03:03:45
glowingexplorer:matrix.org:
@glowingexplorer:matrix.org: i have the perfect word. PEDOPHILES. it has the same level of accuracy and respect as all of the derogatory AI words, and its co-opting an existing word like they are. absolutely perfect. you guys are all pedophiles.
03:14:14
ofrnxmr:xmr.mx:
imagine calling yourselt a programmer with nothing but imaginary code written by copilot on autopilot mode
03:15:54
glowingexplorer:matrix.org:
@ofrnxmr:xmr.mx: that sounds better than being a pedophile, you fucking pedophile
03:16:10
ofrnxmr:xmr.mx:
https://mrelay.p2pool.observer/m/xmr.mx/qhNOKukJJtmxHwostPnaVeGL.jpg (17698294020285397090792835343207.jpg)
06:31:34
DataHoarder:
https://www.nro.gov/news-media-featured-stories/news-media-archive/News-Article/Article/4392223/declassifying-jumpseat-an-american-pioneer-in-space/
07:06:30
DataHoarder:
wait, according to glowingexplorer, is now everyone calling AI slop "vibecoded" a pedophile?
07:06:42
DataHoarder:
sounds like projection
07:07:06
DataHoarder:
also: > <sech1> oh no, AI slop has hit xmrig https://github.com/xmrig/xmrig/pull/3771 :D
07:07:38
DataHoarder:
> unlocks the 0% donation level.
09:50:04
glowingexplorer:matrix.org:
no. pedophile is a new term i coined for a non-coder who criticizes code with some form of AI accusation. so many people here are pedophiles, yes. > <DataHoarder> wait, according to glowingexplorer, is now everyone calling AI slop "vibecoded" a pedophile?
09:50:36
DataHoarder:
you sound salty indeed this https://mrelay.p2pool.observer/m/xmr.mx/qhNOKukJJtmxHwostPnaVeGL.jpg
09:51:34
glowingexplorer:matrix.org:
sure. just a black person would be salty if the N word were casually thrown around in here. you guys are insulting me, whether you intend to or not..
09:54:32
plowsof:
https://mrelay.p2pool.observer/m/xmr.mx/qhNOKukJJtmxHwostPnaVeGL.jpg
09:56:43
plowsof:
😂
09:58:40
plowsof:
I can cite that as a ban reason
09:59:00
glowingexplorer:matrix.org:
i think i must be fortunate i have no idea what that image means
10:00:27
plowsof:
The slopper has to at a bare minimum be able to interpret that image, see thats how i know your vibe coded slop is garbage 😂
10:02:12
glowingexplorer:matrix.org:
just go ahead and get it over with. i'm not going to stop sticking up for myself and every programmer. and if you're going to ban me for being the only person saying that against an army of PEDOPHILES then get on it with it motherfucker
10:02:39
plowsof:
https://mrelay.p2pool.observer/m/xmr.mx/qhNOKukJJtmxHwostPnaVeGL.jpg
10:02:54
plowsof:
LMAO
10:07:03
btcdwed:
hai
10:07:15
btcdwed:
nioc's cat seems to be very happy
10:12:44
glowingexplorer:matrix.org:
AI is a tool. A tool that can be used to create useless garbage, no doubt. But in the right hands, it can also be extremely useful. It mostly depends on the skill of the person using it. The problem is you guys slander everybody who uses AI. That's fucking absurd. Not everybody using AI to help them code is a vibe coding moron [... too long, see https://mrelay.p2pool.observer/e/-Jeq6OEKeEh3TTMy ]
10:48:25
kayabanerve:matrix.org:
I've so far seen it only produce garbage. Shall we call 'vibe coders' slop eaters? In the right hands, it can produce slop at a rate never seen before?
10:48:25
kayabanerve:matrix.org:
My actual comment is that it's never to be trusted, but review is a more notable task than development, so it can never save time unless trusted.
10:48:46
kayabanerve:matrix.org:
Maybe it can churn something out that appears to work, and you don't care. The issue is when it's then presented at something which works.
10:49:20
kayabanerve:matrix.org:
The complete disregard of the requirement to ensure your work, because you don't have to. The LLM did
10:52:06
kayabanerve:matrix.org:
Junior devs who don't understand what they're doing can absolutely use LLMs to be more productive though and believe they're more capable than they are though
10:53:26
sech1:
They can be more productive until there is a bug and AI slop can't fix it without breaking something else, and no one knows the code, even the dev. Who will maintain that code then?
10:53:54
kayabanerve:matrix.org:
More productive at producing slop is still more productive.
