J253

True, but I do have a claude.md as well. Point is, best practice agentic development doesn’t seem to be on display.

I also maintain a popular open source repo and have seen similar. If I receive a low effort single-pass obviously-Clauded PR (many people (agents) don’t even bother using the AGENTS.md file I provide right in the repo?!) with no associated issue I also have no qualms closing. I do generally leave a short note about why I closed (volume of these is around 5/week) showing how far off they (Claude) was and encourage opening an issue where we can start a discussion which rarely happens for the drive-by folks. So I totally feel this but I find myself more disheartened by the practice than angry. It feels like slowly watching open-source go the way of email. Open and free until no-cost spam ruined the inbox for everyone...

So if our universe is inside a black hole from a parent universe, does that mean every black hole in our universe contains its own child universe? We could be living in cosmic Russian dolls all the way down?

Thanks for all the comments so far everyone! I’m quickly getting the impression that I need to work on my resume. Also, I have never used a recruiter. I apply directly to the jobs through their website. Perhaps I need to rethink this approach as well. Again, thanks for the responses! Also I might have to bite the bullet and create a LinkedIn.

Never heard of the race before. Highly recommend reading about some of the quirks surrounding it:

https://en.m.wikipedia.org/wiki/Barkley_Marathons

Agreed. I find Vivaldi’s split pane ability incredibly useful. Surprised other browsers haven’t implemented this kind of this yet.

I’m a home roaster and buy my green beans from a website called www.sweetmarias.com. Everything about this company is awesome but there’s guy there named Tom who knows probably as much as anyone could know about the entire coffee process from growing to wholesaling to roasting. He always travels to farms to meets with growers before purchasing. I guarantee if you get a hold of their customer service and explain your situation, you’ll be able to get a hold of someone there who can give you answers to most of your coffee-related questions.

As easy solution to this is the just not give any credit for homework that goes toward a final grade.

Have a number of in person closed book tests that are difficult enough to prove satisfactory understanding of the topic.

Homework is just a mechanism for a teacher/professor to force a student to dig into and really understand a subject. If the student can use these AI tools to subvert that, then instead of wasting energy or god forbid even more AI to fight the AI, just give students a good enough incentive to actually do the homework. Having the entire grade rely on several of these difficult-enough in-person closed-book tests should be incentive enough.

In my opinion and experience, I think you’re right about “there’s Pandas for that” and “that” can be almost anything. It can do almost anything but making it do almost anything requires constant reference to the docs. And I find maintainability difficult. It seems like there’s 50 kwargs for every method. Sometimes things happen in place by default, other times they don’t. Compound indexes still confuse me. But I’m not a data scientist so I don’t do much ad-hoc analysis that seems typical with pandas users.

This is a great resource for pointing out the things you need to think about in designing a language.

To see the nitty gritty line-by-line walkthrough of everything that goes into actually building a language (all the way down to writing your own VM) I HIGHLY recommend reading Crafting Interpreters[1] by Bob Nystrom. I’m not a language hacker but found everything about this book worthwhile and very interesting.

[1] https://craftinginterpreters.com/

To prevent some of the issues with blurring or compression leaking sensitive info, one simple workaround I’ve used is to just put black boxes over the text as usual in any program and then take a screenshot and save it out from there. No accidental history or compression leaking.

> Netflix is on the very fast downward slope

I definitely don’t disagree with this, but here’s a question for people: if Netflix used their original content budget to instead procure existing content like they originally did, would this save them? This is obviously complicated with everyone creating their own streaming service to charge for their own content, but I’d certainly be happier with Netflix if I had a much better selection of other stuff and not “Netflix Originals”.

Something that greatly helped me was understanding the concept of theory of mind [1]. As soon as I learned about the concept and really understood it and successfully started practicing it I became a much better communicator and listener and it helps in nearly every situation involving humans, be it work, personal, or anything else.

[1] https://en.m.wikipedia.org/wiki/Theory_of_mind

Never heard of that experiment. Just read up on it, though. Fascinating! Thanks for sharing.

Somewhat related, in high school I took a photography class and on the first dark room day, we got paired randomly with someone else in the class to practice taking film out of the canister and a few other dark room things. The person I got paired with and I were from very different circles and would never have exchanged words if it wasn’t for this exercise.

