Presently trying out an account over at @tal@oleo.cafe due to scraping bots bogging down lemmy.today to the point of near-unusability.
An April MIT study found AI Large Language Models (LLM) encourage delusional thinking, likely due to their tendency to flatter and agree with users rather than pushing back or providing objective information.
If all it takes to get someone to believe something is to flatter them and agree with them, it does kind of explain how people manage to sell people on all kinds of crazy things.
Ram is cheap
Kind of divering from the larger point, but that’s true — RAM prices haven’t gone up as much as other things have over the years. I do kind of wonder if there are things that game engines could do to take advantage of more memory.
I think that some of this is making games that will run on both consoles and PCs, where consoles have a pretty hard cap on how much memory they can have, so any work that gets put into improving high-memory stuff is something that console players won’t see.
checks Wikipedia
The XBox Series X has 16GB of unified memory.
The Playstation 5 Pro has 16GB of unified memory and 2GB of system memory.
You can get a desktop with 256GB of memory today, about 14 times that.
Would have to be something that doesn’t require a lot of extra dev time or testing. Can’t do more geometry, I think, because that’d need memory on the GPU.
considers
Maybe something where the game can dynamically render something expensive at high resolution, and then move it into video memory.
Like, Fallout 76 uses, IIRC, statically-rendered billboards of the 3D world for distant terrain features, like, stuff in neighboring and further off cells. You’re gonna have a fixed-size set of those loaded into VRAM at any one time. But you could cut the size of a given area that uses one set of billboards, and keep them preloaded in system memory.
Or…I don’t know if game systems can generate simpler-geometry level-of-detail (LOD) objects in the distance or if human modelers still have to do that by hand. But if they can do it procedurally, increasing the number of LOD levels should just increase storage space, and keeping more preloaded in RAM just require more RAM. You only have one level in VRAM at a time, so it doesn’t increase demand for VRAM. That’d provide for smoother transitions as distant objects come closer.
no TAA
I would like TAA to be available. I think maybe a more reasonable ask is “let me toggle TAA”.
Is he running with antialiasing on?
Looking at their system settings page:
https://borderlands.2k.com/borderlands-4/news/amd-optimization/
Their settings for every single GPU listed there seem to have antialiasing off.
It may be that current hardware can’t do that at a reasonable clip
I found that while the lemmy_server process starts successfully and shows “Starting HTTP server at 0.0.0.0:8536” in logs, nothing is actually listening on port 8536.
Does:
# netstat -ntap|grep 8536
…show anything bound to the port?
I’m not sure how you determined that it’s not binding to the port, but that’s how I’d check.
There isn’t much that should stop a process from listening on a port over 1024 unless another process is already listening on it.
I don’t know why you want a BIOS update.
If you’re wanting VT-x support, looks like it’s present.
https://mobilespecs.net/laptop/Toshiba/Toshiba_SATELLITE_M100-221.html
Processor Model: T5600
https://www.techpowerup.com/cpu-specs/core-2-duo-t5600.c378
VT-x
https://superuser.com/questions/1584771/what-is-difference-between-vmx-and-vt-x
The CPU flag for VT-x capability is “vmx”; in Linux, this can be checked via /proc/cpuinfo
I’d assume so, but more importantly, for both, there’s a cryptographic signature being performed by the card. The credentials never leave the card — there’s a private key on the card, and what goes out is a signature on the transaction, which is useless for doing other transactions.
In most respects, CRTs were technically worse, but a lot of video game art was designed around their characteristics, optimized for them, and thus can appear better on a CRT. We can — and do — try to emulate CRT quirks on LCDs/LEDs to varying degrees, but we’re always going to be just approaching what a CRT looked like.
https://wackoid.com/game/10-pictures-that-show-why-crt-tvs-are-better-for-gaming/
Final Fantasy 6:
Castlevania: Symphony of the Night:
Final Fantasy 7:
Things like blurriness are a limitation in the fidelity of the display, true enough, but in the context of material optimized for that display, it can be a positive rather than a negative.
Well, there’s certainly that. But even then, I’d think that a lot of videos could be made to be more concise. I was actually wondering whether YouTube creators get paid based on the amount of time they have people watch, since that’d explain drawing things out. My impression, from what I could dig up in a brief skim, is that they’re indirectly linked — apparently, YouTube shows ads periodically, and the more ads shown, the more revenue the creator gets. So there would be some level of incentive to stretch videos out.
