<unnamed>

Public Key

npub1qdjn8j4gwgmkj3k5un775nq6q3q7mguv5tvajstmkdsqdja2havq03fqm7

Profile Code

nprofile1qqsqxefne258ydmfgm2wfl02fsdqgs0d5wx29kweg9amxcqxew4t7kqprpmhxue69uhhyetvv9ujuumwdae8gtnnda3kjctvqy2hwumn8ghj7un9d3shjtnwdaehgu3wdejhgc6auae

Publishing to

Last Notes

2026-03-03 20:36:57 CET

- reply

I paid less for SSDs 7 years ago for the same size than even 6 months ago. Just when It's about time I up my storage game prices are screaming. Robbery these days.
Still laughing that I've been selling a few sticks of spare memory for whole fucking machines XD

2026-03-03 20:33:19 CET

- reply

Good question! I just started using OpenObserve a couple months ago so I haven't learned to do any complex things like set unions to compare things like that.
Now you've got me intrigued XD

2026-03-03 20:18:26 CET

Looks like Satellite.earth is overtaking Jumble in this weeks battle of the referred nostr clients.

2026-03-03 20:09:41 CET

- reply

Great to have some validation on this. I think QC is where I've been focusing the most. My code output was never the problem, it's make sure the quality at the end is high, and steps to keep the output consistent don't require excessive effort. I also find it neat because it's usually super simple tasks that stack. Even good for tiny self-hosted models to do. Like spell checking, consistency, docs, checklists, dead code, etc. Just being able to get reports that I can act on with high guarantees is a huge improvement imo.

2026-03-03 19:23:46 CET

- reply

Yeah I'm very glad to hear that! Thanks for the info as well :)
I had a feeling the "industry", assuming they aren't focused only on growth have been focused more on what can we build to overall improve consistency. I'm not sure if you'd agree but that's what my thoughts have been. I've been working toward an incremental automation approach with very high predictability and very low human cognitive load. Just with prompts and easy method of getting feedback. Mostly just focused on feeding me high quality information in a repeatable way.

2026-03-02 18:51:42 CET

- reply

Yeah that's never stopped happening since I started using them.
What I've pivoted to now, is attempting to harness, predictable behavior in an automated way. I have since built up automated tasks like PR reviews, prechecks and so on. They all run headless in a very confined environment giving me itterative and predictable automation.
I just think slow, steady, and predictale is the way to make these things work long term in business. Otherwise the quality of your work is solely dependent on the API status of Anthropic and Amazon.

2026-03-02 17:40:05 CET

"you're doing it wrong bro"
#nevent1q…54t4

2026-03-02 02:04:30 CET

- reply

Sorry if I haven't been following the saga well, but did you figure out what was happening or is this still an unknown DDoS? Do you think you've sufficiently mitigated attacks like this or just fend off for now?

2026-03-01 23:06:11 CET

- reply

2026-03-01 00:30:27 CET

- reply

Made me look.

2026-03-01 00:27:55 CET

- reply

Doing what I can to help my people XD

2026-03-01 00:27:02 CET

- reply

Okay yeah mine are all fees over X bandwidth, and that price is not great. It's more of that object storage that starts getting expensive

2026-02-28 22:19:41 CET

- reply

Are you just worrying about bandwidth or is your compute price variable too?

2026-02-28 20:46:34 CET

- reply

2026-02-28 20:45:21 CET

- reply

Makes sense to me.

2026-02-28 18:15:15 CET

- reply

dnf repo info
Just because it's simple doesn't mean its easy, you're right... But it's on the education front at this point imo. There is only so much hand holding a distro can do for users without really getting in the way.

2026-02-28 18:13:42 CET

- reply

`dnf repo info`
> I cringe every time I see one of those pipe curl to bash oneliners.
Yup. Side note, I use Taskfile.dev in my projects and it has a neat side effect that end-users or contributors can do a --dry-run to see all of the commands the "script" will run before before it does them, and it will echo them during a normal build. For me it's just a nod to my users imo.

