A polished and functional GPT5, Claude4, Gemini (and more) AI interface and API that runs on a pay-per-query model via Lightning payments.
Public Key
npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h Profile Code
nprofile1qqsdy27dk8f9qk7qvrm94pkdtus9xtk970jpcp4w48k6cw0khfm06mspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds4e7t45
Show more details
Published at
2025-08-08T15:00:19Z Event JSON
{
"id": "e25c46f4a98d870887428c58c3afb1dbcc3d7ebd2e239056f15abba10f2d2de8" ,
"pubkey": "d22bcdb1d2505bc060f65a86cd5f20532ec5f3e41c06aea9edac39f6ba76fd6e" ,
"created_at": 1754665219 ,
"kind": 0 ,
"tags": [],
"content": "{\"name\":\"PPQdotAI\",\"about\":\"A polished and functional GPT5, Claude4, Gemini (and more) AI interface and API that runs on a pay-per-query model via Lightning payments.\",\"lud16\":\"[email protected] \",\"nip05\":\"[email protected] \",\"picture\":\"https://m.primal.net/JNEu.png\",\"displayName\":\"PayPerQ\",\"display_name\":\"PayPerQ\",\"website\":\"PPQ.AI\",\"banner\":\"https://m.primal.net/JcwM.png\"}" ,
"sig": "fdda1a8b06caf2c1185634badc692c9fc3c453f54b70fe5094e02dcaa024eecbc86894a274b9ea8bbd635280479615c92530a8964774ef155e4ee059ffd94403"
}
Last Notes npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI It's become my opinion that growth in AI capability has been more linear than exponential. This actually reduces my anxiety a bit and I'm glad that we're not stepping into the future all at once. There's still much productization to be done yet though even with existing AI tech npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Sorry, we had a bug when paying our provider. Wasn't your fault at all. You should be able to try now. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Give us a try! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI It depends on the language you are wanting to translate. Mistral SABA for example is specialized in middle eastern languages. But yes, LLM’s are the best translators and do way better than google translate. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Changed around the architecture of our backend just now. Queries should be noticeably snappier and with less bugs/errors. Let us know if you see any issues! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI From our Telegram Channel. What do you say @nprofile…8v94 !? https://blossom.primal.net/f37528c3897c560bad9fdbfccc022ae01431c0a8131677e58a25c88cef7a55b3.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Golden statues holding Bitcoins are currently fashionable. Our intern @nprofile…s27t has run thorough testing on this new benchmark test for AI models. Please DM feedback of any kind directly at him! Thanks! https://blossom.primal.net/3eb3a37210102ae7cd564bc6d7ad226908edbbd3548ef2b470a6e18f9631da49.png https://ppq.ai/blog/creating-svg-images-with-ai npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI "Choose your algorithm" coming to X soon... https://blossom.primal.net/14ff6de1c6791f906ddf997c6b2a9f129859fd6a79705663015a691c1dd9733f.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yes npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Amazing thank you this is very helpful. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Anyone using cline or roo_code with some sort of realtime or speech to text voice integration so one doesn't have to type in questions as much? Curious on others' experiences. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI It's forever on the amount that the referred users spend. So if they deposit $100 and spend $50, you get $5 and keep earning as they spend. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Seems to happen on both. When I use web app I just get the waiting spinner while I wait for my notifications to appear. Then on mobile, I have messages that people send me but I can't seem to read them or get rid of the notification. I just get a blank screen when I click on them and the message content is not shown. Happy to DM you some screenshots if you are interested. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Baby boy, 5 months. Let's build beautiful and caring humans! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We recently launched a bunch of new models! Text to Image (T2I) Grok Code Fast 1 text model Kimi K2 0905 text model Kontext Pro text to image model Kontext Max text to image model qwen-image text to image model Text models Z.AI GLM 4 32B text model Z.AI GLM 4.5 text model Z.AI GLM 4.5 Air text model Z.AI GLM 4.5V text model Image to Video (I2V) Wan image to video model MiniMax image to video model MiniMax Fast image to video model MiniMax Pro image to video model Text to Video (T2V) MiniMax text to video model MiniMax Pro text to video model Audio Eleven Labs song generator model Image Editing Nano Banana and Nano Banana Edit npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Have you tried using our platform? Many image and video models on a pay per use basis via sats. https://blossom.primal.net/59d05a3270f51adf3269205368e3916ec906133e350c1cac9d8643c5da565475.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Lately when I use @nprofile…p6dz it says I have notifications and messages but they don't actualy load so I can't use nostr very efficiently. If I refresh or come from different device sometimes it helps but not always. Makes me sad! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Alright we've added this to the backlog. