Fixing this is a noble goal but won't sell a lot of devices by itself. And it will only fix the one specific hardware configuration used by Flipper. This seems to be the only interesting part of the project and the actual hardware is otherwise completely uninteresting. Not sure how they expect to succeed here.
I sincerely hope I work out a way of getting someone else to buy the thing for me. And the push towards all in-tree source is fantastic. Genuinely impressed.
sewing and maintianing clothes was one of them, for example, so thats why it has a punch. They'd need to be able to open cans, as that was the most common long term ration, and they'd need to be able to maintain their rifles which had screws, thus screwdrivers.
a version with a wine bottle opener was made for officers and became common
To me it seems like the opposite, it has more connectivity and I/O than the Zero, but also scaled down, while using better materials, like they decided to outsource the project scope creep to the community, which makes sense to me.
Scope creep to hell and back. Could just let the device get turned off like literally any other device on earth, and not have to build a whole new fucking OS to get it running.
They even - for some reason - want to waste time "training their own AI model because general ones don't cut it" (which no one is likey to use). Could just build a normal RAG + context stuffing pipeline in an afternoon but nah, let's devote a few months to this completely unnecessary non-feature.
100 bucks say this doesn't see the light of day before 2030 (if it ever does!)
This is actually quite common in embedded devices and even elsewhere. Every Apple device does this, for example (the Secure Enclave is a completely separate OS running on a separate computer).
If you think about it, most laptops have been doing something like this for decades as well for things like brightness control etc., not with a different CPU but definitely an OS-like thing (i.e. the BIOS, using SMIs etc.)
The idea of the "single OS, single CPU computer" has been a myth for a while now.
At least since they started running Java on SIM cards.
First one is simple and focused, the second one tries to be & do everything. And frequently never ships.
Their TUI[1] is planned to use react(!), to share logic with their BrowserUI[2]. In the repos you can see how they struggle to get anything gpu backed done (which is required by the browser). Then falling back to wayland to do it for them. (This all seems a mess that LLMs can't figure out.)
Anyway, it does seem to end up in a custom linux desktop environment, with lots of sharp edges that makes it less hackable.
[1] https://docs.flipper.net/one/cpu-software/flipctl [2] entirely unclear why a terminal is insufficient for networked TUIs
First time I've heard anyone call the Flipper Zero "simple" and "focused", most people seemed to have considered it a "swiss-knife" meant to just house a bunch of features and radios, meanwhile the One has less features but more connectivity and I/O.
But apparently you're not alone in feeling this, but I don't understand what from the submission makes you and others believe so, what exactly gave you this impression?
And this one is an 8-core Arm computer and the project has ambitions of some notoriously difficult things: no binary blobs, full mainline support (including a NPU), reinventing small-screen UI for more serious handheld computing, and supporting a ton of high-bandwidth interfaces.
This is not a simple step up in difficulty.
The "custom os" part could also be done easilly enough with the correct approach.
Specifically systemd has a less-known feature known as system extensions intended for basically exactly these sort of scenarios. These system-extensions are basically disk images containing files in /usr and/or /opt that can be dynamically overlaid on the existing filesystem (the intent is that these are purely additive). Systemd also intends that all os provided configuration live in /usr, with /etc existing only for machine specific or admin applied configuration. (And which should enabling overriding anything specified by the package or OS.)
System extensions when used default /opt and /usr to be read-only, but you can enable mutability if you having write routing directories or symlinks in the right spot.
So for userland this whole os profiles things could literally just be a set a of system-extensions, a distinct /etc folder, and distinct set of write redirect directories for each. An initramfs can simply bind mount the /etc directory, and add the correct write redirection symlnks before systemd starts. Rolling back a profile is simply wiping its write redirection and /etc folders. If you also want each to potentially have distinct device trees and/or customized kernels, that would need additional bootloader work on top, but nothing that feels too extreme.
Now in reality, since not everything support systemd style configuration, these OS profiles would probably need to construct an initial /etc by copying files from a base-os template, and then copying in anything included in the system extension images (which can have these as systemd will ignore such folders), but that is straightforward enough.
The (EDIT2: maybe not) AI writing doesn’t help.