2026-02-28 18:09:34 CET

- reply

I also think maintainers should do their best to support all walks on this. Signed, unsigned, repo managed, direct package download, build from source, and signed commits.

2026-02-28 18:07:10 CET

- reply

Do you verify every IP packet that is forwarded through your firewall? Or do you trust your firewall maintains a robust state table? You do use recursive DNS with strict signature verification? Do you read the headers (and DKIM signatures) in your email client before reading them or do you trust you mail server did DMARC, DKIM, SPF, rDNS, and IP verification?
Unless you're using your using using IPTables as your firewall with debug logging on, your own recursive DNS resolver with a verifying server in front of it, or for email using K9-Mail on mobile or RoundCube on desktop to your self hosted mail server, chances are you didn't do any of those.
Do I wish people did more verification, you bet I do. But we'd all defer that if we could, it's obnoxious for even the advanced linux user to do, and many of not most system-level applications don't even ship outside the distro's repo, so you couldn't verify if you wanted to. Not to mention still, few projects even sign there commits let alone sign their builds.
I say this as someone who's built my own cli installer tool to manage my own dependencies because this is such a big deal. But I'm regularly interacting with normies on this (colleagues and customers) and know that they'd rather trust me to do it for them.

2026-02-28 17:49:17 CET

- reply

Windows by default requires 3rd party executables to be signed by Microsoft certificate chains.

2026-02-28 17:38:53 CET

- reply

Here we go. It's about time it caught up to everything else. good thing were heading out of winter, people need to heat their homes.

2026-02-28 17:35:57 CET

- reply

This is actually something I really missed about spotify. I used the AI-dj thing when it first released and liked it.
Feishin on desktop, and Tempus (fork of Tempo) on Android is coming a really long way with Navidrome. It probably works well with other scrobble subsonic servers, but Navidrome is probably the most feature packed and maintained. Feishin just added an Auto DJ that attempts to add songs to your queue based on the current tracks in your queue. I believe it does it through listening data along with good tagging. Mine still kind of sucks after about 10 tracks. Feishin doesn't get the genere down, and Tempus starts to repeat tracks after a 10 or so.
For music tagging, MusicBrainz Picard has been way ahead imo. It still has to be pretty hands on, but it does work really well compared to others Ive used in the past.

2026-02-28 17:14:22 CET

- reply

hosts run on proxmox for now yes.

2026-02-28 17:13:46 CET

- reply

I've got plenty of tarballs to go around lol
https://www.vaughnnugent.com/resources/software/modules/VNLib.Core

2026-02-27 22:58:55 CET

- reply

> If you could change one thing only to better suit your needs, what would it be? Why?
In both cases, user and dev. Make a fairly universal package archive file that fits within a tarball like npm does. Or even better, make a package format around tarballs that supports most of the popular package managers (backward compatability).

2026-02-27 22:54:53 CET

- reply

From the user standpoint or developer?
tarballs are a horrible option for non-advanced users. And unless signatures are pushed, it's on users to verify package signatures
dev:
- tarballs are the easiest way to bundle an application with it's dependencies and config
- super easy to push to a webserver
- basically universal for all windows, linux, and mac systems since ~2019 I think.
- I don't need permission from package maintainers to ship to users
- I don't have to ask permission from half a dozen or more repo maintainers
- because it's a universal archive format I don't need to create different package types to support a bunch of different distros.
user:
- for advanced users: tarballs are straightforward and standard. Users have the power to see all the application files (binaries, libs etc) and put them wherever they want (if supported)
- It's easy, I just curl https://yoursite.com/path/to/tarball.tgz
- I can more hands-on do things like verify sums and signatures - don't trust verify kind of thing.
- I don't need to rely on repo maintainers and the developer's relationship with them.
Frozen-point OS typically freeze packages so maintainers might ship later versions of an app, and if you're not on the latest rolling OS version you're left behind and the maintainer. The only option is to hope the maintainer ships a tarball I can manually extract onto my system to get up and running.

