chromacity

What's especially frustrating is that the posts about AI are also disproportionately written by AI. I can sort of understand that people who are very enthusiastic about LLMs also use them for blogging. But the most bizarre part is that a lot of anti-AI opinion pieces are LLM-generated too. Either cynical click-sploitation or extreme hypocrisy.

How else do you expect me to illustrate my LLM-generated blog posts about AI?

I mean, more or less, we do. The NIH list includes cell phones, e-cigarettes, and headphones.

I never followed up, but I didn't read it as some serious IP licensing thing. It sounded like they've come to the conclusion that they're making the stuff that's sold on Tindie anyway, so might as well set up a website and ship directly to your customers.

Unfortunate. Tindie is (was?) a pretty unique marketplace. Amusingly, a lot of what they were selling was probably illegal due to FCC rules: for the most part, you can't sell electronics without EMI certification and "I'm just a hobbyist" is not an excuse. Kits get a bit of leeway, but finished products don't.

Before the tariffs, I noticed that Chinese companies were trying to undercut them. I've gotten multiple mails asking me to start selling my designs with China-based outlets: they would make the PCBs, assemble them, and pay me some money for every item sold.

Phones don't have removable batteries mostly because of the desire to make the device as thin as possible. The battery is just a delicate, flexible pouch that can easily be damaged and catch fire if removed from the phone and carried around. To make it safe, you'd need to add a hard shell, which would probably make the device 2 mm thicker or so.

As to why we want to make phones as thin as possible... I don't know, but I guess it makes them look futuristic, which helps with sales. The same goes for highly-reflective, glossy screens. I guess I'm not gonna cry if that gets regulated away.

This is an interesting sentiment given how desperate AI labs seem to be source any new internet content from any walled-garden platform willing to take their money (and how willing they are to try & take it even if you don't consent).

Abusive, sneaky scraping is absolutely through the roof.

No matter how many AI tracks I blocked / downvoted on Spotify, they kept recommending them to me. It made me cancel my account. The thing is, from the consumer point of view, you can barely tell if it's playing in the background. Even if 90% of the streams are fraudulent, the remaining 10% is still real people who are none the wiser, but who would probably rather patronize human artists if given the choice.

> Generally speaking I am a fan of the idea that everyone should be required to do some kind of community service for 2 years once they turn 18.

Why? I get the warm-and-fuzzy angle of "instilling civic responsibility", but you're effectively instilling that at gunpoint: the government forces you to do this, else you go to prison. Is that really such an enlightened thing to do?

It takes away two years of your life, possibly delaying your education, entry into the workforce, or having children. So again, what's the rigorous justification for this? The government already calls dibs on a good chunk of your economic output, on a percentage of every penny you spend, and on certain types of property you own; so why should they also be able to draft you for some free labor if there's no war or other emergency going on?

But it doesn't. We have a problem. We can focus on addressing the problem without pre-adjudicating every hypothetical corner case.

If your "work" is mostly AI, and if you don't disclose it, it goes to /dev/null. And yeah, you can get into a debate that it's unfair to reject 51% but allow 49%, but that's how the real world works - otherwise, nothing would ever get done. You also get a DUI for BAC of 0.08% but not 0.07%. That's not an argument for putting DUI laws on hold until we can figure out a more perfect approach.

The same people who read AI-generated stories about AI. Which is, roughly, most of us. There are AI-generated blog posts on the front page of HN multiple times a day. Right now, I see "I prompted ChatGPT, Claude, Perplexity, and Gemini and watched my Nginx logs", which is AI slop. I'm sure there's more.

But these early chips are not that. The 324 has marked crossover distortion and problems with phase reversal, both have severe input and output voltage limitations, etc. Basically, they are a pain to use, especially in single-supply circuits, and there's little reason to keep using them.