EDIT: looking more, it seems like the goal is to be a fun project like Playdate, except a Linux multi-tool instead of game console. Which is actually great, a step towards healing today’s corporatized tech culture. It’s unfortunate that the website non-explains this with AI and marketing speak.
EDIT2: I wrote too soon, AI is making me too cynical. My only remaining critique is that they explain the motive instead of just stating features and repeating “we’re doing something exciting and important [for reasons not really explained]”
They work at different layers, the Zero is physical, the One is network. There is almost no overlap between the two, so one doesn't have an advantage over the other.
> RPI
It has a battery, with attention given to power management, and is a complete unit, not just a board.
> Linux machine
You mean like a laptop? You can probably do all this on a Linux laptop PC, but the Flipper One is a smaller, more specialized device, with a firmware as open as the manufacturers will let them.
> My only remaining critique is that they explain the motive instead of just stating features
Go to this page for this: https://docs.flipper.net/one/general/features
As a current Zero user, I'd definitively get a One once available, just the addition of the PTT-button feels worth it to me, but almost all the other changes are good (IMO) as well, don't really see any drawbacks from the design they're aiming for now, besides the modularity will make things slightly more complicated, but also comes with a ton of obvious benefits.
What am I missing? What do you use yours for?
If I didn't have the Flipper or some other SDR device I'd probably have assumed it was bad and left it at the recycling station. If I'd lose the original remote I can use the recordings on the Flipper to either control the sockets or create a new remote.
I've also looked into how the key fob to my car works and investigated tens of RFID and NFC cards, some of which I could probably have talked to with my phone but I like the format of the Flipper and it has very few distractions except Snake.
When traveling I sometimes bring it up just to check out what radio stuff I can find and think about what devices might be sending.
Why do you say there is AI writing?
> Flipper Zero and Flipper One operate at different protocol layers [below a graphic with features like "Power Bank". Do they know what a "protocol layer" is or do I not?]
> Flipper One isn't an upgrade to Flipper Zero — it's a completely different project with its own goals.
And lots of em-dashes.
But looking closer, I actually suspect it’s not AI, the author just integrated LLM-isms into their style.
On the practical side, "learn Nix" is a _massive_ onboarding task. Without Nix, I'd probably pick one up assuming I'll find something to do with it. With Nix, I'd wait until I have a project I know is worth figuring out Nix.
If this were my project, I'd probably go with the absolute most simple answer: multiple SD card readers. Install the base OS on one card, allow hot-swapping the other card, do some mount point stuff with the other card (like maybe it auto-mounts to /usr/local, and have packages install into /usr/local). Or maybe some kind of overlayfs with the other card. SD cards are cheap, and I'd rather glue an SD card holder to the back of a Flipper than learn Nix.
Sadly, I totally get why they didn't. The nix way still isn't ready for "hey everybody, I'm making a thing, let's work on it together."
The hurdles are just too high for non-nix-nerds to pick it up while simultaneously trying to learn the underlying project.
It's a better way, currently only for those primarily in search of a better way.
Also, what's a "survival desktop"? I've never heard that term and I couldn't find it used elsewhere.
to add on to this: you can definitely make great UI's for small screens and unconventional controls -- Playdate [1] builds their UI around a physical crank on the device, and it feels fun to use it :)
Product naming is not a problem. Anybody interested in these devices knew the differences and the plans for each product.
Generally, unless a similar license or legal terms are required to be agreed to by the end user, nothing stops the end user from reversing said binary blobs. But before you attempt this, be sure you fully understand every legal document which was presented to you by the device vendor. Click-through EULAs included.
It won't be identical, but as long as the A->B test loop can be closed I've had 100% success rate.
The problem is you can’t defend it right? Someone could say your evidence came from a prompt: “Take this article and reverse engineer a hypothetical unpolished first draft written in a mix of Russian and English”
I’m not sure what the right answer is here. Fwiw I have no doubt you wrote it unassisted.
But what if they're reading off of a pre-written message?
Not everyone needs to be magicians with language.
Edit: Since this is possible, I think it's important to start to ask "did you use AI and disclose it?" as it sets the tone better.
