HNHacker News
TopNewBestAskShowJobs

rented_mule

1,755 karma · joined July 4, 2021

submissionscomments
rented_mule··on Oracle could face $7B collateral bill for Wisconsin data centre
In terms of what we decide to build, impossible vs. impractical is not a particularly interesting distinction (that distinction can give rise to fun science fiction). Practicality is at the core of good engineering, and possibility is just the first of many steps in evaluating practicality.

If we're going with a distributed approach, there are other ideas that are more practical. For example, send 1 KW boxes to a million people (< 1% of American homes, but there's no reason to limit this to the US). Very roughly, these are beefed-up gaming consoles - you could even start with existing gaming consoles. That box plugs into electricity and ethernet. Pay each of those people $1,000 a month to leave their box running 24/7. For people with a lot of solar at home, it could be a nice income stream. At $1B per year for payouts, it's far cheaper than doing it in orbit. And it doesn't rely on promised but as-yet undelivered technologies. But data centers are probably more practical.

rented_mule··on Oracle could face $7B collateral bill for Wisconsin data centre
> What's the difficulty of a 1GW datacenter in space?

Simple back of the envelope calculations?

Let's assume your numbers are correct and that panels in space can generate 10x the power per square meter per 24 hour period (due to efficiency, lack of night, and lack of atmosphere/weather). That means 10 KM^2 are needed for one such data center. The largest we've built in space is ~3,000 M^2 on the ISS. So one of these DCs in space will require an array ~3,000 times the size of the largest we've built before. Or, it's going to require ~3,000 satellites, each with an array the size of the largest we've ever deployed.

The largest Starlink satellites currently deployed generate less than 30 KW each. We'd need more than 30,000 of those to generate this much power. Starlink has launched ~12,500 to date, with ~10,000 still functioning. So 3x the size of the functioning fleet that's taken 8+ years to deploy. Currently, we are deploying well under 5,000 per year. Let's assume we can come up with enough spare capacity to launch 5,000 per year. That's 6 years to deploy the first 1GW DC. By the time we get to 1GW, the average GPU is 3 years old. So we need to continue to deploy 5,000 satellites per year to maintain 1GW of GPUs that are, on average, 3 years behind current generation GPU designs.

This is all before we get to the economics of launching, the cost of the satellites themselves, heat dissipation, radiation hardening, hardware failure rates (~20% of deployed Starlink satellites are no longer functioning), etc.

rented_mule··on What does privatization of the US Postal Service mean?
I didn't know about it for the first few years I lived here, so ask when you're at the counter.
rented_mule··on What does privatization of the US Postal Service mean?
> Why do rural people deserve subsidies by virtue of living in the sticks?

For the same reason urban people deserve subsidies, for things like public transportation, by virtue of living in the city. The first sentence of the US Constitution lists "promote the general Welfare" as one of the foundational reasons for the creation of the Constitution. There is no "one size fits all" for a nation this size - the welfare of people in the city and the welfare of people in the sticks requires differing allocations of tax money. Also, mail delivery to the sticks benefits people in cities given that mail in the sticks is often sent to or from cities.

rented_mule··on What does privatization of the US Postal Service mean?
I'm not who you're replying too, but I also live in a place with no delivery to the home, despite UPS and FedEx doing so. Everybody in our town gets a free PO Box - I suspect that's how the service obligation is covered. Once a year we have to prove that we still live here to keep the free box.
rented_mule··on Words Are a Byproduct of Consciousness. For LLMs, It's Backwards
There is clearly a level of thinking that doesn't involve language. Participating in competitive sports has been a core part of my family, with some competing at elite levels. When you get to a certain level, if language is involved during the most intense moments, you fail because you act too slowly. Some say you are "reacting" or "using muscle memory" or "using intuition", but some very complex behavior comes from this.

To me, it feels a lot like meditation, or being in the flow when writing code, or being immersed in a video game. I don't have the skills to know, but I wouldn't be surprised if elite chess players are doing the same thing. It's also riding a bike or driving a car, and I think it's why we sometimes can't remember if we stopped at that stop sign behind us, but we almost always did. Language is almost always deeply involved in developing these skills, but gets in the way of top performance once the skills are there.