We were only in that room for 15-20 minutes, but I had some of the deepest most genuine conversation I’ve ever had in that sort period of time. After we left the room, we never spoke again. Something about the safety of blindness made us feel comfortable with each other.

If I had instead struck up an AIM chat or a phone call with this person, I guarantee the conversation topics would have been very different. The physical presence, our actual voices, the absence of light…something about that combination made the experience quite profound for my young self.

> Zillow buying up properties in cash over their own online estimates

A discussion around this probably deserves its own entire page, but as I am currently looking for a home as a first-time buyer, seeing “Zillow owned home” labels all over the area in which I’m looking (as if it’s some sort of positive thing) has been infuriating. All they do is use their scale and data science to push me out of the market several times over!

I present no evidence, to prove this, but from what I can tell, this is what they seem to do:

First, they use scale and data science to find what they consider under-priced homes and buy them for cash. If I happen to find the same home for sale at the same time, what seller is going to go with a traditional loan over a full cash offer who could easily offer more if needed? I can’t compete with that.

Secondly, they don’t do anything to the house (aside from clean it up and taking nice pictures—no value added) and then resell it for multiple tens of thousands more than what they just paid. If this house was originally at the high end of my budget, bought by Zillow, then re-listed for more, I can no longer afford it.

Third, if that house DOES sell for more to someone else, it just drives up the rest of the prices in the neighborhood. A lose-lose all around for me and my family. I don’t see how Zillow being able to flip houses is good for anyone. And by flip, I don’t mean fix up—-I have yet to find a Zillow-owned home that has anything of substantial value added to it—-and saving 1% of closing costs or whatever their incentive is by owning a Zillow house is not enough value add.

FWIW, I’ve spent a LOT of time of the past few years trying to get a hold of reliable MLS market data to do my own sort of analysis to try and find affordable enclaves or diamonds in the rough or whatever and that data is damn near impossible to get free access to. So the fact that Zillow has this data and also has unlimited pockets and FTEs who work on this stuff all day…it doesn’t feel right to me.

And it’s not like they’re driving up prices of arbitrary goods or services—these are single family homes (in probably the most popular price range) they’re inflating when we’re already in the middle of a national housing crisis!

/rant

Edit: spelling/clarification

I’m no expert in all things Apple, but this whole thing seems, to me at least, pretty un-Apple like. It would be easier to rationalize if this was actually a directive from, say, a certain large nation who said “if you want to continue to do business here, implement this capability” and what we’re seeing is Apple trying to spin it as a feature or in some sort of positive light.

> I don’t have a credit card and see no point in it.

If you use a credit card like a debit card (pay it off in full every month) it can be rewarding depending on the card. I make $75-$100 cash back every month on my credit card.

This is one of the nice things about shortcuts. I created a shortcut that will turn off wifi and Bluetooth. You can then add an icon to your home screen to run the shortcut and boom. Both are actually turned off…not just disabled for 24 hours. I also have a shortcut to turn them back on when I need them.

Would an idea like this but applied at increasingly higher rates as you own more homes work? For example, owning a single home, you pay a certain amount, if any. Want to buy a vacation house? Start paying n% LVT. A third vacation house? n*1.2% LVT etc. Are you a pension fund buying up 80,000 single family homes? Tax should be so high on owning this many homes you wouldn’t want to do it.

I’ve never subscribed to the cult of overwork. In fact, I subscribe to the cult of maximizing my hourly rate. If I receive an salaried offer, the first thing I do is take into account company culture and the amount of guaranteed PTO (can’t stand the whole “flexible” PTO thing—always felt like a great way to peer pressure people into not taking PTO) to try and accurately calculate how many hours I’ll work in a year and then divide my base yearly salary by that.

E.g.

Job 1: average 40 hrs/week, 30 days PTO, yearly salary $100k

Job 2: average 50 hrs/week, “flexible” PTO, yearly salary $130k

The maths:

Job 1

40(hr/wk) * 52(wk/yr) = 2080(hr/yr)

2080(hr/yr) - 240 hours PTO = 1840 hours worked per year

$100k/1840 = $54.35/hr

Job 2

“Flexible” PTO is a tough thing to estimate, obviously, but I’ll be generous and assume one takes 20 days.