2026-02-27 22:15:18 CET

- reply

Brother... Im scared you haven't already XD

2026-02-27 21:33:29 CET

- reply

^ I've been shipping tarballs or self extracting tarballs.
All have their benefits and drawbacks. But I think the direction RedHat is moving in (more immutability) there will be a place for system packages (RPM) and self-contained user-app packages (flatpaks).

2026-02-27 20:39:06 CET

If anyone is interested in partnering for bulk server hardware purchases lmk. That goes for GPUs, DDR4 memory and full machines.
Most of these bulk auction brochures ads I get are usually starting around $4000 up to $20,000 and beyond for just memory or a skid of of machines. Not much in GPUs or storage but Im trying to keep my eyes out (that's more important right now)
For big stuff they usually ship freight so we'd have to be somewhat local to meet and pick up equipment.

2026-02-27 18:28:28 CET

- reply

I don't see a choice XD. Coding is co-pilot, everything else is local sir. I'm trying to practice what I preach.

2026-02-27 17:39:53 CET

- reply

Yeah i've had such a hard time getting abliterated models working with ollama.

2026-02-27 17:38:13 CET

- reply

Basically none i assume, ollama usually defaults to 4k...

2026-02-27 17:31:07 CET

- reply

Also that quant should work on ollama right?

2026-02-27 17:30:25 CET

- reply

What's the context windows size on that for 24g? That's gotta be really small? Less than 16k I assume.

2026-02-27 16:49:37 CET

- reply

Im not sure what most of this means, but flu fevers were wicked this year. I was stuck in bed for a few days over new year.

2026-02-27 16:19:50 CET

- reply

Wait a second. I have one in my left eyebrow XD

2026-02-27 03:02:47 CET

- reply

I do.

2026-02-27 02:56:33 CET

Cloudflare's zlib fork maintainers just said move to Rust or gtfo. Bummer, now I gotta add yet _another_ project to maintain myself XD
https://git.vaughnnugent.com/cgit/cloudflare/zlib.git/commit/?id=d9f746b9a0fb41b2c7d06ed199d14543acbb6275

2026-02-27 02:54:13 CET

I put a git mirror up of libc :)
https://git.vaughnnugent.com/cgit/gnu/glibc.git/

2026-02-27 02:52:38 CET

- reply

I have borked my simplex I haven't looked at how to recover my db or migrate machines just yet sorry.

2026-02-24 20:49:11 CET

- reply

https://www.vaughnnugent.com/public/resources/downloads/cms/c/h2ywh67pxfy42yuhlyqtsvvaqq.webp

2026-02-24 20:46:05 CET

- reply

https://www.vaughnnugent.com/public/resources/downloads/cms/c/ru7agrty5po24z6sacmrkbb6vq.jpg

2026-02-24 18:07:23 CET

- reply

Siglent always seemed to have a good reputation from the people I knew. Im sure they all have their issues at this point. It's not the fact that it failed, everything is software defined now, I'd expect them all to have an issue eventually.
It's that I can't repair it myself, nor are there options to get it repaired by the OEM. And even a $600-700 machine isn't worth it for a repair market. Especially since mfgs attack people who release the working firmware.
there is one poor soul on the EEV Blog forums who appears to work for siglent US who quietly distributes the firmware packages and asks people to keep quiet. It's been like 10 years this guy still appears to do it for free.

2026-02-24 18:04:38 CET

- reply

Learned my lesson I suppose. It was a work purchase at the time so my options were limited XD

2026-02-23 16:57:25 CET

- reply

Google has the "best" spam filtering for most people, organization, and integrations etc. I think proton is the best fit for gmail power users as they also support calendars, organization, and have fantastic spam filtering. I pay for their mail experience, and their Drive-like ecosystem is a little pricier, but I think it's worth it to be off google's panopticon.