This feels like a higher level of consciousness to me than language-based thinking. And I think there's a good chance it's the same kind of thinking that we see in animals and primates. Our advantage is that we've figured out how to use language to inform that level of thinking / consciousness more deeply and in more domains.

rented_mule··on SQLite improving performance with pre-sort
This is not unique to SQLite... it should help with any database that uses B+Trees. Batch deletes are also typically faster when the IDs to be deleted are sorted in these same systems.
rented_mule··on We can still stop California's 3D printer surveillance scheme
If this becomes law, it will give rise to a fun new form or protest art in this vain. What is the cutest thing you can design that nobody would consider to be related to guns, but which gets flagged? An obvious example... a llama sitting on the ground, legs hidden, and head held high in the air, chewing its cud. Llamas can be really cute! Sell them on Etsy/eBay/etc., printed by an out-of-state 3D printing service. I just used the EFF form to promise my state senator in Sacramento that I'd send her (and reporters that cover her) one of them if the bill passes.
rented_mule··on Zen and the Art of Machine Learning Research
Sometimes it's a "grass is greener on the other side of the fence" thing or a FOMO thing. But sometimes different roles fit different people better.

I started my career doing mostly full stack work. I couldn't get away from the front end part quickly enough. I had intuition for simplifying UI flows, but none at all for aesthetics. As requests for aesthetic changes came in from our excellent designers, they felt completely arbitrary to me, even though they probably weren't. Most of my career was as a data engineer, data engineering manager, or leading an ML-heavy org. That space fit me so much better.

I loved having a few self-starting front-end devs in my orgs - they could take various tools we were creating for ourselves and make them quite a bit more useful. But it was also always a stepping stone as they typically wanted to work on the public facing part of the product.

rented_mule··on Zen and the Art of Machine Learning Research
Around 2015, I found myself managing back end and machine learning engineers (not researchers) at the same time. Many of the back end engineers wanted to do more ML. Some of them did well when given a chance, but others wanted to revert to back end within a few months. At the same time, one of the ML leaders wanted to step away from ML and only do back end work to support ML.

As I studied these dynamics, something occurred to me... Different people need to see signs of success at different frequencies. Because of the nature of our product, measuring the performance of a new/updated model required the model to be live for at least a full calendar month. So, between initial work and final analysis, it was often a 2 month wait or more. For many back end tasks, you can build a quick prototype, run it to see if it works, and be on your way - the signals come all day long. The varying frequency needs of different people went a long way to determining which of them liked working on ML.

This is sort of a manager's version of feature engineering. ;-) The people on that team taught me a lot!

rented_mule··on DeepSeek Introduces Vision
My understanding is that the core research team is between 100 and 200 people. I don't have a great source for that - a friend of a friend is on the team. By comparison, Open AI's Chief Research Officer said their core research team was about 500 at the end of 2025[1]. With so few people, DeepSeek would have be more selective.

----

[1] https://youtu.be/ZeyHBM2Y5_4?t=483

rented_mule··on How JPL keeps the 13-year-old Curiosity rover doing science
I'm trying to keep this account anonymous so that I'll speak my mind more easily, and she's probably more private than I am, so I don't want to say her name.

Among other things, she worked on mission control software for tracking rover-carrying spacecraft en route to Mars and related ground data systems (I'm probably getting the wording wrong, but that's my outsider understanding of it). Later she worked on software used to analyze interferometer data for exoplanet research on the Caltech side (again, wording?). I'm not aware of her ever being near robotics work.

Part of her shift to Caltech was to try to get to an environment where there was more predictable focus on the science, but a lot of the grant money there was from the same or related budgets as JPL's money, with plenty of strings attached that weren't always pulling in directions she was excited about.

rented_mule··on OpenAI mulls slashing prices as it competes with Anthropic for users
CodeWhale (formerly deepseek-tui) automates this over DeepSeek V4 Flash and Pro. My shallow understanding is that it prompts the model to evaluate the complexity of a given task, then decides on Flash vs. Pro at various reasoning levels for that task. This can help with both cost and speed. If other agent platforms don't already do this, I have to imagine they will at some point.

I'm retired and can't justify spending too much on these things. CodeWhale over DeepSeek is helping me understand this space much better (and have some fun!), and it's quite affordable. I've spent ~30 hours using it over the last couple of weeks, and I've spent $3.89 on DeepSeek in that time. If I don't feel like writing any code for a few weeks, I pay nothing. Looking at DeepSeek's dashboard, about 60% of my requests have gone to Pro and 40% to Flash. I've used 97M Pro tokens and 19M Flash tokens (well over 90% of each have been cache hits, so the price is much lower than it would otherwise be).

