macNchz

The source data, despite the filename, contains Income, Life Expectancy, and Population by Country+Year: https://echarts.apache.org/examples/data/asset/data/life-exp...

I think if I were hiring remotely right now I’d look to create exercises that could be done “open book” using AI, but that I’d validated against current models as something they don’t do very well on their own. There are still tons of areas where the training data is thinner or very outdated, and there’s plenty of signal in seeing whether someone can work through that on their own and fix the things the LLM does wrong, or if they’ve entirely outsourced their problem solving ability.

Here in NYC shore power for cruise ships has been a multi-decade effort. The Manhattan terminal still has no shore power system because it requires an entirely new electrical substation. The Brooklyn one (in proximity to a poor neighborhood) had a system installed some years ago (with an eight figure price tag), but which ships were seemingly not bothering to use. They’ve since mandated that ships actually use it, if they have the capability, and I think they have some kind of incentives for the cruise lines to retrofit their ships for it.

I think most often these days the users themselves are the centralized point of entry for attackers—a focused attacker is not necessarily going after every application a company uses and trying to breach it individually, but instead trying to compromise a user who has access to all of them.

So in that scenario, centralizing auth is desirable because you don’t have that user holding dozens of weak passwords, and you can monitor access and use heuristics to lock the user out of everything automatically if, say, they access the SSO portal from an unexpected IP. You can also always set things up to require re-auth or MFA whenever someone actually signs into something through the IdP, as an extra layer on top.

I’m not sure there’s a clear winner here beyond using languages and libraries that are very well represented in the training data. While types can let the LLM try to solve certain problems on its own, they’re also, as you mention, a source of verbosity that provide more surface area for hallucinations themselves.

> LLMs work best when everything for a domain, including tests, is in a single file

Context management goes a long way to getting better results from LLMs (why I still like Aider over more agentic tools) and this gives you less flexibility to compose the context yourself based on what’s important to a given task.

I have an M1 13” Air and my wife still sometimes uses my old 2016 12” Macbook (920g). Whenever I pick hers up by accident I’m reminded how noticeably thin and light it is. It seems like a small difference but it changes the feel quite a bit.

For sure, barometers are useful, I'm an avid backpacker and make use of the one in my watch for sudden weather changes and altitude, but if I'm buying a $150+ smartwatch to take backpacking, I just wouldn't really consider one without GPS.

It's kind of weird segmentation, but, given neither has GPS, I wonder how many people who'd seriously consider buying one of these really care about having a compass & barometer. Are those often important features to someone who doesn't care about GPS?

This type of situation is why I use Docker for pretty much all of my projects—single package managers are frequently not enough to bootstrap an entire project, and it’s really nice to have a central record of how everything needed was actually installed. It’s so much easier to deal with getting things running on different machines, or things on a single machine that have conflicting dependencies.

In many places in the US you might find it hard to get a plumber to even return your calls or come out to your house for a smaller job, since the good ones have as much work as they could ever want. Not a perfect comparison since plumbers tend not to work for giant companies that can lavish them with perks, but there are parallels in in-demand skilled labor getting to set some of their own rules.

For a long time I was sort of baffled by the (bad) level-of-fitness and restedness metrics I was getting in a couple of popular apps.

At some point I decided to figure out what the deal was with it, and I saw that their calculations were based on resting heart rate, but because I only ever wear my Garmin running watch during actual runs, it was inferring my RHR from the lowest HR it picked up before or after my run, and reporting to every connected app that mine was like 120bpm (my real RHR is more like 50bpm). Everything was seemingly happy to just accept the garbage data and plug it into their formulas to produce garbage output. Really they probably should have just told me to see a doctor if they actually believed I had a 120bpm resting heart rate!

After that I turned off all of the automated metrics—it was a big realization about how easily messed up they can get (even if you wear your watch most of the day, you might have a fairly different RHR based on whether or not you sleep with it on). I can see this exact situation causing someone a lot of undue anxiety.

I bought one of the refreshed ones in 2016 as a sort of "I want something like an iPad but I just can't tolerate not having a real keyboard and OS" thing and was super pleased with it.

The size was really so good—great on a plane, so light I never thought about it in my bag. I took it all over for years. Wrote a little iTunes library export tool in the back seat of a minivan on the drive from NYC to Maine right after I got it^. It's nice to have an actual laptop when traveling rather than just a phone or tablet, but even a little bit bigger and heavier can be annoying. My wife still uses it for web browsing and has only just recently started to mention it being unusably slow (it was never fast, but it was passable). I have an M1 Air now but would absolutely prefer something a bit smaller.

^ https://github.com/drien/itunes-to-sql

The ad actually sticks out in my mind not only for the visuals but because it introduced me to both Jose Gonzalez and The Knife.

They're very similar, but handling both together my old 2016 12" is just slightly smaller and lighter in a way that is really appealing compared to my 2020 13" M1 Air. It's 350g lighter, which makes more of a difference than I would have guessed when throwing it in a bag.

Cooking shows are a perfect analogy for this stuff. For some reason I never connected the highly-edited-mass-appeal "watch someone do skilled work" videos on YouTube with Food Network style content until just now, but you're right they're totally scratching the same basic itch. They make people feel like they're learning something just by watching, while there is really no substitute for actually just doing the thing.

> Is it really the case that it's ok to threaten people or to call for violence against a community in the USA?

In cases like that, courts in the US make the determination as to whether the speech is protected based on whether it may cause “imminent lawless action”—hard to say in the Cheshire case without seeing what they were actually posting, but encouraging people to start violent riots would not be protected in the US either. There are plenty of cases of people getting arrested in the US for Facebook posts inciting violence.

