lynxy's recent activity
-
Comment on WiFi 5 beamforming is able to infer the identity of individuals without a WiFi device on them through passively recording communication in radio networks in ~tech
-
Comment on Anyone playing Subnautica 2? in ~games
lynxy LinkI have maybe 18 hours in it so far! I'm quite enjoying it, though it's most certainly early access and there are a few things that I'm missing that were in the first game. I have not played...I have maybe 18 hours in it so far! I'm quite enjoying it, though it's most certainly early access and there are a few things that I'm missing that were in the first game. I have not played through Below Zero.
Where it is now, it feels like Subnautica 1, but with updated graphics and on another planet, with another story to uncover through the exact same game-play as in the first. Not that I'm complaining.
-
Comment on What programming/technical projects have you been working on? in ~comp
lynxy Link ParentI agree that "quality" is going to be largely subjective- though I wish it weren't so. I have some 2TB of models fetched to local, of different origins and different quantisations, and lordt knows...I agree that "quality" is going to be largely subjective- though I wish it weren't so. I have some 2TB of models fetched to local, of different origins and different quantisations, and lordt knows it might take some time to get an appreciable idea of what feels "good" for me.
By "prompt processing speed" and "response generation speed", I'm assuming you're referring to what is commonly abbreviated to PP and TG (Prompt-Processing and Text/Token-Generation). Benchmarking is certainly a good idea, and I have found a few benchmarks scattered around which use similar-or-same hardware to what I have, at least for the GPU. At the end of the day, this might end up being something I experiment with and then drop for another year or two until the results are actually useful for me- I don't know. I appreciate the answers and the info, though.
As for tooling, MCP servers and such are a little overwhelming from an outside perspective! I understand the concept, but wow has the space absolutely blown-up, and I'm not sure much of it is well thought-through / designed, instead of just functional vibe-coded trash.
-
Comment on What programming/technical projects have you been working on? in ~comp
lynxy LinkI've just pulled apart, cleaned, and updated my rackserver somewhat- replacing the ProArt B650-Creator board with the newer ProArt X870E-Creator, largely because of the jump from PCIE 4.0 to PCIE...I've just pulled apart, cleaned, and updated my rackserver somewhat- replacing the ProArt B650-Creator board with the newer ProArt X870E-Creator, largely because of the jump from PCIE 4.0 to PCIE 5.0 on all slots but the chipset-provided slot, and partly for the jump from 2.5G to 10G networking without needing a PCIE-based NIC of some description. I also jumped from the Intel Pro B50 to the Intel Pro B70 for the 32GB of VRAM, allowing me to experiment with Qwen 3.6 27B locally. On that front, I have a few queries:
- How do you figure out which model will get the best performance to quality ratio given what hardware you have? I've seen a lot of discussion about whether to use the 27B model or the 35B A4B (mixture of experts?) model, and how the former is better for complex programming tasks, whereas the latter will generate tokens a lot quicker due to only loading a portion of the parameters into memory at a time, but at a loss of quality. And that's just two sub-models- what about comparing Llama to Gemma to Qwen, etc?
- Does anybody use Open WebUI? How much more comprehensive is it than llamacpp's built-in llama-ui? Is it easy to switch them out?
On other fronts, I continue to refine a number of tools which I use for my smart home control, various REST endpoints, and my RSS / misc. app bot.
-
Comment on Lifetime Plex Pass will cost $750 USD after July 1st in ~tv
lynxy Link ParentI'm a little confused, so forgive me if this is a silly question- but has Jellyfin not had hardware transcoding capability for a long time at this point? I'm currently using an Intel GPU to...I'm a little confused, so forgive me if this is a silly question- but has Jellyfin not had hardware transcoding capability for a long time at this point? I'm currently using an Intel GPU to transcode media with Jellyfin without issue.
I admit that the Plex app offerings have tended to be more comprehensive, but Finamp has become pretty good for music streaming! And the rest of the Jellyfin app ecosystem seems to be improving too. I entirely understand the choice maybe five years ago, but in a more modern context?
-
Comment on What programming/technical projects have you been working on? in ~comp
lynxy Link ParentAh yeah, I'm referring to a device without a "seat", as it were. A device with no physical input or output- though I'm not sure if rendering to an off-screen buffer and then streaming that over...Ah yeah, I'm referring to a device without a "seat", as it were. A device with no physical input or output- though I'm not sure if rendering to an off-screen buffer and then streaming that over the network nullifies the classification? :P
It sounds like you ran into a lot of edge-case OpenGL version-missmatch problems! I'm glad you managed to work your way around the majority of them.