rented_mule··on How JPL keeps the 13-year-old Curiosity rover doing science
I want you to be wrong, but I think you're right. The cost discrepancy I pointed out is evidence of that. I've heard a lot of things that are consistent with what you're saying from one of my closest friends, who happened to work at JPL and Caltech from 1992 to 2011, much of that time on Mars rover related software.
rented_mule··on How JPL keeps the 13-year-old Curiosity rover doing science
And if we're keeping costs proportional, send orders of magnitude more rovers and that helps the time argument for rovers as well.
rented_mule··on How JPL keeps the 13-year-old Curiosity rover doing science
The total cost of Curiosity to date is well under 5% of the cost of the recent trip humans took around the moon (something like $3B vs. $90B, or $20 vs. $600 per US taxpayer). Imagine the amount of science that could get done if we gave even half the budget of crewed spaceflight to rover / probe style exploration.
rented_mule··on Federal judge blocks H1B visa $100K fee
I think a lot of people don't understand how different it is to live in rural communities in general, and then to your point, how much more extreme rural Alaska is than "regular" rural areas.

I live in rural California. During the pandemic, a lot of people moved here and didn't last six months before leaving (I can relate, there's a lot I miss about living in SF!). There are so many services people expect, and they expect them to be prompt. We have the opposite of economies of scale in our small town. We have zero options for meal delivery. The nearest full size grocery is a 20-30 minute drive. Costco and Trader Joe's are a 90 minute drive. There are so few auto mechanics around that many of us drive 90 minutes for car service if we don't do it ourselves. Power outages often last 3-10 days. Large snow storms (4+ feet) typically make our roads impassable for 3-7 days. When the power goes out, internet follows 90 minutes later. There's no cell coverage inside or out, even when the power is on.

All of that is worse in almost all of Alaska. My brother worked on Alaska's North Slope for a few years - if you've ever seen the TV show Ice Road Truckers, that was his job. They'd fly him up north for a 1-2 week long shift, then fly him back to the little town he lived in for a week-long break. You have to worry about crazy things like hoping the summer doesn't get too warm, because roads will melt and collapse into what was previously permafrost. Uh oh, someone left a loading bay door in the warehouse open and now there's a polar bear in there breaking expensive stuff. Then you go home and can't go inside because a moose decided to go to sleep right in front of your door - better drive 50 miles and fuel up because you might be sleeping in your vehicle tonight and you don't want to freeze to death if your fuel runs out. These are all things my brother experienced. The pay was great, but he finally gave up and moved to Anchorage to drive a truck locally there.

rented_mule··on London's Free Roof Terraces
The equivalent in San Francisco: https://sfpopos.com/
rented_mule··on Corporate America Is Starting to Ration AI as Cost Skyrockets
My initial use was in a repo where I create models for 3d printing using a library called build123d. There are a handful of parametric models and then many instances of those models with parameters (one that's 24 mm in diameter with a cutout, another that's 42 mm in diameter but no cutout, etc.). I tend to be in a hurry when I want a new parametric model, so I've ended up just copying the one that's the most similar and changing what I want to be different.

The first big task was to find the common bits and abstract them out. It did a great job of creating a plan, summarized in a table, that gave a name to shared chunks, the line numbers in various files where they appeared, line counts of new functions vs. removed bits, and some pros/cons about splitting out each chunk. It was very well "thought out", so I told it to go ahead. It did a nice job other than straying from my coding conventions. That gave me a chance to build out my AGENTS.md file (it helped with that, too).

Once that was done, I had it create automated tests for the newly abstracted parts. I think this is probably a bad practice... I believe humans should at least define what the tests are testing so that there is a deeper understanding of what oversight is in place. But I was just trying things. It surprised me how well it did. The biggest surprise was that the tests seemed quite inspired by vision. It would try different parameters and then have comments about making sure the shape protruded in a certain way, then code that did that. I expected it to refactor a bunch of the code to make it more testable. It found a way to not touch the code while testing everything I asked it to with just two simple mocks - I hadn't foreseen that, but it felt quite practical. It was passing around several opaque tuples in the tests and accessing items in them by index. I prompted it to replace the first one with a frozen, kw-only dataclass. Then a second. On the second request, it saw the pattern and did the rest without me asking. It created 44 tests across a handful of files.

The next part is where I was the least happy. I use ruff and ty to check my code with almost all checks enabled. It was mostly good about the ruff issues. But for the type checking, it just wanted to disable 6-8 rules for the entire repo in pyproject.toml, or at least for all the tests. I had to repeatedly tell it not to and it kept telling me it wasn't recommended. When it finally gave in, it fixed most of the type issues (build123d has lots of types specified, but many operations result in type conflicts because things are so deeply overloaded). The things it didn't fix, it just left a comment to ignore type checking altogether on that line. After I did a little more brow beating, it finally changed the comments to only disable specific rules. To be fair, and unlike most of my other repos, I've had to spend way too much time getting types right in this repo myself.