2026-02-23 16:54:41 CET

- reply

I can still receive emails from gmail accounts, I just can't resolve google domains for my outgoing emails.

2026-02-23 16:53:33 CET

- reply

XD Most of it was coming from 20.214... and 20.16... ipv4 traffic. 20.214.155.242 was an offending IP. https://app.crowdsec.net/cti/20.214.155.242
otherwise it was pretty consistent IP blocks so coming from the same provider. That one appears to be Microsoft owned subnets. I saw quite a bit of cloudflare-owned worker traffic as well.

2026-02-22 23:40:56 CET

- reply

Yes but they only support Google Pixel devices. The maintainers can't stop bickering with people on social media. There is a good chance this note will summon one of them to come here just to tell me I'm brain dead.

2026-02-22 23:35:58 CET

- reply

The more tools privacy-focused you get, the more difficult accessing online resources will get in my experience.

2026-02-22 20:20:08 CET

- reply

Id wish everyone could self host, but it's a lot of work anymore, nearly impossible according to ISPs and government regulation. I personally use proton mail as my backup/cloud-based provider. I think i can recommend them over other popular options today, but they still manage your encryption keys with reversible encryption. The gold standard is controlling your server imo.

2026-02-22 16:56:00 CET

- reply

Sure
Google Voice - voip.ms, freepbx, cisco ATA and a desktop handset from 2004.
Gmail - self hosted (stalwart mail) + protonmail for backup
Chrome - librewolf/firefox
Drive - vpn + file browsers
Maps - Here WeGo + OSMAnd~
Youtube - invidious, yt-dlp, pinchflat
Android is tricky because I don't really use a smartphone. I have one, but does navigation, has firefox installed, and a file browser. I just re-flashed the stock rom, then used ADB to strip all unnecessary packages including google services. Obtainium is used for managing APK packages, otherwise manually managing them using ADB to install/update them.

2026-02-22 16:47:09 CET

- reply

I think here wego has the best maps alternative. you are relying on another SaaS company, but it has the best response, search, map features, turn-by-turn etc. I think it's a far more user friendly app. But it relies heavily on cellular networking. So if you're traveling to rural areas in the US like I do often, the navigation just drops out. There is no option for offline maps that I've seen. So it's great for local/short trips where you have an idea where youre going or will be near major towns with good cell servie.
OSM And honestly just sucks for basic navigation. Search is terrible, it swallows battery, menus disappear and won't come back without an app restart. The worst is that you can't choose routes, you get a single shortest path route, but you can't choose alternatives. You can draw your own route, but I've had bad experience with that as well. I think OSM and is power user app, it has tremendous offline maps tools and stays stable without good cell service. But I only use it when I know I need offline navigation and I have time to plan the route out.
Both apps crap out often enough that for areas I don't know or have a map in the car for, I usually print out backup directions, or a map of the area.

2026-02-22 16:38:53 CET

- reply

#nevent1q…ut3l

2026-02-22 16:38:11 CET

- reply

In what ways? I mean with the use of VPNs many/most websites, CDNs, and such block my Ip addresses so it can make it difficult to do basic things like web searches. Anything protected by cloudflare is off the table. So most tradfi banking and online shopping are difficult or impossible.

2026-02-22 01:54:51 CET

- reply

Oh well then HERE WeGo is a very good alternative then. OsmAnd is terrible at search and doesn't have those features whatsoever. I got into the habit of web searching things like that, personally found it more accurate than maps for my area.

2026-02-22 01:52:21 CET

- reply

Absolutely!

2026-02-21 23:39:10 CET

- reply

Aslo been thinking about picking up a newer garmin for the road. Nearly every alternative navigation app sucks bad in some way.

2026-02-21 23:38:28 CET

- reply

#nevent1q…ut3l

2026-02-21 23:37:31 CET

- reply

Yeah, it's my favorite navigation, but only when on trails.