I agree that a GPU in a server can be used for so much. My rackserver, though largely consumer desktop hardware (I couldn't bring myself to splurge for even a second-hand Epyc chip..), does a wonderful job transcoding Jellyfin streams, as it's an Intel ARC and has encode/decode support for most codecs! It also has 32GB of VRAM, so I'm experimenting with hosting Qwen3.6 or Gemma4, though it looks like MTP has issues with the SYCL backend at the moment. I'd love to see how well it works for game streaming, though I don't expect it to hold a candle to a proper gaming-oriented card running on baremetal, haha.
-
Comment on What programming/technical projects have you been working on? in ~comp
lynxy Link ParentI've been tempted to set something like this up! Is the VM machine headless? Do you use any form of HDMI output faker stick? I know that OpenGL has issues with rendering to buffers without a...I've been tempted to set something like this up! Is the VM machine headless? Do you use any form of HDMI output faker stick? I know that OpenGL has issues with rendering to buffers without a display- Vulkan is much better in that respect in that it can render to "off-screen buffers".
I was wondering how possible it would be to dynamically spin up a VM or container containing the streamed application in some kind of well-optimised kiosk DE, with a Sunshine instance, and have it clear itself up again on disconnect, or after a period of inactivity.
My primary computing device these days is a relatively low-power laptop attached to a dock, and it would be neat to run more graphically heavy programs inside my rackserver instead!
-
Comment on David Koepp to write a Westworld film for Warner Bros in ~movies
lynxy (edited )Link ParentI massively enjoyed the first season, and the second season was a lot of fun- even if it got a little confusing. So many scenes were dominated by Jeffrey Wright and Anthony Hopkins, and I loved...I massively enjoyed the first season, and the second season was a lot of fun- even if it got a little confusing. So many scenes were dominated by Jeffrey Wright and Anthony Hopkins, and I loved Thandie Newton! The latter seasons sort of lost my interest, unfortunately.
-
Comment on Tildes Gardening Group: Week 3/5/26 in ~hobbies
lynxy LinkNot a lot from this end- I'm down to one lavender seedling, but it's looking quite robust, so I'm honestly not too worried. We have managed to root some sweet (Thai?) basil which we got as...Not a lot from this end- I'm down to one lavender seedling, but it's looking quite robust, so I'm honestly not too worried. We have managed to root some sweet (Thai?) basil which we got as cuttings for a salad, which will complement the number of holy (I think?) basil plants which I have managed to propagate.
The strawberries are moving slowly, but still alive. I think I just need to water them more often. They're in a large balcony planter, and they're only a centimeter or two tall each, and I think the amount of soil dries quickly in this early summer heat.
And lastly, I've been dumping tomato plants on any neighbours who will take them. I had close to 26 germinate and survive to a reasonable size, and it's just too much for me. I only need 4- 2 of each kind. Luckily my neighbours seem to be quite happy to receive free plants for their balconies. And it gives me a good opportunity to practice my German (I have the A1 exam on Tuesday, and then we'll be moving onto A2/1 before the end of semester). I might have to start giving out succulents, too, as mine are dropping so many leaves which each bud into a new plant if I give them water. They're cute, and relatively hardy.
-
Comment on Happy Birthday David Attenborough, 'the voice for nature,' turns 100 in ~enviro
lynxy LinkI had no idea that both Obama and Billie were a part of the royal family? On a more serious note- I grew up on Attenborough documentaries. His voice, and his approach to nature, were and are a...Counting Britain's royal family, Barack Obama and pop star Billie Eilish among his admirers..
I had no idea that both Obama and Billie were a part of the royal family?
On a more serious note- I grew up on Attenborough documentaries. His voice, and his approach to nature, were and are a formative part of who I am. I wish him good health for as many years as he wants, and given his love for his work, I'm sure he wants many more.
-
Comment on Trying to find other uses of a specific sample in ~music
lynxy (edited )LinkIt, or a similar sample, seems to also be used in "Game" by Rizha. I don't know if this helps with your search. Edit: A Reddit comment links to the following broken NASA archive link, commenting...It, or a similar sample, seems to also be used in "Game" by Rizha.
I don't know if this helps with your search.
