, but the sun doesn't get above the trees in the winter). The batteries/generator would have to be much bigger to replace that furnace with a heat pump. They will power the propane furnace (mainly the forced-air part). About 1/3 of our heating is firewood, but during deep cold spells, it's hard to keep it above ~55°F in rooms other than the one with the fireplace without help from the furnace.

We do plan to install a mini-split system before too long, and use that when we do have electricity, relegating the central propane furnace to power outages. We already use an evaporative cooler for cooling, which uses far less electricity than AC. We'll only use the AC function of the mini-splits when we have a lot of smoke from wildfires in the area (evaporative coolers blow in lots of outside air).

rented_mule··on Gridfinity: The modular, open-source grid storage system
It's been MIT licensed for a couple of years now:

https://x.com/zackfreedman/status/1650629770156326912#m

or for those allergic to X:

https://lightbrd.com/zackfreedman/status/1650629770156326912...

rented_mule··on Generative AI's failure to induce robust models of the world
> this depends on the antecedent claim that much if not all of reason is strictly representational and strictly linguistic. It's not obvious to me that this is the case

I'm with you on this. Software engineers talk about being in the flow when they are at their most productive. For me, the telltale sign of being in the flow is that I'm no longer thinking in English, but I'm somehow navigating the problem / solution space more intuitively. The same thing happens in many other domains. We learn to walk long before we have the language for all the cognitive processes required. I don't think we deeply understand what's going in these situations, so how are we going to build something to emulate it? I certainly don't consciously predict the next token, especially when I'm in the flow.

And why would we try to emulate how we do it? I'd much rather have technology that complements. I want different failure modes and different abilities so that we can achieve more with these tools than we could by just adding subservient humans. The good news is that everything we've built so far is succeeding at this!

We'll know that society is finally starting to understand these technologies and how to apply them when we are able to get away from using science fiction tropes to talk about them. The people I know who develop LLMs for a living, and the others I know that are creating the most interesting applications of them, already talk about them as tools without any need to anthropomorphize. It's sad to watch their frustration as they are slowed down every time a person in power shows up with a vision based on assumptions of human-like qualities rather than a vision informed by the actual qualities of the technology.

Maybe I'm being too harsh or impatient? I suppose we had to slowly come to understand the unique qualities of a "car" before we could stop limiting our thinking by referring to it as a "horseless carriage".

rented_mule··on Philadelphia Transit System Votes to Cut Service by 45%, Hike Fares
There are huge benefits derived from economies of scale that can be built at the appropriate levels of government. You can end up wasting a lot of resources by trying to do things too locally or not locally enough.

I live in a rural area and it would bankrupt our county if we had to maintain all the miles of road (far more per taxpayer than urban areas) if the state wasn't doing so much of it. The state uses expensive machines to do much of that work efficiently. It wouldn't make sense for our county to buy that equipment and have it sit unused much of the time. So the county would be less efficient at it. And that's before we get to things like duplication of administrative roles around that work.

Sometimes that means urban areas are helping fund some of my local roads. And sometimes that means I'm helping fund their public transport. When done well, we all get far more for less tax money.

rented_mule··on Dev jobs are about to get a hard reset and nobody's ready
> My UML professor said very similar things 25 years ago.

Yeah, it feels like a lot of these stories are the new version of the methodology / database / framework hype we've seen for so many years... with our new tool, you can write Reddit / Yelp / Wikipedia / S3 / etc. in a week. Sure, you can write a prototype that kind of replicates the surface level functionality. How about the actual hard part? Things like scaling, optimization of computing resources across large fleets, evolution of the system to attract more customers, fighting spam, increasingly complicated integrations with 3rd parties, maintenance, etc.

So much of this will turn out like all those things we've seen before. It will take skilled engineers doing a lot of thinking to figure out the right way to leverage the tools to make better systems for less.

rented_mule··on Ask HN: What cool skill or project interests you, but feels out of reach?
Will Pirkle has a couple of books that go through this. I've found them quite helpful. https://www.willpirkle.com/
rented_mule··on Uber Just Reinvented the Bus Again
This varies. UPS and FedEx deliver to my door. The USPS will only deliver to my nearest post office, not to my door.

There is no home mail delivery at all in our town, nor in much of our county. Each affected home gets a free PO Box.

rented_mule··on Getting Past Procrastination
> Action leads to motivation, not the other way around.