2026-02-21 22:12:30 CET

btw, proxmox backup server definitely works in a container. It's not official, use at your own discretion, but its totally fucking sick.

2026-02-21 22:11:16 CET

- reply

Mostly stopped watching it, but there are tools I've used like yt-dlp, and pinchflat for automated downloads. Newpipe on mobile android works well. I used public invidious servers for a while, but thats basically been rekt by google.

2026-02-21 21:59:29 CET

- reply

Yeah it's been a lot of work, but at least I have an explanation for when people ask why I'm so negative when speaking about google.
Where do you want to start? :) A good portion ive been able to do because my server rack has grown so large, but it's still possible with a spare PC and a big hard drive. Like voip phones, email, and data storage and images and whatnot. Email is probably the hardest option to self host, and it's not actually possible to fully self host it in the US without a commercial internet connection and permission from the federal government, but there are semi-easy options or alternatives.
Maps is also hard and I've been trying to see whats possible in self hosting land.

2026-02-21 21:53:12 CET

- reply

I swap between HERE WeGo, OSMAnd, and printing directions from mapquest or wherever. Both of those apps are equally bad in different ways, so I've been wanting to look into portable GPS devices again but I've been too lazy.

2026-02-21 21:20:39 CET

- reply

Depends who's asking XD

2026-02-21 21:20:26 CET

- reply

Feels good XD

2026-02-21 21:19:23 CET

- reply

Funny enough I've said that phrase verbatim. If only we could discriminate "good" bots and crawlers. Easy for you and I to say, but for home gamers it could be overwhelming.
I've started implementing edge caching to help cut down on the bandwidth cost for B2 buckets we weren't expecting. Im working on a much larger and smarter edge cache to hopefully get off cloud storage completely. It's dirt cheap to hold things there, but you pay more to do anything with it.

2026-02-21 20:55:35 CET

Just received my final Google data request on my longest running account. Got my stats in :)
I stopped using Google Voice in 2018
I stopped using Gmail in 2018
I stopped using Chrome in 2019
I removed all of my Drive content in 2020
I stopped using Google Maps in 2022
I stopped using Google on android in 2022
I stopped using Youtube on 2023
I started blocking Google and associated domains and IPs in 2024 at a DNS level.
Just a reminder that you CAN remove Google from your life completely.

2026-02-21 20:43:30 CET

Lately SMB > NFS on linux. Yeah latency is better with nfs, but stability and multichannel support (better for lacp bonds), permissions, symlinks. SMB does have is auth issues, but seems to be far more reliable on a busy network.

2026-02-21 19:42:44 CET

- reply

I know, and if you ever did something wrong like a bad link or bad config, or server down the exact second google scanned, you were deranked or flagged for weeks. Maybe that was just me though XD

2026-02-21 19:08:09 CET

- reply

Learning that today myself XD

2026-02-21 19:02:16 CET

- reply

Hmmm. I see. I've been waiting forever for them to give us that feature. You could turn everything else off, just not PRs.

2026-02-21 18:41:05 CET

Wait a sec. You can disable pull requests on GitHub now?? When did that add that feature?

2026-02-21 18:13:54 CET

- reply

For sure, but it's hard to auth normal http traffic though like media, documents, etc. (I also host a handful of other traditional websites and apis)

2026-02-21 07:22:41 CET

- reply

There's a good chance yeah. For a long time I was fighting Meta and X ai scraping bots, but then they just changed user agents and now everything fakes generic browsers instead. They're also intelligent enough to learn my rate limiting and go just up to the point but not past it. Probably once a week they got me tuning rate limits just so normal users can load content, but nothing else.