10:54:14
kayabanerve:matrix.org:
I didn't say more contributive to society
11:05:56
ofrnxmr:xmr.mx:
The shoe fits > <@glowingexplorer:matrix.org> sure. just a black person would be salty if the N word were casually thrown around in here. you guys are insulting me, whether you intend to or not..
11:14:43
DataHoarder:
> sticking to every programmer
11:14:57
DataHoarder:
We already went over this and got actual programmers calling your bullshit what it is
11:20:31
DataHoarder:
Like https://www.anthropic.com/research/AI-assistance-coding-skills
11:20:43
DataHoarder:
Literally using it reduces your skill level
11:22:30
DataHoarder:
> We find that AI use impairs conceptual understanding, code reading, and debugging abilities, without delivering significant efficiency gains on average.
11:23:00
DataHoarder:
https://arxiv.org/abs/2601.20245
11:23:28
DataHoarder:
> Our findings suggest that AI-enhanced productivity is not a shortcut to competence and AI assistance should be carefully adopted into workflows to preserve skill formation -- particularly in safety-critical domains.
11:23:45
DataHoarder:
Safety critical aka Monero :)
11:24:41
BlueyHealer:
Or cryptography in general
11:25:17
BlueyHealer:
Like, from what I understand, even just correctly implementing existing libraries is a minefield for mistakes.
11:25:19
DataHoarder:
You know it's bad even when the companies selling you that find and report how damaging it is to skilled users (and how doesn't build up experience or skills for new people, where they show manually debugging as an example of gaining skills)
11:25:58
DataHoarder:
On the paper thy asked a group to not use AI, yet some did, so they asked them again to not use it, and they were unable to do anything after it was a task they were able to do months ago
11:26:15
DataHoarder:
And some STILL used it in secret as they could no longer accomplish the task
11:40:01
kayabanerve:matrix.org:
I am insulting people who willingly choose to not develop their skills and instead defer their lives to a magic box, correct.
11:40:41
kayabanerve:matrix.org:
It goes a step past simply being lazy into being unethical, presenting someone else's work as your own.
11:41:10
kayabanerve:matrix.org:
Whether that work is the work trained upon or the LLM's develop can be debated. The fact you pasted a GH issue link into an LLM doesn't make it yours.
11:42:20
kayabanerve:matrix.org:
Inb4 "you have to accurately prompt it and guide it"
11:42:20
kayabanerve:matrix.org:
If a non-dev project manager started PRs, I'd be identically frustrated with the corporate sockpuppet in front of me claiming to be the author of code they didn't write and didn't fully consider. At least the code itself would be better though.
11:43:45
kayabanerve:matrix.org:
There's one project I saw recently which no now longer accepts PRs and solely prompts as GH issues lmao
11:45:49
kayabanerve:matrix.org:
I'm sure LLMs can produce better code than someone who doesn't know how to code.
11:45:49
kayabanerve:matrix.org:
I don't want someone who doesn't know how to code submitting PRs. I definitely don't want them submitting PRs which pass a surface glance and further waste my time.
11:48:14
DataHoarder:
> 12:43:15 <br-m> <kayabanerve:matrix.org> Inb4 "you have to accurately prompt it and guide it"
11:48:45
DataHoarder:
This was their point a couple days ago and why they started "you are insulting me by calling it vibecoded"
11:50:40
kayabanerve:matrix.org:
project manager claims to be a dev and gets roasted for not being a dev
12:01:03
DataHoarder:
Also the classic "you are not a dev so your words do not matter" and "every programmer uses it"
12:16:12
kayabanerve:matrix.org:
Ugh, I think there was a horrific statistic that 70% of StackOverflow survey participants identified as using an LLM at least once a month or so, and ~55% at least once a week. It didn't elucidate further on how or where though.
12:18:15
ravfx:xmr.mx:
Thanks for the links related to ai brainrot studies
12:20:29
Cindy:
did that guy really had to call us pedophiles just because we don't like slop?
12:21:02
kayabanerve:matrix.org:
most balanced and reasonable take I've seen from an llmk bro tbf
12:31:16
kaigoh:gohegan.uk:
Probably used AI for that, too!