-the opening the port process makes sense. It seems like if I have a backend on my rig, I’m going to need to open a port to access that backend from a front end of a phone device.
Yes. Or even if you run a Web-accessible front-end on the LLM PC — the Web browser on the phone needs to reach the Web frontend on the PC.
Or possibly even access that same backend on the phone device via a mirror?
Well, the term wouldn’t be a mirror. In your shoes, it’s not what I would do, because introducing some third host not on your network to the equation is another thing to break. But, okay, hypothetically, I guess that doing that would be an option. thinks. There might be some service out there that permits two devices to connect to each other, though I’m not personally aware of one. And, say you got a virtual private server for $10 a month or whatever the going rate is, yeah, that could be set up to do this – you could use it as an intermediate host, do SSH tunneling from both the PC and the phone of the sort that another user in this thread mentioned. I guess that that’d let you reach the PC from other places, if that’s something that you want to do, though it’s not the only way to accomplish that. But…I think that that’s most-likely going to add more complexity. The only scenario where that would truly be necessary is if the wireless access point — which I assume your ISP has provided — absolutely does not permit the LLM PC and the phone to communicate at all on the WiFi network, which I think is very unlikely, and even then, I’d probably just get a second wireless access point in that scenario, put the PC and the phone on it.
In general, I don’t think that trying to connect the two machines on your home network via a machine out on the Internet somewhere is a great idea. More moving parts, more things to break, and if you lose Internet connectivity, you lose the ability to have them talk to each other.
-it seems like it would be easier if I could connect to the rig via an android phone instead of an iPhone. My end goal is to use Linux but I’m not ready for that step. Seems like android would be an adequate stepping stone to move to, especially if we have to go thru all this trouble with iPhone. Shall we try on the android instead? If not I’ll follow the directions you put above and report back on Saturday.
If you have an Android phone available, that would probably be easier from my standpoint, because I can replicate the environment; I have an Android phone available here. But it’s not really the phone where setup is the issue. Like, it’s going to be the LLM PC and potentially wireless access point that require any configuration changes to make ollama reachable from the phone; the phone doesn’t need anything other than Reins installed and having an endpoint set or just using a Web browser and using the correct URL there. I’m just mostly-interested in that the phone has to be able to talk to the PC, has to be able to open a TCP connection to the PC, and so having diagnostic tools on a phone is helpful. I don’t have to guess how the diagnostic tools work in Termux on an Android, because I can use them myself locally.
I wouldn’t suggest going out and buying an Android phone to do just that, though. I mean…this is a one-off diagnostic task, just trying to understand why the phone isn’t able to reach the LLM PC. If you can open a connection from the Android phone to the LLM PC, then you should also be able to open a connection from the iOS phone to the LLM PC. If you do have one already available, though, then yeah, my preference would be if you could install Termux on it for the diagnostic tools rather than install iSH on the iOS device. It should still be possible to get the LLM PC reachable on the iOS device either way.
I don’t mind trying to diagnose connectivity on the iOS device. Just keep in mind that I may have to guess a bit as to what the behavior is, because I can’t actually try the device here, so we may potentially have a few extra rounds of back-and-forth.
If you do want to use an Android phone, then just put the phone on the WiFi network, install Termux, open Termux, run the apk command to install telnet (apk install telnet) and then try the telnet command I mentioned and just report back what error, if anything, you get when trying to open a connection to the LLM PC — hopefully it’ll be one of the above three outcomes.
Deregulation might give some amount of an edge, but I really don’t think that in 2025, the major limitation on deployment of AI systems is overbearing regulation. Rather, it’s lack of sufficient R&D work on the systems, and them needing further technical development.
I doubt that the government can do a whole lot to try to improve the rate of R&D. Maybe research grants, but I think that industry already has plenty of capital available in the US. Maybe work visas for people doing R&D work on AI.
The oldest two mechanisms of authenticating on credit cards.
From oldest to newest, they are:
Printed data on card.
Magstrip (which basically has the same data in machine-readable form).
Smartcard chip with contacts.
Wireless.