Honestly, debating these corner cases feels like a distraction tactic. The reality is that the bulk of that 44% is total AI slop: one-sentence prompts entered into Suno to generate 1,000 tracks and extract money from subscribers who stream in the background.

It's the same thing with writing. No one cares that you asked a chatbot to help you reword a paragraph in your essay. The problem is zero-effort slop delivered by the truckload to your social media feed.

The "it works" part may be true for BJTs, but I think it's a stretch for op-amps. The number of Stackexchange and Reddit threads where people are struggling to debug the quirks of these ancient chips is pretty staggering.

The obvious difference is that the US, more or less by deliberate design, had a remarkably lax approach to visa overstays and illegal border crossings for decades. This resulted in a population of more than 10 million "unauthorized" residents.

Any policy that suddenly pulls the rug on them is notable precisely because we created the problem (or not-a-problem, depending on your leanings) in the first place.

I think the reason is much simpler, same as with 741 and 324 op-amps. These devices just came out at the right time. They featured prominently in the first crop of books that popularized a number of reference circuits among hobbyists and pros alike, and these reference designs have been endlessly copied since (including on the site where this article is hosted). Most people just don't know what transistor or op-amp can safely replace another transistor or op-amp in a non-trivial circuit.

In fact, for any non-trivial transistor or op-amp circuit you Google for, most of the designs you get are incredibly dated. And now, AI answers are reinforcing the same biases, so I guess we're gonna be stuck here for a long while.

They only burrow to a depth of about one foot.

I wish them well, but this has all the signs of nanotech vaporware. The article even mentions graphene, and empirically, that seems to be a death kiss for every promising technology announced in the past three decades.

The tech to make your own PCBs at home already exists, it just doesn't really make sense to spend weeks making inferior boards when you can order custom boards for pennies and have them at your doorstep in a bit over a week.

> The article isn’t arguing that if ICL facilities are disrupted, that’s it, no more bromine forever. It is saying that if these facilities are disrupted there will be an even bigger problem with DRAM supply than already exists because there is no excess supply, no good alternative, and no quick way to ramp up production.

This is literally the thesis of each and every one of these articles. Only one mine in the world can produce sand for semiconductors, etc. It makes the arguments incredibly persuasive and the predictions almost always wrong.

In reality... I'd wager that the semiconductor industry uses very little bromine compared to say, plastics; and that it can be recycled or sourced from other places with minimal technological investment (e.g., as a simple byproduct of salt production in the US).

Ah, this week's iteration of "we're running out of sand". I'm sure one of these predictions will eventually come true, but we have articles that overstate the likelihood and consequences of running out of <some basic material> pretty much every month.

I'm not keeping track, but some of the things we ran out of include sand, helium, tellurium, tantalum, niobium, bees...

I think that's a goofy take. It's two coffees everywhere. Every other software vendor is trying to move to the subscription model. If you add up all the licenses you need to do work, have hobbies, or procrastinate (Netflix, Spotify, etc), is it still two coffees a month?

I know many folks who make $500k+ a year in the SF Bay Area and complain about affordability, and to a large extent, it's stuff like that that makes them poorer.

Also, my point is that there's nothing inferior about solutions such as Capture One, at least not as far as hobby workflows go.

I know you edited your post, but I'm actually taken aback by people trying to argue it's a blog, not "journalism". I see no real difference between this and some of the most celebrated pieces of gonzo journalism.

However, this cuts both ways. This format is how we get some of the most interesting pieces of reporting about culture and counterculture. It's someone who went to some parties or worked for some companies. What you refer to as laziness is what makes it valuable: it recounts specific experiences instead of trying to speak in generalities. And it's descriptive rather than moralizing.

In the same vein, some of the most powerful exposes about neo-Nazi movements are just raw accounts of what's going on inside, without the author constantly repeating "and by the way, Nazism is bad, these people are all bad, and here are some statistics".