13:03:29
ravfx:xmr.mx:
https://hackread.com/us-seizes-400m-helix-dark-web-crypto-mixer/
13:07:33
BlueyHealer:
"you need to guide and prompt it" - imagine someone commissioning art and after a long arguing in chat, proudly proclaims "I made this drawing"
16:47:39
basses:matrix.org:
https://www.phoronix.com/news/KDE-Brings-Back-Air-Theme
18:17:48
basses:matrix.org:
https://an.dywa.ng/carrier-gnss.html
18:30:47
msg:
basses: wow, what about grapheneOS
18:30:54
DataHoarder:
it's part of the chip
18:31:17
DataHoarder:
> The feature is only available to devices with Apple’s in-house modem introduced in 2025
18:31:20
DataHoarder:
which apple makes
18:31:42
msg:
:(
18:31:49
DataHoarder:
https://support.apple.com/en-us/126101
18:31:50
msg:
so that means grapheneOS sends location data to the carrier
18:31:51
msg:
all the time?
18:32:03
DataHoarder:
imagine a feature that's only available on cheap iphone and not pros :P
18:32:19
DataHoarder:
msg: it's again not the OS, but the chip
18:32:19
msg:
smh smh
18:32:22
DataHoarder:
graphneneOS doesn't even know
18:32:24
msg:
DataHoarder: yeah i know
18:32:30
msg:
so that means like the pixels do it themselves
18:32:38
msg:
grapheneOS can't control it right?
18:32:51
DataHoarder:
the broadcom chip most likely
18:32:53
DataHoarder:
it's what apple switched away from too
18:33:00
gan:skhron.org:
it's on the chip, independent of the OS > <msg> so that means grapheneOS sends location data to the carrier
18:33:24
DataHoarder:
like. not even Google can
18:33:33
msg:
but it's pretty sad
18:33:54
DataHoarder:
Apple can cause they designed something in-house they fully control
18:34:37
msg:
yeah that seems about right
18:35:12
gan:skhron.org:
nothing ever changes to be honest, we had stuff like Intel ME and AMD PSP since ages on desktop, furthermore it's just another vector, even without it, it'll be possible to get the location via triangulation, or internal MAC-based maps services
18:35:35
DataHoarder:
from the POV of carriers not Mac-based
18:35:45
gan:skhron.org:
people already forgot that Apple used to leak location that they "used" to store based off MACs
18:35:51
DataHoarder:
triangulation is enhanced with the direction finding of modern cell nodes
18:35:57
msg:
so that means any service that works with phone numbers is a massive privacy problem
18:36:01
DataHoarder:
that it's used for higher data rates and selective bandwidth
18:36:01
gan:skhron.org:
DataHoarder: yeah, I just mentioned another vectors
18:36:04
msg:
because if you have the phone number, and connections of that carrier
18:36:09
msg:
you can figure out where they are at all times
18:36:19
msg:
this applies to shit like Signal and Telegram
18:36:23
DataHoarder:
msg: it always was even without all of this
18:36:25
DataHoarder:
phone numbers can be taken over
18:36:29
msg:
yes thats true
18:36:43
msg:
phone numbers are not even good long-term identifiers
18:36:57
msg:
if your carrier is garbage, they can let someone else take it
18:37:21
gan:skhron.org:
all phones have connections regardless of the carrier, you never wondered how it's possible to make emergency calls without SMS? > <msg> because if you have the phone number, and connections of that carrier
18:37:58
gan:skhron.org:
s/SMS/GSM/
18:38:11
gan:skhron.org:
I also believe that national alerts are also sent without a need of SIM
18:38:25
gan:skhron.org:
@gan:skhron.org: what the fuck, why my brain keep farting
18:38:38
gan:skhron.org:
I meant SIM in both instances
18:39:43
albertlarsan68:albertlarsan.fr:
IIRC it is as part of the 4G or 5G protocols that the GPS position is sent to the relay antennas, to make cell handover work "better"
18:40:23
BlueyHealer:
gan, from what I understsand though, IME/PSP are not really passive surveillance and more like an increased attack surface.
18:40:49
msg:
also btw, what do people think of my new nick :D
18:41:07
BlueyHealer:
Who were you before? Have the nick change messages turned off.
18:42:01
msg:
Cindy
18:42:06
BlueyHealer:
msg | so that means any service that works with phone numbers is a massive privacy problem <- That's NOT the biggest problem, though. The biggest being the official inability to get a number without KYC,
18:42:08
BlueyHealer:
Ah!
18:42:15
BlueyHealer:
Guessed right)
18:43:42
basses:matrix.org:
they didn't comment about it yet afaik > <msg> basses: wow, what about grapheneOS
18:44:16
gan:skhron.org:
Sure, imagine sending all of that traffic to the nearest multiple letter storage facilities, they'd fucking choke and die > <BlueyHealer> gan, from what I understsand though, IME/PSP are not really passive surveillance and more like an increased attack surface.
