regardless of whatever else you think about AI, it’s a good idea to be very wary about security-sensitive apps that are written using LLMs.

there’s a self-hostable S3 replacement called RustFS, with development that leans heavily on LLMs.

they had a security vulnerabilty last year.

one LLM-written commit added a complete authentication bypass to the code. literally, you could just send rustfs rpc as the auth token instead of an actual token.

it was fixed in a pull request named fix: Prevent panic in GetMetrics gRPC handler on invalid input which claimed to fix a different auth problem and just coincidentally removed the hardcoded auth token.

security stuff is hard, and LLMs are prone to over-confidence. that’s a bad combo.

this shows all the hallmarks of being vibe-coded slop (emoji-studded readme being the first dead giveaway)

it’s “open-source” but https://www.neatmail.app/ has a Pricing tab

the self-hosting instructions mention a “DodoPay account (payment processing)” with no explanation of what payments you’d be processing if you’re self-hosting it.

and one of the listed “AI Integrations” is:

In-House Model: neatmail_model — our proprietary classification model built and maintained in-house (still under work)

which means, unless proven otherwise, you should assume that this is feeding the entire contents of your incoming emails not just to OpenAI, but also to this “Neatmail” company for processing with their “proprietary” model.

hard pass.

For the past month or so, I’ve been getting “RDSEED32 is broken” and it seems to be an issue with AMD’s drivers?

https://www.amd.com/en/resources/product-security/bulletin/amd-sb-7055.html

it sounds like the kernel is just working around a known CPU microcode bug. it would probably be using the 64-bit RDSEED operation anyway, so disabling the 32-bit option probably doesn’t actually change anything.

also, the kernel’s random number generator is very robust (especially since Jason Donenfeld, the author of Wireguard, took over its maintenance) and will work perfectly fine even in the complete absence of RDSEED CPU instructions.

“Hey” is an email thingy run by the company that DHH owns.

the rest of those “apps” are probably thrown in to make the list seem more complete. the real goal is to promote his paid email service.

best of luck to his replacement, Greg Yoshi

kindergarteners through second graders clamor for Green Eggs and Ham after binge-watching the 2019 Netflix series of the same title

I don’t have kids, and most of my friends don’t, so I’d never heard about this…

the book is 65 pages long according to Amazon.

the series has 23 half-hour episodes.

moving on from Dr. Seuss seems totally reasonable to me - but “Netflix adapted a 60-page picture book into 12 hours of TV content” shows why it seems unlikely.

everything in media is more or less required to be a franchise now. if you want a book to read to your kid that’s as good or better than The Cat in the Hat, I’m sure there are plenty. the article gives several possibilities. but as far as media executives are concerned, they don’t want just a book - it needs to be a whole-ass ecosystem.

they’re making a Cat in the Hat movie. they already made one in 2003. it had a budget of $109 million, but only made $134 million at the box office…which means it’s considered a failure. I suspect the bigger problem, besides making only $25 million in profit, is that it was so bad that Seuss’s widow prohibited them from developing a sequel.

so they’re making another one, this one entirely animated. I was going to make a joke about a “Dr. Seuss Cinematic Universe” but then I scrolled down in that Wikipedia page and I read:

In October 2020, it was announced that the film intends to launch a shared universe with two other animated films. The first, Thing One and Thing Two, is expected to be released in 2026. The other, an adaptation of another Dr. Seuss book Oh, the Places You’ll Go!, is scheduled to be released on March 17, 2028. In April 2025, the name of the shared universe was revealed to be “The Seussiverse”.

a cheap docking station with two SATA slots (currently housing hard disks) and putting them together on a RAID0 almost doubles a single one’s performance.

you can buy a 50cc moped and attach a NOS cylinder to it. that might be a fun hobby project, if you’re into it.

but in a drag race, you’re going to get beat by a 10 year old Toyota Prius. because there’s only so much you can eke out of a 50cc engine.