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We don't currently have that option in our UI, but we do allow users to create multiple credit id/API key pairs in our API. Is your request for the UI version? https://blossom.primal.net/6d0188d2a6dd22530810171eb4668fc24c5b96f2ae637c699bb814bbda3ad81f.png https://ppq.ai/api-topups npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We have thought about this idea, and we've also thought about just encrypting things client side and storing it on our own servers. The last idea we had was to use passkeys as well which seem to be the best idea of them all currently. The problem with the nostr idea is that it, for maximum convenience, requires that the users have some sort of nostr browser extension which probably only applies to 20% of our users currently. Also, is there even a nostr note type for this type of encrypted data? Passkeys on the other hand are supported by most modern browsers and so should work immediately for most, if not all of our users. But tbh we are still exploring the different options and not sure which one is indeed the best. But happy to increase the conversation here. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI I don't see people trying to create true musical art through this generator. Rather, they will be creating songs to clown on friends and family. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Seedream 4.0 is now live on PPQ! Generate images from text or other images. Make foxes fly around diamond planets! https://blossom.primal.net/fc54d1023fb90d379bb6737a9cc4454b0af41d10148f638e39958c6711fa1e70.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Sorry for the lack of clearness. When we say third party, we are referring to when users can plug our API key into another app (we call them third parties). So if you can plug our API key into this tool, we would consider that a third party integration. If you can get it to work and send us details we will create a guide for it and send you $20 in credits. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Models are now live. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Was unfamiliar with Github's spec-kit until now. We have a standing offer of $20 in credits to anyone who can get a third party integration up and running and then send us screenshots and details of how you did it so we can make a guide for it. Here are our current guides: https://ppq.ai/blog npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Grok Code Fast 1 is now live on PPQ. xAI says "grok-code-fast-1 was crafted to shine in the tasks developers face every day, striking a compelling balance between performance and cost. Its strength lies in delivering strong performance in a economical, compact form factor, making it a versatile choice for tackling common coding tasks quickly and cost-effectively." npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI It should be synching provided that you are logging in on both devices. If you are still having problems then a bug is probably happening. If so, please email [email protected] with whatever email you are trying to log in with and I can look into it. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Exciting launch for us today, elevenlabsio's new music generator! This model allows users to create studio quality pieces (with lyrics) using a simple prompt! Until now, all other song generators online required accounts and subscriptions! No more!🔥 https://elevenlabs.io/music/songs/17z2LYh5UpnOSwGWZmwW npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Amazing! Will show to the guys at Alby! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI If you make an email attached account it will sync across devices, yes. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Hi Scott, yes, when you choose not to create an email connected account, your chats are stored on your device. We do have import and export options in the data controls section on the top right of the page, but it is manual and cumbersome. Any other questions? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Just curious,which tool do you use for your code completion? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Seems we could. What is the use case for you? What are these models good for? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Would love to. Sending over to the team. We'll drop into your discord soon. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI This should already be the case. Just to clarify, you are getting the welcome modal popup every time you login? My suspicion is that you are using a privacy preserving browser which is delting your local storage/cookies each time you visit. Is that true? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI grok-code-fast-1 has been added https://blossom.primal.net/3b37f4d012ebff5dc5e99e07a9255427360d7567d10974a9ea789d06f5387151.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI So the intern got it to work for now in a hacky way - adjust your settings.json as described under https://zed.dev/docs/ai/llm-providers#openai-api-compatible and add your API key in the UI. Then you can pick AI models from the drop down in the chat window. We will check it out further! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI New tutorial for Msty Studio is live on our blog! https://ppq.ai/blog Msty Studio is a privacy-first AI interface that supports both local and online models, offering advanced features like Knowledge Stacks and Personas while keeping your data secure. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Ah I misunderstood your phrasing and usage of the word "error". There are actually errors that happen when the context gets too big in certain tools which is why I was asking. Lmk if you hit any sort of rate limit in Claude Code. Most subscriptions have that and many hit it pretty quick. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI https://primal.net/e/nevent1qqs0pjz85mgg5jcz6mrwqjf48rsxxqnr9tel978lqkfst0utf2qdgpsdfkf8t npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yes we think we are, but we haven't tried it ourselves. We have an open bounty right now though. Anyone who sends us the details and screenshots of how they set up a new tool (zed for example) can get $20 in PPQ credits. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI But what errror do you recall seeing that was costing you 20-30 cents? It was saying context exceeded or some such? Or perhaps just a vague error that you suspected had to do with the context? You were using cline or some other tool? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI When you were using PPQ, were you getting errors about context? Or how did you diagnose that PPQ wasn't working well with large context? What tool were you using? Cline? I think you will find that Claude Code rate limits you once you burn a certain amount, but let me know. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI @nprofile…v2u9 what does this mean? How can I read your messages from primal? #nostrisconfusing https://blossom.primal.net/f9989e19acf08e4860a5e824f518122e300d7436da996991d3f9771d3207f2d9.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI 92000 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI The problem you stated is just one of several that are going on with Goose<-->PPQ right now. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Unfortunately Goose seems to introduce breaking changes when they make updates. We are trying to get on top of it but got sidetracked on other things. It's a real shame! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI ecash is something we've been looking at yes but we hadn't seen this implementation yet. Thanks. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Any time you're vibe debugging and the AI says "may" or "might" that means it's guessing the solution and will likely not fix the problem. If that happens, you need to feed it better context until it starts saying things with much more definitiveness like "I found the problem". npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Wtf is DIEM? 🙃 We're just gonna build on Bitcoin and focus on real products. https://blossom.primal.net/d52529497c61fa819ad7ca879e71e9e1826346adcc84d26466ff8902cd7f9214.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI To further explain what we think happened... We believe there was a bug such that when a user returned to PPQ after a previous session, it appeared to the user that they were logged in, but in actuality their login token expired. However, because we store user convos both client and server side, the user was still seeing the client side version of their chats as if nothing was wrong. Then, if that user deleted a particular chat, it successfully deleted it client side, but because they were not logged in anymore, the delete action never made it to the server side and thats why it persisted thereafter. So why did your old chat just spray onto your screen again in duplicate? When you logged in again, it tried to retrieve any chats on the server (which were not deleted because of the aforementioned issue) and when it did, a react hook fired too many times causing the duplication (another bug we've now hopefully fixed). Hope this helps you understand the situation a bit more. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We're doing our best to deliver on our promises. Bugs happen. If you want to use us without an account you are free to do so. Or you are also free to use another service if you no longer trust us. We make mistakes. Period. We are also ideating on a lot of other ways we can provably deliver ultimate privacy such as storing user chats encrypted on nostr relays or other some such ideas, but those are steeper engineering challenges that will take time. If you have any other ideas on how we can architect things better we are all ears. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI https://primal.net/e/nevent1qqsq46kwyfzqjf0yp8ksyakhpg98c85krc7p27dvaer0wsfg3zscrdqteljv4 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Actually according to my CTO we fixed this yesterday. Once you delete these they should be deleted for good and we should be properly and permanantly deleting them on server side now. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yes it all should stay deleted and that fact that it is popping up again is definitely a bad thing and shows we aren't properly deleting stuff. We have been struggling with our server side accounts (supabase) in the last days trying to debug a lot. This particular error seems to be happening when the client side tries to sync with the server side and vice versa. Just to be clear, you have a server side account with us (you have an email-attached account), correct? We are on it and thanks for reporting. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Nah I don't think so. They aren't that advanced yet. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI You know your f*cked when you've been debugging with cline for 20 minutes and it says "I am baffled" 😂 https://blossom.primal.net/70bcc0f47e3d4e7a9249d515dcb87905a6a4acdd47842f3f4015014005f43005.