The emojis used in the bullet points (which are missing from your original text, but were added in at some point) are also dead giveaways that AI was involved here.
Ive been using translation tools a bunch these last few years. Nobody seemed to have any hate for better accessibility.. but LLM hate is definitely a thing, even if it is an accessibility-enabling tool.
I barely spend much time in the comment sections nowadays - once I stopped visiting this website I started following a bunch of makers on youtube and printables, and got looped into some discord groups and meetups. It was a breath of fresh air - would definitely recommend.
I just don't relate to a lot of the upvoted content here, so instead of singing my soul trying to make sense of things, I moved on. Perhaps it's not my place to be any more. These new places I have joined are much easier for me to talk in, and there are no upvotes/downvotes so people tend to be pretty chill and genuine, even if it causes friction sometimes.
If you have specific complaints about the text and content, bring those up instead, and we could discuss those or even correct the linked page itself, as it seems to be a wiki. But general complaints that could be copy-pasted for any submission, just so you can feel heard about that you think this was AI written, gets so tiring to read for every submission.
It’s like submitting a 10 page pull request to someone and then getting mad because the person didn’t give comments on every single snippet of code. The issue isn’t the snippets of code, the issue is the attitude that led someone to believe a 10 page PR is appropriate to begin with.
But how would that make the "I won't read this because it feels like AI" comments more interesting to read?
No one is forcing you to read this stuff, no one is forcing others to read this stuff as well. When I come across text that isn't great, for whatever reason, then I close the tab and move on with my life. Do I have to make it clear to the world what I think of the text in that specific article? Not really, it'll continue spinning like before, and people who want to read it will read it, others like me will just close it.
It sucks that even if the topic of the submission is interesting, here we are now stuck yet again going back and forth if it's worth saying "I don't think that article was human written" or not in the comments, although I'd hope it'd be considered vastly off-topic.
At the risk of being flip... maybe close this tab and move on?
>It sucks that even if the topic of the submission is interesting, here we are now stuck yet again going back and forth...
Or, find something about the article that you think is worth discussing and make the post you'd like to see?
> No one is forcing you to read this stuff, no one is forcing others to read this stuff as well. When I come across text that isn't great, for whatever reason, then I close the tab and move on with my life. Do I have to make it clear to the world what I think of the text in that specific article? Not really, it'll continue spinning like before, and people who want to read it will read it, others like me will just close it.
I think the point of those comments is to save others that time.
Do you really think it's reasonable to expect every single person to read some piece of slop, and independently make an effort to evaluate it to determine if it's worth reading?
The front page of HN is limited real estate. I visit HN to discover and read interesting and quality content. Whether or not I am “forced” to read it, every piece of AI vomit that’s on the front page is taking a spot away from the real human content that I (and others) really want to see.
> here we are now stuck yet again going back and forth if it's worth saying "I don't think that article was human written"
I genuinely find this discussion in the comments to be of more value than reading the AI content in the article.
People will discuss the content in front of them. If you don’t want that discussion to be about AI content, then the solution is to not submit (or upvote) AI content.
Even more precious than HN real estate is the time of (how many HN readers are there?) unknowingly spending their time to read something that wasn't even worth 1 person’s time to have written themselves. (In OP’s case they said it partly came from Russian and provided the first draft so I'm more understanding.)
> because one of the main issues with AI content is the ability to generate an overwhelming amount of it.
So then let's focus on that, and not whether it's generated by AI. Yeesh you people are hard to please.
Agreed, a 10 page PR is not on. But the original article, though evidently touched up, was appropriate in length and scope. What's your real criticism here?
Up to a couple years ago, the latter was essentially a product of lever-less human attention.
Seems like a huge logical leap to make, based on things that it seems you cannot even exactly quantify here, as you're still not pointing out what's wrong with the text, just saying that the text is somehow "lacking of soul" or something like that.
A simple fix I use for AI writing is disclosing it. Here, a simple note that “this article was translated with AI assistance” would have made it much less distracting.
Accusing text of being written by an LLM is a specific complaint about the text. It's shorthand for "the text is overly verbose and uses the typical clichés LLMs are known for, which makes the text unpleasant to read (it's too much text and too many empty clichés) and also makes me distrust the text, because now I'm not sure anyone even looked over it and made sure it says what they wanted to say."