My last task involved a small library management system for our little town library (tracking library cards, books, DVDs, check-outs/check-ins, etc.). I inherited it from someone who had built the entire web app out of bash/awk/troff scripts with the data in text files burdened by a lot of schema changes that he didn't really know how to deal with. I'm halfway through moving it to Python/FastAPI/SQLite. I asked it to do a security audit of the entire code base, both the newer parts and the old parts that are still in bash/awk/troff. It found everything I knew about and a few things I didn't know about. It made a decent assessment of the risks/impact of each issue. It also called out design decisions that were good security practices. One of the next big tasks will be to see how it does at continuing the migration - it has enough examples of how I've done it that I suspect it can do something fairly consistent with my thinking. I'll probably have it do one or two web pages. When I feel like it understands what I'm after, I'll tell it to use sub-agents to do the rest. I'll be very happy if I don't have to tease apart any more troff scripts that are generating PDF files!

rented_mule··on Corporate America Is Starting to Ration AI as Cost Skyrockets
I retired a few years ago, but I still write a fair bit of code. I was using Copilot's code completion before I retired, but coding agents hadn't come around yet. I've been wanting to try them, but I kept putting it off, and now the price increases make it hard to justify.

So I just started trying CodeWhale (https://github.com/Hmbown/CodeWhale) with DeepSeek V4. I expected to be impressed by the abilities (which still require plenty of oversight). I didn't expect to be completely shocked by how cheep it is. After most of a week of using it 4-8 hours a day, which would amount to a full week of coding in many jobs after you account for non-coding activities, I'm about to hit $3 in total usage. So we're talking $10-20 per month for single-agent use by a full time software developer? And I'm sure some of my usage is waste as I'm still getting my head around things like compaction. If I take a break for a few weeks, I pay nothing because there is no subscription.

If DeepSeek and Xiaomi MiMo stay within a few months of the US-based models in terms of capabilities and US companies don't figure out how to drastically cut prices, I can't see how China hasn't already won. Protectionism would be one reason, but that might be ceding 50-90% of the total addressable market, and bring us closer to moving knowledge work out of the US the same way we did with manufacturing because it's too expensive in the US.

rented_mule··on Print with dozens of colors: Our new open-source ColorMix for PrusaSlicer
I worked on a predecessor to that at Xerox at the start of the 1990s. The first was the Xerox 4850 which was the size of 3-4 clothes washers, cost several hundred thousand dollars, and printed at 50 pages per minute in black plus one color (Xerox called it Highlight Color), but no CMY. It was exactly for the logo use case you mention. A big customer was AT&T for printing phone bills with their blue logo on it. It let them replace many millions of pages of letterhead not sold by Xerox with blank paper sold by Xerox. Xerox loved paper phone bills.
rented_mule··on Print with dozens of colors: Our new open-source ColorMix for PrusaSlicer
> You still need a prime tower when switching tool heads, but that is significantly less than a full purge. But I believe the INDX has the same restriction.

INDX no longer needs a tower. They say there is 13 milligrams of waste (which they call less than a grain of rice) on each filament change. So a print with 1,000 changes wastes 13 grams of filament. Details:

https://blog.prusa3d.com/prusa-core-one-indx-orders-now-open...

rented_mule··on Print with dozens of colors: Our new open-source ColorMix for PrusaSlicer
I can imagine filament vendors making bundles of filaments with interesting mixes of colors. CMYKW is an obvious one, but there must be other color combinations that will mix in interesting ways.
rented_mule··on The four-day workweek in Australia: insights from early adopters of 100:80:100
Chinese tech companies' 996 policies, and large Chinese tech companies in general, are newer than that ethic in the US. What I hear about 995.5 (every other Saturday off) from my friend at Xiaohongshu in Beijing sounds remarkably consistent with what I heard from my Google friends 20-25 years ago, from working hours to on-site amenities that kept you at work. You're spotting a correlation, but I think causality probably goes in the US -> China direction on this.

In the early to mid 90s, I worked at a Silicon Valley based software startup. We had something called "The Century Club". You made the club if you'd done 3 consecutive months in the last year without working less than 100 hours in any week of those months (averaging 100 hours was not good enough). More engineers were in the club than not. We were told that making the club was not mandatory, but nobody in the club was ever fired and most not in the club were eventually fired.