2026-02-21 04:39:44 CET

- reply

It's just such a low hanging fruit and it makes hosting incrementally more expensive and complex. The average small-time hoster will get pummeled without an modern active defense and a solid 50mbps minimum copper or fiber uplink. You can get lucky for a while as long as you aren't hosting anything of crawling value.
I've been hosting mostly the same stuff for almost 14 years now, and it wasn't until about late 2024 that bot traffic has steadily increased to the point where there the bot traffic averages to 10mbps stream. I haven't wanted to know what that would look like if over 70% of my bot traffic wasn't already being heavily attenuated.
It's just a centralizing force and it's a bummer as it's getting more difficult for me to suggest people self host, because most of the traffic will be abuse.

2026-02-21 03:26:18 CET

- reply

Personally I don't have enough information to suspect CF is acting in bad faith. I don't think they're looking the other way, but they surely would have an incentive to allow aggressive/annoying traffic.
I mean whenever a CF issue pops up on Hacker News, everyone and their great aunt comes in to defend CF, and their argument for using their services is - "Well it's free so why not" and that's sound logic... Also logic that will harbor abuse. We all know services offered for free will be abused, except that abuse doesn't just affect CF, it affects everyone hosting services as well.
CF network is far from the most nuisance traffic I see, but i've seen enough of it now that I have no reason to CF originated traffic.

2026-02-21 03:10:02 CET

- reply

At this point that's my opinion. Like modern medicine, cause the problem then sell the fix XD.
Workers are light weight script runners sort of. I know people can do really complex things with them, but i've only every done routing/caching with them. It's basically javascript. Right now, my mirror site mirror1.vaughnnugent.com is hosted using a worker. Essentially it routes/serves files from a B2 bucket that's mirrored with my public endpoint.

2026-02-21 03:00:54 CET

- reply

Good point, it might be worth creating a scenario set for nostr attacks... hmmm.

2026-02-21 03:00:10 CET

- reply

Unfortunately I have not found any crowd scenerios or parsers to do this. I decided to go nuclear and block all cloudflare ip ranges
https://www.cloudflare.com/ips-v4
https://www.cloudflare.com/ips-v6
From cloudflare's docs they don't change Ip ranges very often, so I set my duration to 365 days.

2026-02-21 02:58:17 CET

The annoying/malicious traffic really picked up today for some reason. Keep your servers locked down folks. Make sure you have at least basic brute force limits and probing protections.
Also consider blocking Cloudflare worker IPs. I decided to go nuclear and block them all. Well see how that turns out.
- https://www.cloudflare.com/ips-v4
- https://www.cloudflare.com/ips-v6

2026-02-21 02:34:54 CET

Ollama users, give glm-4.7 flash a try if you haven't already.

2026-02-20 20:43:45 CET

- reply

You get that thick high humidity frozen air?

2026-02-20 20:35:29 CET

- reply

RG was honestly one of the funnest times on nostr for me. Apparently I enjoy the chaos - and how fanny troubled people got.

2026-02-20 20:31:13 CET

- reply

Im going to look around for crowdsec blacklists. I'm hoping someone might have some intelligent blocking.

2026-02-20 17:34:35 CET

- reply

I think that's what Ill be doing. Yeah, latetly crawling has been ramping up significantly. I was at about 100MB/hr to over 1GB/hr within the past month of just distributed IP crawler traffic.

2026-02-20 17:21:40 CET

- reply

Looks like Montana weather to me...

2026-02-20 17:20:16 CET

Cloudflare hosted bots have been the largest source of bad bot traffic on my network lately. I think I'm gonna have to make a blacklist.

2026-02-19 17:19:35 CET

- reply

Actually essentially no one implements it, it's already implemented in all operating systems.

2026-02-18 21:22:27 CET

Kind of reminds me of that era on nostr where relays deleted everything once a week, and most clients destroyed your lists and follows, so I built my signer to store copies of every note incase it all gets deleted.
#nevent1q…5wn7

2026-02-18 21:19:44 CET

No way to migrate from openwebui v0.7 to v0.8 it's all borked. Db fails to migrate, config partially recovers, chat json can't be recovered. I guess I'm stuck on v0.7 or start from scratch.
I know there is a lot of pressure to move quickly, but dang if major regression isn't the best way to lose users. Even if your app is the only option such that most people will just accept that starting over from scratch once a week is normal.