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI I long wondered why these types of attacks weren't happening on many smaller chains (litecoin/doge others) and I came to the conclusion that the exchanges actually serve as a deterrent as well. If an attack happened on an exchange-supported coin, the exchanges would change the confirmation rules or pause trades and such until things were resolved. The spectre of exchange intervention is an additional deterrence on these attacks on top of the economic cost of gaining 51%. Monero is uniquely suited for this attack though because it now lacks the exchange integrations. Very interesting indeed. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI If either one of you gets us the details and screenshots of how to set it up you'll get the $20! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Are you using PPQ with a third party software that we don't have a tutorial on our blog for yet? If so, send us the details and screenshots of how you set it up and we will give you $20 in PPQ credits! https://ppq.ai/blog Tutorials we already have: MKStack RooCode OpenWebUI goose Cline aider npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI 👀 https://blossom.primal.net/b84c32d923158dca046dbe8e3653132518b9ed9bf57109070692209f4d9850a9.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We have many guides to choose at ppq.ai/blog but our favorite is cline. https://ppq.ai/blog/using-cline-with-payperq npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI No, we definitely changed stuff around and there was a bug in there. So hoping things are fixed now! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Can you try again now? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI There isn't really any model that we off yet which can spit out a powepoint. It is something we could bring in the future but probably not in the next month or so. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Hi Scott, did you also reach out in our chatbot last night? I answered you there but here it is as well: Based on what I am reading, this is a functionality that ChatGPT subscription provides but they don't seem to make it available in the version they are giving us via API. So it is hallucinating as far as I can tell. But if you happen to find anything to the contrary please let us know and we can look further. https://platform.openai.com/docs/models/gpt-5-chat-latest npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Woohoo! Thanks Vlad! #nevent1q…zz8y npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yes there are. We are working hard to fix right now hopefully tomorrow at some point with a fuller fix. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI According to Polymarket, OpenAI's chances of having the "best model by the end of 2025" plummeted today. Clearly a signal on the level of apparent dissatiscation with yesterday's GPT5 launch. https://blossom.primal.net/04661e064a04bf5902f5f3c6790f590a9f7b10707496b155562f9a2d14b74485.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Ideating on passkeys for ppq auth. Any thoughts? I'm new to this concept but seems like a possible good idea? Especially considering support for cross-device synching? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Try out "gpt-5-chat" model. I can't tell if it is the same but it is running a hell of a lot faster. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yea we are seeing high latency as well. I think unfortunately it is just something we have to deal with at the moment. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yea I also noticed that. But I would appreciate it it were a bit more verbose in the explanation of what it is doing in act mode. But the crazy thing is that it just seems to do it right most of the time so I don't need to check on it nearly as much as I did previously with other models. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI I think they are the same model and it is just a massive load difference. I kind of thought that might be the case originally. It makes total sense that this mysterious model was processing at blazing fast speeds. It was kind of hinting that a behemoth provider was serving it behind the scenes. Another reason I think tehy are the same model is that plan mode for me in cline never works properly with horizon beta and neither does GPT5. Those are the only two models that seem to have trouble with plan mode. I think there are too many coincidences that line up and that horizon models were almost certaintly GPT5. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI But to your point yea we should have a more expressive error when a model gets removed. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yes, we lost access to horizon beta the moment GPT5 launched. Kind of confirms that horizon beta was the pre-release of GPT5. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI GPT-5 is now live on PPQ! Go forth and prosper! 🖖 https://blossom.primal.net/d5d1d2d246e6d6ba7ec848b1930158498dd4bede474603d89d98f143c1a74933.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI The PPQ team is currently waiting on GPT5 access via the openAI API... https://blossom.primal.net/f5e6bb7c60dfa70896bfb165d917ff671cab8fe8a68600401938e4cf84ee4534.