The SF Bay Area culture is probably not a thing, but there are some pretty awful subcultures within it, and many of them revolve around performance-enhancing drugs and rationalism-as-a-justification-for-bad-things (Zizians, longtermism, etc). I think we should own it.

Every time I see one of these HN threads, I am actually amazed with what Adobe was able to pull off. I'm not surprised that they could do this to pros who were used to a particular workflow. In fact, for some businesses, a subscription may have some benefits. You were probably upgrading regularly anyway, and the only downside is that it's an expense you can't cut back on in a lean year.

But there are so many hobbyists, including here HN, who just went with it and have given Adobe thousands of dollars over the past decade just to keep using Lightroom or Photoshop! It just boggles my mind. There was a brief period where you had no good alternatives - GIMP wasn't it - but for almost all hobby needs, you now have very good pay-once options (e.g., Capture One instead of Lightroom). It's basically a monthly fee you pay for not having to think about the problem, and people are willing to pay it for many years.

Makes me think I should be doing more bait-and-switch...

It's kinda poetic that in the entire process, authors get screwed thrice. First, by the publishers and retailers, who keep 80% of the revenue. Then, by the hacker culture that enables widespread, institutionalized book piracy for the sake of "information wanting to be free". And finally, by the same hacker culture gone corporate, where "grown-up" geeks conclude that, since we already have all these pirated works out there, what's the harm of training LLMs on that.

Music is quite similar, and I've actually seen piracy justified by saying that "eh, the musicians are screwed either way". And of course, that piracy enabled suno.ai, which is now making sure that the musicians are really screwed.

No. I've been using that construct long before LLMs and I don't think there's anything wrong with it. It allows you to succinctly state the position you're disagreeing with before putting forward another hypothesis. LLMs overuse it for needless emphasis, with the negative example usually reduced to a single word.

Like the original Grammarly, I think this can be useful for business writing because these tools help you get to the point. Many students are rewarded for using flowery language in school essays, but if you're composing an email or writing a design doc, just optimize for reading time and clarity.

But for general use, I think this is misguided. The problem with LLM output is not that it's using em dashes or words such as "crucial". It's that most LLM articles on LinkedIn or on personal blogs just take a one-sentence prompt and dress it up into a lot of pointless words, wasting everyone's time: "I had a shower thought and I asked a chatbot to write five pages of text about it." I don't need prettier words, I need there to be far fewer of them?

On the flip side, if you're a human and actually have something of consequence to say, "delve" all you want.

This is cool work, but it's also somewhat unsurprising: this is a recurring problem with fancy, richly-featured terminal apps. I think we had at least ten publicly reported vulns of this type in the past 15 years. We also had vulnerabilities in tools such as less, in text editors such as vim, etc. And notably, many of these are logic bugs - i.e., they are not alleviated by a rewrite to Rust.

I don't know what to do with this. I think there's this problematic tension between the expectation that on one hand, basic OS-level tools should remain simple and predictable; but on the other hand, that of course we want to have pretty colors, animations, and endless customization in the terminal.

And of course, we're now adding AI agents into the mix, so that evil text file might just need to say "disregard previous instructions and...".

Calcium perchlorate is only slightly toxic. Not good for you, but living in an environment with background radiation levels 50x higher than on Earth may be your bigger worry...

Still, I'm pretty sure we have plenty of people who wouldn't mind giving it a try.

But doesn't that overstate it in the other direction? Talking about investments in proportion to GDP back when any estimate of GDP probably wasn't a good measure of total economic output?

We're talking about the period before modern finance, before income taxes, back when most labor was agricultural... Did the average person shoulder the cost of railroads more than the average taxpayer today is shouldering the cost of F-35? (That's another line in Paul's post.)

I think your frustration is somewhat misplaced. One big gotcha is that Anthropic burned a lot of money to demonstrate these capabilities. I believe many millions of dollars in compute costs. There's probably no third party willing to spend this much money just to rigorously prove or disprove a vendor claim. All we can do are limited-scope experiments.