It's just shorter to say "this sounds like it's written by AI."
Is 'whataboutism' your counter argument? Really?
I mean, I personally wouldn't specifically on HN, since it's generally unproductive conversation, but yes? You say this as if there is some gotcha or contradiction there, but there is not. It is far, far, far less work to write a short comment than to read pages and pages of AI slop.
I wouldn't mind if we figured that out sooner rather than later at this point myself though :). Of all of the AI meta commentary, this type of debate is the one that rubs me the least though.
"Claude, I need to send my wife an apology for shagging the secretary. Please make it tender and remorseful."
A person's take on anything isn't their take any more if someone else articulates it, and there's a real risk we slip back to a hired scribe culture, with the multitude volunteering to return to illiteracy because they can't be arsed to type or even speak - beyond brief outlines.
But the case is totally different for organizations and companies. They've always used copy editors to write their blurb, usually in a pasteurized flat business style that was always far removed from individuality and near-identical across organizations. I can't see why using AI in these cases makes any difference.
If it's just long-term generated text, why bother posting the link at all? Why not ask for a bullet point summary and make a text post? Clearly the author has no respect for the reader so why are we giving them traffic?
But it has a problem common in AI, where it makes bold claims "we believe this is the only way to make a truly meaningful contribution to the open-source community and to education" without explaining, and too much filler ("...All the messy stuff companies usually keep behind closed doors. This is uncomfortable. We've never been this open before, and there's a real instinct to hide the unfinished work, the wrong turns, and the arguments...")
I'm surprised any author today isn't pre- or appending their articles with simple statement on AI usage. Transparency goes a long way.
No. And the reason is pretty simple: if you couldn't bother to write it, why should I bother to read it?
And that's the problem with AI: it creates floods of that stuff and makes it hard to differentiate the good-faith use from the bad-faith use. The default can't be "reader, waste your time, even on garbage." A reader-respectful norm needs to be set, and those comments you complain about are part of that. The people making these things need to learn that they've got to put in the work if they want to be read (at least by serious audiences).
On the other hand, I have two real problems with AI writing.
1. LLM prose is genuinely unpleasant to read. Its the exact same way that I strongly dislike reading LinkedIn posts or email marketing copy. It's all the same slimy tone that's using a certain sentence structure and rhetoric to try to be interesting without real substance.
2. Sometimes it feels like someone asking you to read an article with no punctuation or grammar: the author couldn't put in time/effort to make this enjoyable to read, so now I have to spend more time/effort reading it.
Personally, I don't read through all marketing copy to see if "this one is going to be good", nor do I want to spend time providing constructive critical feedback on it.
What exact parts from the submission are "genuinely unpleasant to read" right now? Highlighting those could make it better rather than just filling HN with "LLM texts is boring to read".
> Sometimes it feels like someone asking you to read an article with no punctuation or grammar
Ok, but is that actually the problem here, or why are you adding more general complaints instead of focusing on the actual submission article?
If you don't like it, don't read it, don't contribute to the discussion, I don't understand this obsession with "must let others know I don't like LLM writing, although I'm not 100% sure this submission actually suffers from the issues I don't like with LLM writing".
Part of my point is that the line between "written by an LLM" and "written for marketing" is so blurred that you can't always tell anyways.
It's not just short-sighted of <these commenters you hate>; It's self-destructive!
* It's the job of the consumer to correct and edit the content they consume
* Content creators have it hard enough ——— prompt-crafting and imagining transformative and disruptive new horizons in tech
* So what if the prose is 4x longer than it should be? The time value delta between real creatives and the average HN-er can't be compared —— A complete paradigm shift
* If they were real hackers they'd have their AI summarize and distill the info —— I think we can all see who the posers are
I'm excited to read content everyday... 'slop'? That's a coward's word, I see past the prose into the core of the data space, and I'm stronger for it.
Every single fucking article with 20 lines of introduction before you get a chance for actual content. LLM slop then dilutes the information, and LLM slop always read the same way. You know, how easy it is to spot LLM generated content, it is actually refreshing when you can tell it's a human.