Edit: A Reddit comment links to the following broken NASA archive link, commenting that it is a "ringtone" from or for the Astronaut known as Cooper. Unfortunately the archive links all seem to be a little messed up.
-
Comment on Linux privilege escalation (CVE-2026-31431) in ~comp
lynxy Link ParentI imagine it was fixed in 7.0, but has been backported to 6.19 and 6.18 as of the subversions 6.19.12 and 6.18.22. I would think that 6.19.14 has the fix- the distros have had over a month to...I imagine it was fixed in 7.0, but has been backported to 6.19 and 6.18 as of the subversions 6.19.12 and 6.18.22. I would think that 6.19.14 has the fix- the distros have had over a month to respond to the CVE.
-
Comment on Networking: DHCP+VPN, LAN communication query in ~comp
lynxy Link ParentYes- split tunneling, as you described it, is correct. This is on both Linux and Android. I'm hoping for a solution that would carve out all LAN ranges as described by RFC1918, and work across any...Yes- split tunneling, as you described it, is correct. This is on both Linux and Android. I'm hoping for a solution that would carve out all LAN ranges as described by RFC1918, and work across any LAN (no matter what subnet is used, or what IP the gateway has). A solution that is implemented on the network-infra-side would also be acceptable, as I wouldn't have to think about implementing it on all devices I own in the house.
WG Tunnel sounds like it might be a good tool to look into- I'll pull it and check it out. Thanks!
-
Comment on ps5-linux-loader: Linux on a PlayStation 5 in ~games
lynxy LinkI wonder how soon it'll be until we see SteamOS supported- I note that the ps5-linux-image repo states that it:I wonder how soon it'll be until we see SteamOS supported- I note that the ps5-linux-image repo states that it:
Supports Ubuntu 26.04, Ubuntu 24.04, Arch, and Alpine
-
Comment on Linux privilege escalation (CVE-2026-31431) in ~comp
lynxy Link ParentYeah, unfortunately I think the whole page is AI generated (or at least edited by AI), but this seems to be a legit issue with an actual CVE and the page provides information on what devices are...Yeah, unfortunately I think the whole page is AI generated (or at least edited by AI), but this seems to be a legit issue with an actual CVE and the page provides information on what devices are most at risk.
-
Comment on Linux privilege escalation (CVE-2026-31431) in ~comp
lynxy LinkThe kernel mailing list entry can be found here.If your kernel was built between 2017 and the patch — which covers essentially every mainstream Linux distribution — you're in scope.
Copy Fail requires only an unprivileged local user account — no network access, no kernel debugging features, no pre-installed primitives. The kernel crypto API (AF_ALG) ships enabled in essentially every mainstream distro's default config, so the entire 2017 → patch window is in play out of the box.
-
Linux privilege escalation (CVE-2026-31431)
49 votes -
Comment on Networking: DHCP+VPN, LAN communication query in ~comp
lynxy Link ParentOh, totally- I used that site to generate the AllowedIPs chunk above. It just feels somewhat clunky as a solution, and for some reason, one some Linux devices, it causes all sorts of instability....Oh, totally- I used that site to generate the AllowedIPs chunk above. It just feels somewhat clunky as a solution, and for some reason, one some Linux devices, it causes all sorts of instability. The PreUp and PreDown technique is also outlined above, but I'm not satisfied that it works as a global solution across different LANs. Maybe I shouldn't be searching for perfect, as always, though..
-
Networking: DHCP+VPN, LAN communication query
Preface: I know not what I talk about. INSTANCE: A Unifi network with multiple VLANs, each with their own subnet. A Linux client that is assigned to a single VLAN, connected to the network via...
Preface: I know not what I talk about.
INSTANCE: A Unifi network with multiple VLANs, each with their own subnet. A Linux client that is assigned to a single VLAN, connected to the network via Wi-Fi, and running a full-tunnel Wireguard config which tunnels data to the provider's endpoints.
QUESTION: While running full-tunnel VPN configurations supplied by a commercial VPN provider, how might the client device talk with other devices on the same (local!) LAN as it is, including devices that are on a different (local!) VLAN, and thus a different subnet?