Check out Amazon Sidewalk—if any of your neighbors connect a device supporting it to the internet, your devices will find them and use their internet connection if you don’t provide one! Apparently their network covers like 90% of the US population (at a low data rate), and they’ve made it possible for non-Amazon devices to implement it.

Densities have changed quite a bit across American suburbs over time—Brooklyn Heights, considered to be one of the first commuter "suburbs" in the 1800s is like 75k/people per square mile, Levittown, perhaps the first real planned cul-de-sac community in the late 40s: 7.5k, whereas a modern single family subdivision suburb is more like 2.5k.

Something that turned me on to these issues in the first place was reading through the P&Z committee minutes for my rural New England hometown, where simply opening a hair salon in an existing, vacant commercial space in the village center has been a multi-year affair because their variance for not having off-street parking expired when the previous business closed.

It's in a 150 year old commercial building and has had a variety of even higher-traffic businesses in it for decades, but now the town has created this artificial barrier to filling the existing space, which clearly functioned just fine before (the town has plenty of parking), and is dragging the process out through meeting after meeting. It's frustrating to watch.

Funnily enough, the fact that these drones are being employed to deliver "toothpaste and batteries" is really kind of a perfect microcosm of the kind of American planning and zoning foundational to modern NIMBYism—in a different environment, daily necessities like that would be available at a corner store very close to where people live, but we've found it preferable to make such restrictive residential areas and require businesses to have large parking lots such that corner stores have disappeared in favor of strip malls and big box stores, and now we're trying to replace those with annoying drones.

Reasoning models pretty reliably seem to do worse at exacting output formats/structured outputs—so far with Aider it has been an effective strategy to employ o1 to “think” about the issue at hand, and have Sonnet implement. Interested to try various approaches with 3.7 in various combinations of reasoning effort.

My rural American hometown has surprisingly expensive homes because of highly restrictive building codes, many of the houses are owned by retired people or wealthy people from out of town, there's little industry to speak of, and multiple people from my 30-student 8th grade graduating class who stayed in town and did not leave for college have now died of heroin overdoses.

> The wealthy fat cats that are making money producing their stuff overseas or simply operating as middlemen for overseas manufacturers are going to have a reduction of income and profits.

The companies who moved manufacturing abroad and pocketed the savings for their investors and CEOs aren’t likely to just say “oh well” and take a big hit to their margins—they’ll just mark up their products to make up the difference. Maybe, eventually, domestic competition will emerge but it’s not a given and likely takes time. In the meantime, consumers are paying the increased costs.

I bought a small HP laser printer a couple of years ago, not realizing that the “e” in the model number meant it’d require an HP account and an always on internet connection.

I did the initial setup and declined to create an account, after which it successfully printed 20 pages before refusing to print anything other than a sheet of paper telling me I needed to create an HP account and finish setting it up with the app.

There was absolutely no way it would print without an HP account, and, even if I’d made one, it’d actually refuse to print if it ever didn’t have an internet connection (even over USB). Naturally it’d taken me three weeks or so to print the twenty pages and it was outside the return window, but it made me so upset that I just put it away and bought a new printer from a manufacturer that’s not overtly hostile.

HP lost me as a customer for life with that thing, and actually made me a vocal critic.

The print quality wasn’t as good as I’d hoped for a laser printer, either.

There are some iOS apps that work as Safari extensions and enable Greasemonkey scripts. I use one called Userscripts.

While it's totally possible to build competently, my impression is that, in the US at least, there are tons of existing houses—built roughly between the 70s (energy crisis) and the mid-2000s when ventilation requirements became more common in building codes—that are fairly tightly sealed but lack any sort of real ventilation systems beyond like, kitchen and bathroom exhaust fans.

Having grown up and lived primarily in (uncomfortably!) drafty old houses, I've noticed the phenomenon ever since I was little kid because they have a distinctive stuffiness and smell of furniture/carpet/plywood even if nothing inside is actually new. I think many people are just used to it/consider it a normal smell of a house because so much of the housing stock is in this category.

They’ve got a 19 year old engineer with some clear links to the cybercrime world seemingly plugging in to things, for one. https://www.wired.com/story/edward-coristine-tesla-sexy-path...

I've been saying for a while now that I'd love to see rules restricting the use of the term "Buy" such that it can only apply to digital products when they are DRM-free and fully downloadable. Anything where the seller retains the right to claw back their product post-sale is more of an indefinite lease or purchase of rights rather than "buying" the product itself.

I think a relatively small proportion of people buying media online fully comprehend that—based on a contract negotiation gone wrong or just the whim of a senior exec—the things they've "bought" can simply be taken away from them. Sellers should be required to make it fully clear (e.g. not just in their 73 page ToS) that they're selling something impermanent and entirely unlike owning physical media.

Totally—on the hardware platform piece, I've owned Macs since the 90s, first installed Linux on one in like 2003ish, tried it on all of my various Macs since at some point, but switched to a desktop Linux machine full time in 2020. I'm generally not one to crap on ambition, but Linux is just so much more enjoyable on hardware that isn't totally closed/undocumented/hostile/unsupported by the manufacturer. With the Macs there was just always something slightly unsupported, regularly broken by updates, or otherwise not quite right, and that was before the architecture switch.

Apple makes great hardware (I have an M1 laptop I use away from home), but if I'm intending to run Linux as my primary OS, I'm buying from a company that is more open to it.

I don’t follow the Linux kernel development process closely, but on most projects—open source or proprietary—that I’ve been involved in, dropping a big diff without having proposed or discussed anything about it beforehand would have been fairly unwelcome, even if it’s something that would be infeasible to break down smaller. I’d also argue that even quite substantial changes can be made with incremental patches, if you choose to do so.