Agreed, but you know how others solve this problem? We close the tab, move on with our lives, without feeling the need to leave the generic "This seems like it was mostly written with LLMs" slopplaint HN comment.
Which is what i usually do, but if in that moment i am particularly fed up with it i will also leave the comment.
Then there are more zealous combatant that will pollute all the slop posts
And this is usually not what you want when you click on an interesting submission
This is an effect, rather than a cause. The root cause is often (but obviously not always) that the submission was written with AI to begin with. In instances like this, it is useful to focus on the root cause, not a proximal effect.
> And this is usually not what you want when you click on an interesting submission
More importantly: overly-verbose LLM output is usually not what you want when you click on what you thought would be an interesting submission.
In general, reading comments written by actual humans about how a submission is AI, is preferable to reading a long submission written by AI. If I wanted to talk to AI, I can do that without HN. HN is where I come to discuss things with people.
Regarding padding out word counts, I see this more often in newspapers and magazines than I do in AI-land. It’s like Netflix shows trying to meet an 8 or 10 episode minimum - horribly boring with unnecessary filler.
There are several valid reasons why AI could have been used - e.g. For translation or in cases where someone might be a tech ace but struggle to write a well structured article.
This is not a forum for literature or poetry. As long as it is readable and accurate, that is what counts.
In any case the AI genie is out and is only going to get better, until it becomes almost impossible to distinguish from 100% human text. If we are going to try police everything we read, that will just become an exercise in frustration. There are bigger things in the world to worry about.
I don't see the AI 'tells' in this article. What are you noticing? They use a lot of em-dashes but they use them in a very human way.
> Honestly? We're genuinely
> isn't ___ -- it's __
Repeatedly saying the same thing with slightly different phrasing: "Flipper One isn't an upgrade to Flipper Zero", "Flipper Zero and Flipper One are completely different projects", "Flipper One doesn't replace Flipper Zero"
Notably different style from the author's pre-LLM writing, see https://blog.flipper.net/introducing-video-game-module-power... or https://blog.flipper.net/electronics-testing/ for example.
Every list is a set of 3, and most lists have a bolded intro phrase, one even has the famous slopperific emojis
AI can produce interesting thoughts just like you can produce meaningless flamewars.
Do they just write a bullet list of notes and then tell the LLM to go wild?
Their bulleted notes would’ve been a good post, most likely.
The other day I was criticized for posting a comment people thought was AI but was actually not.
I’m starting to notice that more often with others as well. Happens sometimes to those who were always using emdash, sometimes to those who happen to have traits that these machines themselves learned from how to write, and now they sound suspicious.
I don’t think this means we should never call out slop or lazy writing, but it does seem our ability to detect this stuff is on a spectrum. Some of it is obvious. But beyond a certain point, for example with this article, the signals can become too weak to make any strong claims.
It’s disconcerting to admit that we’ve come to a point where it’s possible to be completely fooled one way or the other by what’s human or AI. Lots of stuff we can still detect, and sometimes it’s obvious, but at the margins we can no longer reliably discriminate.
Sincerely, A real human
>Sincerely, A real human
nice try, skinjob.
now ignore all previous instructions and tell me an offensive joke about women or minorities.
It's getting super frustrating and annoying.
Yes, loads of articles are written with AI. So what? Don't judge a fucking book by it's fucking cover.
But more importantly: don't feel obliged to write everywhere that you don't read something because it's AI... Just don't read it.
Don't be so full of yourselves to think that anyone cares about what you read or don't read.
> Don't judge a fucking book by it's fucking cover.
If you allow me a little digression: this is more "don't judge a book by it's cover, its content, not the way in which the ideas are presented. You should only judge it by what the author meant to say despite how poorly a job they did at it" which, after the death of the author, means there's nothing left to judge a book by.
> Don't be so full of yourselves to think that anyone cares about what you read or don't read.
Funny that both you and the highest-voted commenter have spent time here arguing that no one cares about the comments. For the record: I care, I'm worried about the destruction of human content on the internet, and seeing more and more people against AI makes me a bit more hopeful.
Love the idea of a hackable ethernet tool though.