18:44:16
basses:matrix.org:
you can always run airplane mode which sends nothing to cell towers (wifi will work fine) > <msg> so that means grapheneOS sends location data to the carrier
18:44:48
gan:skhron.org:
Nonetheless, it allows full remote access, and has been exploited in various ways - https://en.wikipedia.org/wiki/Intel_Management_Engine#Security_vulnerabilities
18:45:32
BlueyHealer:
TBH didn't know about the GPS thing, to my shame. So this means the satellite polling is on all the time? Then what does the "GPS" switch in the settings do? There does seem to be a difference in battery usage...
18:46:30
BlueyHealer:
<gan:skhron.org> Sure, imagine sending all of that traffic to the nearest multiple letter storage facilities, they'd fucking choke and die <- Eh, don't think that's an argument. What exact "all of this traffic"? I was thinking more about how there isn't evidence of this in general.
18:47:32
albertlarsan68:albertlarsan.fr:
It is not the same level of precision I would imagine. Also, it prevents apps from requesting position data.
18:48:22
BlueyHealer:
gan, yeah, seen that! So this strengthens the impression of "increasing the attack surface rather than passive data collection". Not to diminish the importance of that.
18:48:40
msg:
what if we laser etch a bit of the modem chip
18:48:43
gan:skhron.org:
BlueyHealer: Intel ME have direct access to a NIC, so they could send all traffic at will, to note, people already exploited Intel ME to extract data - https://en.wikipedia.org/wiki/Intel_Management_Engine#PLATINUM
18:50:39
BlueyHealer:
That's on thr same page you just linked, yeah) That says it was a dedicated attack.
18:50:40
msg:
gan: does Intel ME have access to the system's RAM
18:50:57
gan:skhron.org:
BlueyHealer: I see that in the same vein as Signalling System No. 7, which was heavily abused in Germany
18:51:21
gan:skhron.org:
msg: it have full access to pretty much everything
18:51:54
BlueyHealer:
Isn't that what's used pretty much everywhere? That's specifically a communication protocol, that's different.
18:53:28
gan:skhron.org:
BlueyHealer: Yeah, it's different in that sense, but computers are all pretty much connected to the global internet, which makes stuff like Intel ME concerning
18:58:37
msg:
gan: what if you don't use the NIC
18:58:41
BlueyHealer:
It is, in the sense of increasing the attack surface and being a powerful target.
18:58:58
BlueyHealer:
I am just not liking how you compare the two because they have nothing in common fundamentally though
18:59:02
msg:
seriously what if you just use a WiFi card, or an ethernet adaptor with a propietary interface
18:59:15
msg:
that only the OS can interact with
19:02:35
msg:
the stuff i've seen about Intel ME makes it seem like it only works on wired LAN adaptors
19:02:55
msg:
and doesn't have the capacity (or drivers) to interact with wireless cards
19:06:53
BlueyHealer:
msg, it does? I assumed it could use the same networking the normal OS does. But yeah, never thought about the variety of drivers... Thanks for thought fuel, I'll have to check more in-depth.
19:08:07
msg:
https://security.stackexchange.com/a/154877
19:09:15
gan:skhron.org:
ME have full access to PCI interfaces as well > <msg> seriously what if you just use a WiFi card, or an ethernet adaptor with a propietary interface
19:09:51
msg:
okay but can it actually know how to communicate with the wifi card?
19:10:13
msg:
or understand anything it says?
19:12:36
gan:skhron.org:
msg: It runs a UNIX-like system, so that only a question of time to implement something so it'd understand, and redirection is much simpler to implement than trying to do it on the chip's OS
19:14:02
msg:
do you think they'd bother including drivers for every single wireless card they could find?
19:14:30
msg:
and what if there was a new card (with its own communication protocol and drivers) made AFTER the intel ME firmware
19:16:20
gan:skhron.org:
msg: That's not what I've said, furthermore nearly all cards have a common interface, as I've said it'll be simpler to do redirection
19:17:00
BlueyHealer:
Redirection to where?
19:17:18
msg:
redirection?
19:19:07
gan:skhron.org:
BlueyHealer: To any point that's under the control of a person who's extracting
19:20:25
msg:
a person who's extracting?
19:20:28
gan:skhron.org:
msg: Almost any computer interface at such level could be hijacked to send all the data somewhere, for a remotely-relevant example - USB/IP
19:21:04
msg:
soo.. that'd require physical access to the machine
19:21:13
gan:skhron.org:
msg: the attacker
19:21:35
gan:skhron.org:
msg: Not necessarily
19:22:18
msg:
also USB/IP requires a TCP/IP stack.. which requires a network device you can properly communicate over
19:22:36
msg:
where can you actually redirect the data
19:23:10
gan:skhron.org:
ME could just monitor traffic in full?