I've found this to be very true. A trick I found that made this easier for me is to leave a trivial task to start tomorrow with, often with notes to remind myself what to do. Ideally the trivial task is on the way to something bigger, not finishing something. That gets me into my editor, gets me running the code / tests / etc., and gives me a trivially easy way to get moving. Then the motivation kicks in and I can start moving for real.

The same approach helps me with tasks outside of software development, and even outside of work.

rented_mule··on Faulty 120W charger analysis (Anker GAN Prime) [video]
There are many different philosophies. My neighbor is 84 years old and worked for the US Forest Service for 30 years doing timber management and fire prevention / suppression. He has thousands of hours of experience using chainsaws, much of that on trees that are 100+ feet tall. Even in his 80s, he prefers to use a chainsaw wearing nothing other than shorts and a t-shirt - no chaps, no long pants, no socks, and no shoes.

His argument is that safety equipment makes people complacent, especially when they are experienced enough to move quickly. If you have bare feet / legs, you are much more aware of the dangers and hence more careful. He still has all his limbs and the fingers / toes at the ends of them. Too many of his former coworkers do not. He is nearly deaf, so using hearing protection might have been better.

That's all anecdotal at best, but experts often disagree on the safest approach for themselves vs. others. And what's best for experts is often not what's best for novices, and vice versa.

rented_mule··on Faulty 120W charger analysis (Anker GAN Prime) [video]
Some examples:

- the one that came with your laptop no longer works

- you use your laptop at home and office, and don't want to carry the charger back and forth every day

- you travel a lot and want something smaller/lighter than your stock charger that can charge your laptop and phone at the same time, both at high speed

rented_mule··on Why Property Testing Finds Bugs Unit Testing Does Not (2021)
Is the root of the problem Unicode? Or is the root of the problem the complexity of the union of written human languages? To the extent that it's the latter, Unicode is just the messenger.
rented_mule··on What makes a good engineer also makes a good engineering organization (2024)
I once heard a great question to illustrate this... When you are inspired by a painting, is this your first thought:

"I need to figure out what brand of paints and brushes this painter used so I can paint like this."

Of course not! It's the painter that determines the quality of the painting and not the maker of the tools. A great painter (or filmmaker or cook or software engineer or ...) will understand the unique attributes of the tools available and leverage those to do great work, whether that's in oil or water, black & white or color, Python or Rust. A difference between experts and non-experts is that experts will find opportunity in the constraints of the tools they have available, and not only frustration.

I believe this extends to engineering managers. I want managers who are good at spotting the unique advantages of each person on the team and are good at structuring the roadmap and individual projects to let the team get the most out of those advantages. And I want them to spot where advantages and disadvantages can be paired to enable learning. All this helps teams see how valuable their teammates can be and how much they can learn from each other. I've seen it result in teams achieving remarkable outcomes while having a ton of fun doing it.

Scaling this up further... I want senior management that is in tune with the unique advantages of different parts of the org, and who then use that knowledge to make better decisions about direction. We can go in direction A, B, or C... that group over there is great at A, so let's pick that and give it to that group. Some see Conway's Law as a sort of "doom", but I've seen it used to the org's advantage with this kind of thinking.

rented_mule··on InventWood is about to mass-produce wood that's stronger than steel
The picture at the top of the article appears representative of the surface appearance: https://www.inventwood.com/superwood-beams
rented_mule··on Computer Architects Can't Find the Average
Metrics can be useful when they are thoughtfully chosen for the context in which they are used. The more abstract or synthetic a benchmark is, the less context there is to inform such a choice. Said another way, metrics characterizing synthetic benchmarks inherit/expose one of the key problems of synthetic benchmarks - a lack of context.

In practical settings, the way to characterize performance will be more clear from context, often giving one (-ish) clear metric for each dimension you care about. For example, if an endpoint has a hard timeout of 100 ms, then it's pretty interesting to look at the percentage of real world requests that are faster than 100 ms. If the same solution is also used in a setting where throughput is more important than latency, then an additional metric is probably needed for that use case. Multiple metrics are needed for multiple use cases to capture trade-offs being made between those use cases.

rented_mule··on Using uv and PEP 723 for Self-Contained Python Scripts
Why would you use `uv run file.py` instead of `python3 file.py`? It seems to me that downloading dependencies is the point of doing that.
rented_mule··on Using uv and PEP 723 for Self-Contained Python Scripts
It's a package manager. The job of package managers is to download code that you then run. That certainly has security implications, but that doesn't differentiate uv from pip, Poetry, Cargo, CPAN, npm, RubyGems, ...
← PreviousPage 5 of 12Next →