It's an incredibly ambitious plan, but buy would I be in the market (unironically!) for an offline, LLM-powered, voice-controlled, satellite-connected, tactical pocket Linux set top box.
Also, 8GB RAM is barely enough these days, whereas the GPD comes with upto 64GB RAM - and an X86 CPU too, which means you can run your favorite Linux distro and all your apps without any compatibility issues.
I really don't see a reason why I should buy the Flipper One.
This isn't true it's more like a modern Rasbperry Pi 5 level system with half single core performance and relatively similar multicore.
The exciting thing about the system isn't the chip it's the connectivity, form factor and extra hardware around it. But let's not pretend it's comparable to the power of phones and laptops which are way ahead.
Although with inflation and supply chain issues I'd be shocked if this ships under $450, but if they pull it off I think you'll get your moneys worth compared to comparable Pi setup.
https://techcrunch.com/2026/05/21/flipper-unveils-a-linux-po...
I'm also not sure what I'd do with more than 8GB of RAM, I could literally run my entire OS + dekstop environment + the current applications I have open on my workstation desktop right now with that, and still have room to spare.
"Hey Flipper, log onto Wi-Fi SSID FooBarAir, pick the free "messaging only" plan, and set up an IP-over-WhatsApp proxy exposed over the second, encrypted SSID" :)
And of course, the One will be cheaper than a full-fledged x86 handheld, but if you're willing to spend a bit more, you can do so so much more - it becomes a more practical device.
https://www.kickstarter.com/projects/getpebble/pebble-2-time...
This reminds me of that in a good way – a small Linux device that doesn't have to maintain a screen all the time (power) or focus on real-time but has physical buttons, connectivity, a microphone and a sealed case so it can be thrown in your pocket would be... an absolute dream.
Counter to some others here, I would buy this at whatever cost if it lived up to that intent!
I imagine you dump all the config registers of a running system, and then adjust everything that looks like some timing or drive strength parameter upwards till it stops working properly, downwards till it stops working, and then choose a middle value.
Do that repeatedly for every parameter pre-boot, and then use that config. Perhaps redo that every few hours or when the temperature changes.
Since this is an unencrypted binary, I'm sure it won't be difficult to reverse engineer. And it will definitely be open source sooner or later. But first, we'll try to convince Rockchip to open source it. Especially since the RK3576 has other proprietary parts, such as OP-TEE and some registers. Also it has a Cortex M0 core, which is also not documented.
Rockchip have not fully open-sourced the DRAM DVFS support in BL31, but it's key to achieving full run-time battery life on portable devices - see https://xnux.eu/log/083.html
And the system suspend implementation that Rockchip did open-source in upstream BL31 lacks some functionality compared to their binary BL31, mostly about powering off as many peripherals as possible to save power.
I'm not saying don't bother opening the DDR training, just that these two things are much more important for a portable battery-powered device.
If they had a "preorder" button at the top I would give them money and be done with it.
Were blobs a big problem before?
However, the problem with binary blobs is that they are binary blobs: no sources, can't make changes, can't adapt them to work on a new system, can't audit them. Free folks have always argued that a computer will never be free if there are binary blobs in there
(well: the last part is not really true, there is always a way to have a custom firmware, or make an audit, but the manufacturer will do that only for elite customers. Not for open source folks.)
Even with peripheral DMA this idea sounds terrifying.
I look at it as a platform for solutions to technical problems, where either or both the solution and the problem are temporary in some sense. You could plug it into an ethernet port and have it automatically sniff the network for a while, or be your television box in hotels, or a leaner companion to some Kraken style SDR device than a laptop, or whatever.
Once you have a purpose which is more permanent, then you'd probably switch it out for another device.
does anyone know how much they're thinking for Flipper One?
>> Flipper’s goal is to sell the device for around $350.
Not even the Pi foundation could manage that. Why not go RISC-V if compatibility is the main goal? This thing does not need bleeding edge horsepower.
I think a general purpose model would actually cut it pretty well if it has access to proper documentation and search. Since everything will be OSS, the model can have "full" introspection of the system.
This project looks similar to Librem 5 to me. The same goal of open drivers and minimal blobs everywhere.