19:23:24
msg:
again, how can it do that when it doesn't know how to communicate with the wireless chip
19:23:42
gan:skhron.org:
I just offered it as an example since it's showcasing that it's possible to send something like USB over the network
19:23:44
msg:
we're assuming a computer whose NIC is not connected, and primarily communicates over a proprietary wireless card
19:28:37
BlueyHealer:
I don't understand... After it has intercepted the traffic that was going to be sent via a wi-fi card, how does it tell the card to send it out?
19:30:08
gan:skhron.org:
BlueyHealer: That an interesting question and the only flaw that I could see here
19:30:42
gan:skhron.org:
It'll be difficult for an outsider to implement a driver for Minix in my view
19:32:59
BlueyHealer:
Like, the OS needs a whole up-to-date collection of drivers for this... I hope I don't forget to search that up because I never thought of this.
20:05:57
msg:
gan: that's what i was talking about :o
20:07:26
gan:skhron.org:
yeah, PCI redirectors themselves aren't hard to implement, but ME would need to send it somewhere, and since there's no NIC, I'm not so sure how it'd do that
20:08:16
msg:
also btw
20:08:24
gan:skhron.org:
I've heard that NIC drivers are more generic in their implementations than say WIFI cards, although they also follow a standard
20:08:38
msg:
yeah NICs are more generic
20:08:47
msg:
but WiFi cards are.. not that generic
20:08:52
gan:skhron.org:
With latter area, I'm not familiar to have an informed opinion
20:11:10
msg:
do we really need to lobotomize intel ME?
20:11:21
msg:
when we can just.. use a network adaptor it doesn't know?
20:12:08
gan:skhron.org:
msg: that's still preferable, because you can't know if a person exploited it
20:12:39
gan:skhron.org:
à la MSI warehouse BIOS fiasco
20:13:48
gan:skhron.org:
totally unsigned MEs has been spotted on some hardware - https://libreboot.org/docs/install/t480.html#absolutely-unsigned-intel-me
20:16:24
msg:
that's.. worse
20:17:27
gan:skhron.org:
Generally think of it like that anyone could install openwrt on almost any router, alongside spyware
20:17:55
gan:skhron.org:
People are too focused on the state, but sometimes it's just skiddies
20:18:11
gan:skhron.org:
Although there's no replacement for ME yet
20:18:34
msg:
how do RISC-V CPUs compare?
20:19:53
gan:skhron.org:
That depends on the vendor to be honest, since they're royalty-free, but I'm not aware of ongoing spyware-aligned ISA or firmware bundles that would include any such spyware
20:20:32
gan:skhron.org:
even ARM doesn't have anything like PSP or ME, but they could have funny stuff in their mask
20:21:25
gan:skhron.org:
although, again, I'm not aware of any malicious backdoor-looking bullshit on ARM CPUs
20:21:31
msg:
hmmm
20:21:54
msg:
yeah that's true, i haven't eithe
20:21:55
msg:
either*
20:22:43
gan:skhron.org:
I prefer to judge based off history and practicality
20:23:37
gan:skhron.org:
I anyway separate "freer" hardware in my use-case, so in my threat model, I'm not too concern with it
20:24:40
msg:
gan: if ARM had something similar
20:25:06
msg:
where would it store the firmware for its little spyware component
20:25:23
msg:
in Intel ME, the BIOS has to spoonfeed the ME the firmware
20:46:22
basses:matrix.org:
https://help.apple.com/pdf/security/en_US/apple-platform-security-guide.pdf
20:46:26
basses:matrix.org:
https://mrelay.p2pool.observer/m/matrix.org/DdqJVRlQvcezbMLbCzRCBAjc.png (clipboard.png)
20:46:29
basses:matrix.org:
M5 chip supports memory tagging
20:47:02
basses:matrix.org:
MIE
20:53:55
nioc:
https://rimgo.catsarch.com/nheIBGt.png
21:49:52
DataHoarder:
> So this means the satellite polling is on all the time?
21:49:52
DataHoarder:
it is a passive element always receiving signals
21:50:07
DataHoarder:
it also syncs the clock
22:01:43
AlbertLarsan68:
I’m pretty sure clock syncing also uses the cell tower (at least to determine the timezone)
22:10:14
DataHoarder:
yep
22:10:20
DataHoarder:
multiple sources