The first two mechanisms hand over all the data required to impersonate the cardholder whenever used, which isn’t very secure. Yes, there’s value to keeping a mechanism around for a while to permit transition time, but we should have had tap-to-pay hardware on PCs and phones and the like a long time ago.
When the current leader Kim Jong Un came to power in 2011, the escapees who were interviewed said they had hoped their lives would improve, as Kim had promised they would no longer need to “tighten their belts” – meaning they would have enough to eat.
IIRC from past reading, he did try to increase the amount of meat available, which for North Korea is a big deal.
kagis
While North Korean diets have historically been plant-heavy, there have been efforts to increase the availability of protein sources, especially since 2005. Despite these efforts, structural and practical limitations prevent major protein farming expansion, including the competition for food stocks, resources and land allocations, much less the ability to acquire seed animals and raise them.
Prior to 2000, except for North Korea’s elites, the country subsisted principally on vegetarian diets. To have meat as few as two to three times a year was the apparent norm. Under Kim Jong Il, that began to change as efforts to expand the availability of animal protein to more of the population began around 2005. Under Kim Jong Un, there has been an even greater emphasis on animal husbandry, including poultry, pig, rabbit and larger grazing animals such as sheep, goats and cattle.
Yes. I’ve never had a bad liquid incident, but I don’t want my phone to be one spilled drink or dropped-into-toilet-or-puddle away from being fried.
I choose Ollama because it was supposed to be one of the easier loca AIs to set up.
Well, the ollama bit is up, which is why you can use it on the PC. The problem is network connectivity between the Windows PC and the phone.
Opening a port between two things on the local network is going to be pretty much the same for anything. Some software packages — I dunno about LLM chat stuff — make use of a third, outside system as a point to coordinate, so that software only has to open outbound TCP connections to the Internet. But for local communication, it’s gonna look pretty similar. If you put koboldcpp or llama.cpp or whatever on your machine, you need the same connectivity, though it might default to using a different port number.
I’m happy to keep banging away if you’re also willing, though. I mean, this does kinda narrow it down. If you don’t want to do so though, remove that firewall rule that we added earlier from the Windows PC. If you do:
considers
The next step is seeing where the break in connectivity is.
I’m not familiar with iOS, but let me see if there’s a software package for it that will let it open a TCP connection and preferably ping (and ideally show the ARP cache to see whether Ethernet packets are getting from the phone to the Windows machine at all, though that may not be viable).
Basically, would be nice to see whether packets can currently get from the phone to the PC and back.
kagis
Looks like Windows Firewall blocks ICMP by default, which is traditionally used by ping, the simple protocol to see if one host can reach another on the network. Mmmmf.
And it sounds like ARP isn’t available on a non-jailbroken iPhone, which would be the simplest way to see whether a packet is making it from the iPhone to the PC. I was worried would be the case.
Hmm. This is a little less convenient in that I don’t have tools that I normally would when trying to troubleshoot network problems on a Linux system.
thinks
I guess the simplest thing available, cuts things down as far as possible in terms of connectivity between the two that should be able to reach from the iPhone to the PC should be a TCP connection.
I don’t know the iOS software library well, but lemme search for a telnet client. I’m sure that it’ll have one; every platform does.
searches
Oh, this is even better. It looks like there’s some iOS app, “iSH”, with a tiny Alpine Linux environment for iOS, kinda like Termux on Android. That’ll have telnet and probably other network diagnostic tools, and those I am familiar with, so I don’t have to guess from screenshots how things work. You should be able to try to open a TCP connection from the phone to the PC with the Linux telnet client in that.
goes looking around
Okay. If you’re willing to give this a shot, it sounds like the way this works is:
https://apps.apple.com/us/app/ish-shell/id1436902243
Install that from the iOS store.
When opened, it should show a Linux terminal. If it works like Termux, it’ll have basically nothing from Alpine Linux installed, no telnet client, just a few simple commands. You’ll be looking at a prompt that probably looks something like iPhone:~.
Then if you run (don’t type the pound sign — it’s just a convention to include it, to show that it’s something to type at a prompt):
# apk install telnet
That should install the Linux telnet client inside the iSH app using the Alpine Linux package manager.
Then to try to open a TCP connection from the phone to the Windows PC, you want the private IP of the Windows PC, the thing you see in ipconfig (which I’ll type as 10.1.1.2 here, but replace with yours):
# telnet 10.1.1.2 11434
That’ll try to open a TCP connection from the phone to port 11434 on the PC.