50(hr/wk) * 52(wk/yr) = 2600(hr/yr)

2600(hr/yr) - 160 hours PTO = 2440 hours worked per year

$130k/2440 = $53.28/hr

So there ya go. Even with a 30% pay “increase”, overworking yourself as a salaried employee at Job2 ends up earning you less per hour than the salaried position at Job1. Obviously, there are all sorts of fair arguments for what _potential_ benefits you might reap from putting in those extra hours, so it’s entirely personal whether it’s worth it or not to you. Depends on how valuable 600 extra hours per year of free time is to you. For me, it’s Job1 999 times out of 1000.

I know that CA (LA, SF, SD) housing has always been expensive but it seems like in the past 20 years or so it’s become cumbersomely so. Unless you rode up on the wave of raising house prices through those years (presumably, mostly boomers, or you’re already ultra-wealthy) buying a decent single family home is out of the question.

But with the boomer generation approaching EOL (sorry, a bit macabre) does anyone think prices will have to cool a bit as supply slowly starts to trickle up a bit? Or will that just attract other wealthy people from elsewhere into the market? The current price trend can’t possibly keep up at its current pace or no one will be able to buy…

> on a sonogram, they appear as a single line.

Which is akin to dragging a single finger across different piano keys. Only a single frequency, or note, played at a time. This is common among songbirds.

Contrast that with the sound of a crow. The sonogram is much more broadband in signature. This is akin to mashing a bunch of keys on a piano all at the same time. Many frequencies present at simultaneously.

I have a (perhaps unpopular) pet theory that college as we know it going to cease to exist sometime in the next several decades. In its current direction it feels like it’s becoming more about gaming the system. Those who do the “best” are really just the best at looking like the best.

I saw so many of my peers in college “learning” strictly for the sake of getting good grades, not to understand the material conceptually. I’ll try an analogy—-it feels like college is social media and students are aspiring influencers, all jockeying to find the right aesthetic and be that perfect shining light with all the followers but it turns out they’re all optimizing to the same local max and end up as copies of copies of copies. In the end all their resumes look the same. A bit drab, sweeping, and cynical, admittedly, but that’s what I think.

Anyway, my theory is that with ubiquitous high speed Internet, remote work being more widely acceptable, and continued suburbanization, future generations will look less favorably at taking on a huge debt burden to attend in-person college and something _like_ highly niche online trade/vocational schools or guilds will become popular and widely acceptable as a substitute for a four year degree in certain positions.

> The sad part of online ads is that the internet allows an advertiser, with Google's help, to potentially annoy hundreds of millions of people in order to make some small number of sales.

This sentence suddenly made me see all online ads as essentially modern, “intelligent” robocalls at scale.

Ugh...I have a HN Twitter bot that just tried to post this and it got locked.

https://www.twitter.com/hackernewstop10

A great example of Idyll recently featured on HN[0] and the direct link[1].

[0] https://news.ycombinator.com/item?id=23722275

[1] https://rreusser.github.io/explorations/sphere-eversion/

The San Mateo County Health Officer recently published a statement that rhetorically asked this and other similar questions about CA, and I suppose human behavior and response to various types of SAH orders and the like. It was a very well-written piece and was featured here on HN 3 weeks ago[0]. Here’s a direct link[1]. I highly recommend reading to get an honest perspective from someone who has to make these kinds of decisions at the county level.

[0] https://news.ycombinator.com/item?id=25341289 [1] https://www.smchealth.org/health-officer-updates/december-7-...

Just last week, a song from this album popped on my Spotify discover weekly and I really liked it. Until I read your comment right now, I didn’t understand the context of the album. When I checked the screen to see what I was listening to I first thought it was done in parody or in a mocking fashion or something. I had a mental note to go back and research it a bit more to figure out what it really was, and now I know. And I’m suddenly excited to go back and listen to the entire album. Thanks!

In my experience I’ve seen R used in more exploratory/ad hoc type analysis and algorithm development by “non-developers”—-statisticians, scientists, etc. usually without performance consideration—-and that code is then turned into production code with the dev team using Python or C or something more performant or maintainable.