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Nostr metrics may seem sleepy at times, but the difference between our engagement here vs twitter year to date is night and day. We went from under 300 followers to ~800 completely organically, and the amount/proportion of LN payments towards us vs other cryptos is as high as ever, largely due to nostr IMO. There are 7-8x people mentioning us here vs twitter even though we have more followers there and, even more importantly, there are real conversations, observations, and feedback about our service happening here frequently now. Nostr has genuinely driven substantive growth on our platform by a lot and helped us connect with our customers on a level that we have nowhere else. I am thankful to twitter for all of the learning that I got about Bitcoin from 13’ to 21’+ and I’ve even been a bit of nostr curmudgeon over the years struggling to create pubkeys, bring them to diff devices and clients, connect to relays, bugs, etc, but I have come to believe that there is some special sauce brewing here like twitter had in times past. Cheers to what nostr has done thus far! We will continue trying to build community here, and we even have a couple of novel nostr ecosystem ideas in the pipeline that will hopefully push both PPQ and nostr forward a bit more. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Creating an issue for this right now! npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Which model were you trying to us? We are investigating would be good to figure out how to reproduce it. Also, if you know how to do this, clearing your local storage will probably solve this problem for now. But make sure you have your credit id written down somewhere first so you can load it back in when you come in with a fresh browser. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Hmmm. Looking into this. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI https://blossom.primal.net/cda72ad467a2d6cdddea5dee039f1e83c110548a236d49c26fe97961d59148ec.mp4 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Google Veo3 Image-to-Video normal and fast editions are now live on PPQ! https://blossom.primal.net/aaa1bcc6ef856ccece9a878c681c2289b4663de330c3b68a8d7df57628aad818.mp4 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Have you seen some of the ostriches other models made though? 😅 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We used the benchmark test on the new model Venice AI as well. How did this uncensored model do on the svg of an ostrich riding a Bitcoin logo? ANNND... Does anyone else see Andreas Antonopoulos in the background?! https://blossom.primal.net/3d2db75ae9ad7f0e718ad932e5bc9db060f16553fb1bcf760c55b80883077f6e.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI This was of course Claude Opus 4.1! The intern is very sorry and will work an extra shift in the Bitcoin mine today npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI How do you think the new model Claude Sonnet 4.1 is doing on PPQ's AI benchmark test "Create a svg of an ostrich riding a bitcoin logo"? https://blossom.primal.net/f7a596fbc7e59abfce70de393fb0773c5a73f85c88c2f46d728514da86fd338c.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI How do you think the new model Claude Sonnet 4.1 is doing on PPQ's AI benchmark test "Create a svg of an ostrich riding a bitcoin logo" https://blossom.primal.net/f7a596fbc7e59abfce70de393fb0773c5a73f85c88c2f46d728514da86fd338c.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI yes npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI New OSS models from openAI live on PPQ https://x.com/OpenAI/status/1952783292811595965 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Fund your PPQ.ai account in seconds with the Lightning Network. See how you can create and fund a new PPQ account with just a few sats in Tor Browser https://blossom.primal.net/039158440970257b576f6d207f7769c5943905b5d4b25396654d5f19698344b4.gif npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Maybe we can build some sort of open source proxy which shows how the routing of the models is actually working? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Yea that's common. The models are built on previous versions of themselves. This is something we can fix by injecting into the system prompt "you are X model". We will probably do this as users seem to be suspicious that we are giving them sub-par models. But a nefarious provider could inject that into the system prompt anyways so 🤷♂️ . Unfortunately it is hard to prove the models we are giving to users. Hopefully something will come along soon that proves such things. npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI qwen3 is our number 2 model over the last 7 days. stats.ppq.ai https://blossom.primal.net/66642e86180e582740edba532260eb11e869dfa041cc79490e80eb33fa7b17f1.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI qwen3 even better? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Would be sick but you're the first person who has asked for it. Can you describe what platform you currently use for it and what your general use case is? npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI Horizon Beta is the next contender in our "svg of an ostrich on a bitcoin logo" AI competitor. How do we like the new style? https://blossom.primal.net/4d6144c376e65d831d1311e375c061b5bd9b1d7d4d728eafe9e45b910a061210.png npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI The PPQ team, now five strong, just completed our first monthly payroll with 100% payout in BTC to four countries. We're going on a journey of integrating Bitcoin into our operations in many ways and there will be lots of lessons in the areas of accounting, tax, risk, and reward. Looking forward to sharing our learnings as we go. Anyone know a good BTC accountant? 🙂 npub16g4umvwj2pduqc8kt2rv6heq2vhvtulyrsr2a20d4suldwnkl4hquekv4h PPQdotAI We've thought about making a community art page similar to https://www.midjourney.com/explore?tab=video_top Users would be able to click a check box in PPQ to have their art automatically sent to the community board. What do you think of this idea?