“RAID0 using a cheap 2-slot external enclosure” is one of the more cursed things I’ve ever contemplated. firmly in “just because you can doesn’t mean you should” territory.

short answer: buy NVMe. plug it directly into your motherboard, don’t use an enclosure. forget about wonky RAID0 crap.

longer answer:

SATA SSDs (which you say in the comments below are all you’ve got) are an evolutionary dead-end. they’re SSDs pretending to be very fast hard drives. they end up being bottlenecked by the assumptions that the SATA protocol makes about how fast a hard drive can be.

look at this chart for example. SATA (AHCI) limits a device to having 32 commands queued up at once, which means the operating system needs to jump through hoops in terms of maintaining its own queue of pending reads & writes and issuing them to the device as queue space becomes available.

NVMe raises that limit to 64k, which for any non-server workload is effectively unlimited. the NVMe drive can respond to IO requests pretty much as quickly as the OS can dispatch them.

if you want to know more nitty-gritty details, Scaling ZFS for NVMe is an interesting talk, much of it isn’t specific to ZFS, but instead is about how NVMe devices are so fast that they’re forcing filesystem developers to rethink long-standing assumptions about drives being slow.

I am basing that on both what I see on the news and what is happening to people all around me.

what news sources are you consuming?

because if you’re getting the message from the news that economic collapse is imminent and all currencies are going to be worthless and we will need to fall back to a barter-based economy…that is a function of choices you’ve made in your news diet, much more than it has anything to do with anything actually happening in the real world.

and what specifically is happening to people around you that you’re referring to? do you have a pen-pal in Weimar-era Germany who you’re communicating with through a time portal? or are you talking with other people who have the same news diet as you do and forming a self-reinforcing worldview?

at the risk of being “guy who pretty much only plays Factorio recommends you play Factorio…”

you can easily put 50+ hours into a single savefile, especially with the Space Age expansion

Encryption lengths are getting long so you’d think it was high time.

that’s unrelated - AES-256 for example can be executed just fine on either a 32- or 64-bit machine. in theory there’s nothing stopping you from running it on an 8-bit or 16-bit CPU (although other considerations related to the size of AES’s lookup tables make this unlikely). from some random googling, here is an implementation of Chacha20, another 256-bit encryption algorithm, for 8-bit microcontrollers.

when we talk about 32 vs 64-bit CPUs, in general we’re only talking about the address space - the size of a pointer determines how much RAM the computer is able to use. 32-bit machines were typically limited to 4GB (though PAE helped kick that can down the road)

CPU registers can also be sized independently of the address space - for example AVX-512 CPUs have a register that is 512 bits wide even though the CPU is still “64-bit”.

it might be more complicated than you’re looking for (requires a self-hosted server instead of just a desktop app), but take a look at the ecosystem surrounding Subsonic

Subsonic did some licensing shenanigans, but there’s an actively-maintained GPL3 fork called airsonic-advanced

there’s also alternate implementations, Gonic and Navidrome, that maintain compatibility with the original Subsonic API

because they all work with a common API, there’s a variety of clients that can work with the backend.

I’m also a big fan of Beets for music organization, it’s not tied in to the Subsonic ecosystem so you can use them completely separately if you want. it handles tagging, can fetch lyrics, and can also transcode the library (or an arbitrary subset of it) if you want to send it to a portable device. (not sure if this is what you mean by compatibility)

I currently have Beets organizing everything, run Navidrome on my server pointed at the Beets library directory, then Ultrasonic on my phone, and the Navidrome web interface on my desktop. the combo is especially nice for streaming to my phone - Navidrome will transcode FLAC to Opus on the fly, and Ultrasonic has an option to cache those files locally, and to pre-download them over wifi instead of mobile data. so I have my full collection available on my phone, can stream it from anywhere, and the songs I listen to frequently are already downloaded and I don’t have to waste mobile data, or wait for them to load if I have poor cell signal.