The next startup I was at had a similar culture without the cute name. I remember my most exhausting stretch there was coming in on a Saturday morning, for a database migration that had to happen outside business hours, and working straight through without sleep (other than nodding off at the keyboard) until Monday afternoon. Our CEO was kind enough to bring us food. Even in regular times there, I would go exercise from 10-11 PM, and more often than not I'd go back to the office after.

A decade later I was at Amazon. Our entire group of ~100 engineers was required by our VP to work weekends, in the office, for months at a time when approaching ship dates. The VP would send an email every Friday during this period to remind us to be there. Of course he wasn't there.

Those were all pretty counterproductive, but didn't seem that unusual. The difference in the US back then was that even asking about such things during an interview would often result in no offer because the candidate didn't have a "good" work ethic. Things have gotten a lot better in the US in the last 10-15 years, but a lot of that came from competition for talent. The more that competition eases, the more likely it is that we'll go backwards on this.

Relating back to the article... For the last 3 years of my career (I retired a few years ago), I worked 4-day weeks, and it was all remote. This is just as anecdotal as the article, but I felt I got far more done, with higher quality, than at any point in my career. It was such a revelation.

rented_mule··on Numexpr: Fast numerical array expression evaluator for Python, NumPy, Pandas
I don't know how you'd define niche, but there are many applications where multidimensional arrays of a uniform data type are needed and data frames are not. Image processing would be a simple example where you might have 2D arrays of floating point brightness values that you want to do operations on. Medical imaging is often 3 or 4 dimensions (spatial + time). Analysis of a spatially arranged set of sensor readings over time can easily be more than that. The increase in speed of deferred evaluation is nice in applications like this, especially given that very little change is needed.

You're right that it has trade-offs, like challenges with linting. But many practitioners in these domains are experts in the area of science or engineering involved, not in software development. The ease of adapting an existing script is a big deal for many of them. Many don't even know what a linter is, and numexpr predates (by several years) the high quality linters like ruff that so many of us rely on today.

rented_mule··on Numexpr: Fast numerical array expression evaluator for Python, NumPy, Pandas
numexpr is for different use cases than Polars or Taichi, which themselves are quite different from each other. numexpr is more akin to numba - it speeds up numpy usage.

numexpr speeds up different kinds of usage than numba does. numba is best at speeding up non-vectorizable usage of numpy like repeated operations on arrays inside of for-loops. numexpr speeds up regular numpy expressions, like `5 * x + 7` where x is an array, by avoiding intermediate allocations. It calculates the entire expression for each cell, rather than doing each individual operation into intermediate arrays. It uses strings for expressing calculation so that Python will not break down the expression and hand it off to numexpr one operator at a time like it does with numpy.

rented_mule··on California's Battery Array Is as Powerful as 12 Nuclear Power Plants
Yep. PG&E is the electric company for most of Northern California. Their "time of use" plans typically specify 4 PM to 9 PM as the peak period.
rented_mule··on I believe there are entire companies right now under AI psychosis
Here's a 26-year old post on the exact topic of messiness you raise:

https://www.joelonsoftware.com/2000/04/06/things-you-should-...

A decade ago, I was sitting in on a meeting about a rewrite and, before I could say anything, someone in the first year of her career asked why anyone thought a rewrite would be any cleaner once all the edge cases were handled. Afterwards, I asked her where she learned this. She said "I don't know, it just seems kind of obvious." She went on to be a great engineer and is now a great manager.

rented_mule··on GeoJSON
I'd take your point further... No format is useful for everything. But we have to be aware of the trade-offs of each format (or language or tool or ...) in order to make the right choice of what to use for a given use case. We do that by sharing knowledge of where a given tool succeeds and where it falls down. Pointing out something a format doesn't handle well is not condemning that format for all use cases (I happily choose GeoJSON over other formats for many things).
rented_mule··on GeoJSON
I think what the original comment is pointing out is that GeoJSON lacks a concept of a shared boundary. Shared boundaries expressed in GeoJSON get around that by duplicating data. Whenever data is duplicated, there's a risk that the copies will not be exactly the same. That makes the task of modification more challenging given that the real world is full of messy data, like duplicates not matching.

20-25 years ago I worked a lot with map data from otherwise high quality, and sometimes authoritative, sources like the USGS and NOAA that had this non-identical shared boundaries problem (in formats other than GeoJSON). If the format doesn't allow such mistakes to be expressed, then they have to fix their data to publish it in said format.

← PreviousPage 2 of 12Next →