Let's say, for instance, that I have the following network architecture:
Name Subnet Gateway Mask Internal 10.0.0.0/24 10.0.0.1 255.255.255.0 Hosted 10.5.0.0/24 10.5.0.1 255.255.255.0 Private 10.5.1.0/24 10.5.1.1 255.255.255.0 Guests 10.5.2.0/24 10.5.2.1 255.255.255.0I have a device that is connected to the network via Wi-Fi, and is contained within the "Private" VLAN. It can also talk to devices that are in the "Internal" VLAN (by necessity), and devices in the "Hosted" VLAN.
Once I spin the VPN up, using a configuration gained from OVPN / PIA / NordVPN / whatever, the client can still communicate with devices on the same VLAN as it- for example, if the client is 10.5.1.132, it can still communicate with 10.5.1.42, but it cannot communicate with, say, 10.5.0.11. One would assume that is because the DHCP server has told it that it can access devices within a specific range through the correct gateway- and in-fact, this shows in the IP routing table:
Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface default unifi.localdoma 0.0.0.0 UG 1024 0 0 eth0 10.5.1.0 0.0.0.0 255.255.255.0 U 1024 0 0 eth0 unifi.localdoma 0.0.0.0 255.255.255.255 UH 1024 0 0 eth0There are, as far as I can tell, a number of solutions for this problem. The first is to not use a full-tunnel VPN, and instead build a set of AllowedIPs which carve out the LAN ranges which you do not want tunneled. This, I think, is known as a split-tunnel VPN. If one wants to carve out all of the possible LAN IP ranges, as specified in RFC1918, it starts to become somewhat cumbersome:
AllowedIPs = 0.0.0.0/5, 8.0.0.0/7, 11.0.0.0/8, 12.0.0.0/6, 16.0.0.0/4, 32.0.0.0/3, 64.0.0.0/2, 128.0.0.0/3, 160.0.0.0/5, 168.0.0.0/6, 172.0.0.0/12, 172.32.0.0/11, 172.64.0.0/10, 172.128.0.0/9, 173.0.0.0/8, 174.0.0.0/7, 176.0.0.0/4, 192.0.0.0/9, 192.128.0.0/11, 192.160.0.0/13, 192.169.0.0/16, 192.170.0.0/15, 192.172.0.0/14, 192.176.0.0/12, 192.192.0.0/10, 193.0.0.0/8, 194.0.0.0/7, 196.0.0.0/6, 200.0.0.0/5, 208.0.0.0/4, 224.0.0.0/3This might be the only solution that currently works on an Android device running the basic Wireguard APK, as found on Github.
An alternative solution is to carve out the LAN ranges you want to avoid routing through the tunnel using the PostUp/PostDown entries. For example:
PostUp = ip route add 10.5.0.0/16 via 10.5.1.1 [dev IFNAME?] [metric INT?] PostDown = ip route del 10.5.0.0/16 via 10.5.1.1This would work, on your local LAN, as long as you stay on the same VLAN and can connect to the correct gateway. Unfortunately, Android Wireguard cannot handle PostUp/PostDown entries, AFAICT. Small bother.
One final solution that I have heard mutterings about is known as "Classless Static Routes", or DHCP Option 121/249. This uses "CIDR" notation, and I'll be honest- here I am a little bit lost. I like the sound of this solution, as it means that I can configure the networking infrastructure itself to provide the required information for valid routing, and not have to faff about with the many, many Wireguard configurations on my devices.
Does anybody have any experience with this problem? Am I approaching this problem in an entirely stupid way? Is there a better way (insert infomercial here)?
8 votes -
Comment on Looking for early users to try my app in ~tech
lynxy LinkJust a quick heads-up- I gave the site a read-through, and not once on the primary page is the browser or platform mentioned (unless I have somehow missed it). It would be a good idea to make it...Just a quick heads-up- I gave the site a read-through, and not once on the primary page is the browser or platform mentioned (unless I have somehow missed it). It would be a good idea to make it immediately obvious which are supported, and how (browser addon, mobile app, et cetera).
I was interested in running it in my smart-home, but it seems too good to be true, and the whole README looks like AI generated garbage- so I did a little searching online, and it seems the general sentiment is that it's all a bit of a scam to sell the "Seed" sensor for a couple of hundred dollars (each piece of marketing material shows a different price..). Allegedly, the whole project is vibe coded, and it uses a bunch of hardcoded data? I've decided that I'm not interested enough to actually look into the code myself, and if you come to a different conclusion I would love to hear about it!
EDIT: I'm pretty sure what is outlined in the paper is largely possible, to varying degrees of confidence- I just don't think this project is it, chief.