Now, what would happen if everything were working correctly, is that the phone would send an ARP request saying “what is the MAC address — the Ethernet address — of the machine with IP address 10.1.1.2 on the local network?” The wireless access point would hand this to the PC. The PC would respond. The phone would then send a series of packets to that IP address to open a TCP connection on port 11434.
My guess is that you’ll see one of several things at this point.
First, it might be that the wireless access point is refusing to let packets from the phone reach the PC at all — they only let the phone talk to the Internet, not to the PC. Some wireless access points can be configured to do this or have a “guest” wireless network that impose this constraint. Then the phone won’t get an ARP response, since the PC will never see the ARP query. That’ll look like this (using a network I’m on at the moment to demonstrate):
$ telnet 192.168.1.35
Trying 192.168.1.35...
telnet: Unable to connect to remote host: No route to host
$
Second, it might fail because I dicked up in some way and Windows Firewall is still blocking the phone from connecting to the PC. The ARP request is going out, the response comes back from the PC, the phone tries to open a TCP connection to the IP address on the host with the specified MAC address, and never gets a response. If that’s the case, it’ll probably look like this:
$ telnet 192.168.1.126 7500
Trying 192.168.1.126...
telnet: Unable to connect to remote host: Connection timed out
$
If that’s what you get, the problem is likely the Windows Firewall configuration (well or theoretically the wireless access point could be configured to do that, but I doubt it).
Third, it might succeed. That’ll look like this:
$ telnet 192.168.1.126 6000
Trying 192.168.1.126...
Connected to 192.168.1.126.
Escape character is '^]'.
If you see that, you can open a TCP connection from the phone to the PC, and whatever issue you’re hitting with Reins isn’t a network problem. Maybe I gave the endpoint syntax wrong, for example. But the issue will be at the application level, not the network level.
There are 3 lines with the :11434 in them. No brackets or anything like that. -1 has 0.0.0.0 in front -2 has 10.#.#.# in front and has a foreign address that is something other than 0.0.0 -3 is like the 2nd but a slightly different foreign address
Okay, that…should be okay. As long as all of the addresses that it’s listening on are IPv4 — of the format “x.x.x.x”. No colons in them (other than the colon preceeding “11434”). Not IPv6.
The subnet mask is 255.255.255.0
Okay, gotcha. In that case, go ahead with the instructions above, just instead of “/8”, do “/24”. So:
For “local IP addresses”, you want “These IP Addresses”, and enter
10.0.0.0/24. That’ll be every IPv4 address on your Windows LLM that has “10” as its first number and the first following two numbers the same as yours.For “remote IP addresses”, you want “These IP Addresses”, and enter
10.0.0.0/24. Same thing all addresses that start with a “10.” followed by the same following two numbers, which should include your iOS device.
Oh yes. I’m on windows 10 as well.
Okay. I think that the interface to add the firewall rule there looks the same as the one I Iinked to above. I went searching for screenshots of adding a hole for a port on Windows 10, and the control panel looks identical to me.
So, yeah, should be good to go ahead with the above instructions, just using “/24” instead of “/8” in the two places where I mention “/8”. Hopefully after that it’ll be working; if not, then we’ll need to troubleshoot.
copyright
This isn’t a copyright, but rather a patent.
-A backend is where all the weird c++ language stuff happens to generate a response from an AI. -a front end is a pretty app or webpage that takes that response and make it more digestible to the user.
Yes.
-agreed. I’ve seen in other posts that exposing a port on windows defender firewall is the easiest (and safest?) way to go for specifically what I’m looking for. I don’t think I need to forward a port as that would be for more remote access.
Yes. I’d like to confirm that that is not happening, in fact.
The ipv6 was identical to one of the ones I have.
Hmm. Okay, thanks for mentioning the IPv6 thing. It is possible to have ollama reachable from the Internet via IPv6, if it’s forwarded. I should have thought of that too and mentioned that. Shouldn’t need to open an IPv6 hole in the Windows Firewall, but would rather not rely on the Windows Firewall at all.