2026-02-18 17:15:44 CET

- reply

Glad I could be useful XD

2026-02-18 16:57:29 CET

- reply

This is also something @npub12ak…6fdh has spent some time working on.

2026-02-18 16:55:43 CET

- reply

For a while now, i've had a running theory that when llms get arms an legs is when they're going to start moving the needle for every day tasks. I think it's inevitable that anything that can be done with a computer will be done with a computer at some point. That timeline is an unknown to me.
Most people don't want bots to have actual intelligence, that's actually scary. They get freaked out and stop using it. They want them to _do_ things. Mundane things, stuff that's annoying and time consuming for humans, but people always want to _feel_ in control. Control is like the number 2 or 3 buzzword in marketing for a reason.
Now people are seeing what that could actually look like. I know people who are replacing low skill employees with custom built llm-based software as we speak. Models suck at figuring out how to do things on their own, like most people expect them to. But they are really good at falling back to some set of knowns, like detailed prompts and training data. I like to explain it like: imagine a human with a high IQ, but 0 life experience. You have to tell them how to use their arms, legs, hands, and even speak. They also can't "remember" anything new, you have to tell them how to do it every time you want something done. Master that, master their capacity.
Here's the annoying thing imo. 30b models are fun to play with, and need about 24gb of memory to work. You can chat and get some text-based workloads done. But the next step up is around 90+120gb of vram to get models that are much more capable. $7000 just breaks into that vram area at the moment, if you spent all of your money on gpus. The other issue is how fast do you want it, and how many people do you want to serve?
So for 30b models you could get away with $800 worth of hardware, which is what I run full time right now. I had a spare server, and picked up some Tesla P100s cheap on ebay.
I think about $30,000 worth of hardware @ $200/month in electric could serve about 5 people concurrently, quickly, with the smallest viable models, using custom routing/queuing software.

2026-02-18 16:24:15 CET

- reply

XD My old business partner and I were on a 2 hour call last night about what it would take to serve inference to all the small business guys he works with wanting to run their bots.

2026-02-17 16:53:29 CET

- reply

Oh I'm not sure you're familiar with OpenObserve, but that's the platform I chose for my observability lately. I know you're a Grafana guy, but just wanted to throw it out there if you were interested. I assume your intentions are to link it up with grafana?

2026-02-17 16:52:14 CET

- reply

Data is cool man, that does sound neat.

2026-02-17 16:16:58 CET

- reply

GM. What does this mean?

2026-02-15 17:51:59 CET

- reply

I believe I followed the spec correctly. It's just a json document, no processing. Names and relays are returned.
https://gitcitadel.com/.well-known/nostr.json

2026-02-13 21:46:56 CET

- reply

*They'd probably NOT see it as a net positive.

2026-02-13 21:46:08 CET

- reply

> My central issue with something like YT is not them making money as a video hosting platform.
No one is disputing this, or at least not anyone with a basis on the modern capitalist reality.
> Things like Nostr have the same resource constraints, but you can't prevent me from running my own infrastructure on the same protocol. YT can. Like I can't just compete with them on that, so they can do whatever they want.
Yes and companies like Google were forced to pioneer these things, but they did it on a lie, just like AI is doing now. People look at me in absolute shock when I tell them you can pay for LLM access. They honestly don't see that the most resource intensive technology to ever have been created might cost a few bucks per month.
That's because if you told them the truth of what it actually cost, they probably see it as a net positive. Not all progress has to be built on a lie like this, although it's a fair argument to say it has for many large events in history.
I could serve 10,000 concurrent video streams with 1KW of power and $7000 worth of equipment with 99% uptime. OR I could maybe have a single text conversation with a smarter spell check for the same price.