It shouldn’t be an issue if ollama is only listening on an IPv4 address. You only see the “0.0.0.0:11434” line, right? No other lines, probably with brackets in the address, that have a “:11434”, right? That could be an IPv6 address.
goes to look for an example of Windows netstat output showing a listening IPv6 socket
Here:
https://www.configserverfirewall.com/windows-10/netstat-command-to-check-open-ports-in-windows/
Can you just make sure that there’s nothing like 0:[::]:11434 in there? That’d be what you’d see if it were listening for IPv6 connections.
Sorry, just don’t know oollama’s behavior off the top of my head and want to be sure on this before moving ahead, don’t want to create any security issues.
The ipv4 was not identical. (But I don’t think that matters moving forward.)
Yeah, that’s expected and good. The one from the website is your public IP address, anf the one from ipconfig your private one, that you’ll use to talk to the machine wirh your phone.
I had to go into the settings in the ollama backend app to enable “expose Ollama to the network”.
Great, yeah, that was the right move.
Okay, then just want to sanity check that your iOS device is in the same address range on your WiFi network, that the 10.x.x.x address on your LLM PC isn’t from a VPN or something (since it’s a little unusual to use a 10.x.x.x address on a home broadband router, and I want to make sure that that’s where the address is from). Go ahead and put the iOS device on your WiFi network if you have not already.
This describes how to check the IP address on an iOS device.
You should also be seeing a 10.x.x.x address there. If you don’t, then let’s stop and sort that out.
If that’s a 10.x.x.x address as well, then should be good to go.
Oh, one last thing. In the ipconfig output, can you make sure that the “Subnet Mask” reads “255.0.0.0”? If it’s something different, can you provide that? It’ll affect the “/8” thst I’m listing below.
Okay, if you’ve got that set up and there are no other “:11434” lines and the Subnet Mask is “255.0.0.0”, the next is to poke a hole in Windows Firewall on IPv4 TCP port 11434.
kagis for screenshots of someone doing this on Windows 11
https://windowsreport.com/windows-firewall-allow-ip-range/
I’m assuming that this is Windows 11 on your PC, should have asked.
You’re going to want a new inbound rule, Protocol TCP, Port 11434.
For “local IP addresses”, you want “These IP Addresses”, and enter 10.0.0.0/8. That’ll be every IPv4 address on your Windows LLM that has “10” as its first number — you said that you had a “10.” from ipconfig.
For “remote IP addresses”, you want “These IP Addresses”, and enter 10.0.0.0/8. Same thing all addresses that start with a “10.”, which should include your iOS device.
And you want to select “Allow this connection”.
Okay. Now you should have a hole in Windows Firewall. Just to confirm that port 11434 isn’t reachable from the Internet, I’m gonna use one of the port-open-testing services online. My first hit is for one that only does IPv4 and another that only does IPv6, but I guess doing two sites is okay. Can you go to this site (or another, if you know of a site that does port testing that you prefer)
https://www.yougetsignal.com/tools/open-ports/
Plug in your public IPv4 address there (not the private one from ipconfig, the one from that website thst I listed earlier) and port 11434. It should say “closed” or “blocked” or something that isn’t “open”. If it’s “open”, go back and pull that firewall rule out, because your router is forwarding incoming IPv4 connections to your LLM PC in some way that’s getting to ollama, and we gotta work out how to stop that.
https://port.tools/port-checker-ipv6/
Here’s an IPv6 port tester. Plug in your IPv6 address there (which you said was the same from both the website and ipconfig) and port 11434. It should also say “closed” or “blocked” or similar. If it says “open” — I very much doubt this — then go back and pull out the firewall rule.
If both say “closed”, then go ahead and install Reins.
Based on this:
https://www.reddit.com/r/ollama/comments/1ijdp1e/reins/
It’ll let you input an “endpoint”.
Plug in the private IPv4 address from your LLM PC, what was in ipconfig, in the form of an http URL on the ollama port, like “http://10.something.something.something:11434/” and you should, hopefully, be able to chat.
If all this is working and you’ve given your Windows PC a name, you might want to go back to that endpoint setting and replace the IP address there with the name of your LLM PC. I don’t know for sure what the mDNS situation is on iOS or Windows, but if that works, that way, if your Windows PC loses its DCHP lease and gets a new IP address at some point from your broadband router, it won’t break connectivity for Reins as Reins tries to use the old IP address.
One, if it got lucky.