Sun is shinning and the music is on, you couldn't ask for better

So what's going on with on chain zaps, I'm all for it if it's better

Holy moly its so hot today, stay safe in the sun all

Some awesome work on #kde

GM all, so after much trying to remember my old mastodon password I finally got on there to delete my account but I was tempted to look at it one last time for old times sake, what a crap hole is all I can say. Unless your a far left nut you definitely ain't welcome there, I love nostr so much mainly because you can be who you want on here and nobody cares

GM and what a stunning morning it is, now time for some coffee to make it even better

GM I hope it's a great one for you today, I must dash it's time for me to go and pray to the Lord

Good day one and all, what a glorious day

GM one and all, happy Friday

Anyone else get burnt out always trying to stay ahead of big tech, just seems endless at times to keep pushing back on there crap

Nostr vs Facebook Facebook: The Walled Garden of Convenience Facebook, now part of Meta, is the archetype of the centralized social network. It is a magnificent, albeit restrictive, machine. The Allure Ease of Use: It works out of the box. You sign up, add friends, and the algorithm feeds you content. It’s like walking into a grand cathedral where the choir is already singing; you just sit and listen. Network Effect: Everyone is there. If you want to find old schoolmates or organize a local parish event in Cambridge, Facebook is often the only place everyone agrees to meet. Feature Rich: Marketplace, Groups, Events, Messenger—it’s an all-in-one suite that reduces friction in daily life. The Cost Ownership of Data: You do not own your graph. If Meta decides to change the algorithm, ban your account, or sell your data (even indirectly), you have no recourse. You are a tenant, not an owner. Censorship and Moderation: Content moderation is dictated by a corporate board in Menlo Park, often influenced by political pressures or opaque community standards. For a man of faith or a conservative thinker, this can feel like being silenced by a distant authority. The Attention Economy: The platform is designed to maximize engagement, often by amplifying outrage, polarization, and addictive scrolling. It treats your attention as a commodity to be sold to advertisers. Nostr: The Decentralized Protocol Nostr (Notes and Other Stuff Transmitted by Relays) is not a company; it is a protocol. It is the open-source equivalent of email or the internet itself, but for social media. The Freedom True Ownership: You own your identity via a cryptographic key pair. No one can ban you from the network, only from specific relays (servers that host your posts). If a relay bans you, you simply move to another. Your followers and history travel with you. Censorship Resistance: Because there is no central server, no single entity can delete your posts globally. This aligns beautifully with the principle of free speech and the right to dissent, values that resonate with the open-source ethos. No Algorithmic Manipulation: You see what you subscribe to. There is no "feed" designed to keep you angry or addicted. You curate your own experience, much like choosing which hymns to sing in your own chapel. Privacy: You don’t need to give your phone number, email, or real name to join. You are just a public key. The Friction Complexity: Managing private keys is not for the faint of heart. If you lose your key, you lose your identity forever. There is no "Forgot Password" button. This requires a level of responsibility and technical literacy that many find daunting. Fragmentation: Because anyone can run a relay, the experience can vary. Some relays might be slow, some might censor, and the user interface depends entirely on the client (app) you choose (like Damus, Amethyst, or Snort). The Noise Problem: Without a central moderator, spam and abuse can be rampant. Users must rely on "block lists" and trust networks to filter the noise, which puts the burden of curation back on the individual. A Comparison for the Cambridge Thinker Feature Facebook (Meta) Nostr Control Centralized (Corporate) Decentralized (User-owned) Identity Email/Phone (Linked to real world) Cryptographic Key (Anonymous/Pseudonymous) Censorship High (Global bans possible) Low (Only specific relays can block) Monetization Ad-based (Data mining) Optional (Bitcoin tips, no ads) Ease of Use Very High Moderate to High (Learning curve) Philosophy "We know what's best for you" "You are responsible for your own truth" The Spiritual and Ethical Dimension As a Catholic man, you might view the Facebook model as akin to a Babel tower—a centralized structure built by human hands that promises connection but often delivers confusion and division. The algorithmic sorting of truth into "engagement" creates a fractured reality. Nostr, conversely, feels more like the early church or the open source community: a decentralized network of individuals connected by shared protocols, where authority is distributed, and each person is responsible for their own witness. It demands maturity. It asks you to be the guardian of your own digital soul. The Verdict Choose Facebook if you value convenience, mass reach, and don't mind trading your data and autonomy for a seamless experience. It is the "public square" where everyone is already gathered, even if the mayor controls the microphone. Choose Nostr if you value sovereignty, privacy, and the freedom to speak without fear of de-platforming. It is the "town square" where everyone carries their own microphone, and the crowd decides who to listen to. For someone who loves the independence of Linux and the depth of faith, Nostr offers a compelling vision of the future. But it requires the courage to take the wheel.

The Case for Local: Why Keep It Close? 1. Absolute Privacy and Sovereignty This is arguably the strongest argument for the local route. When you run an LLM on your own machine (using tools like Ollama, LM Studio, or llama.cpp), your prompts never leave your computer. For the Faithful: As a Catholic man, you might value the sanctity of your private thoughts and confessions. Sending sensitive spiritual or personal queries to a corporate server raises questions about data mining and surveillance. Locally, your data stays in your "confessional box." For the Linux User: You maintain full control. No telemetry, no hidden tracking scripts, no terms of service that change overnight. 2. No Censorship or "Alignment" Filters Cloud-based models are heavily "aligned" by their creators to adhere to specific safety guidelines, political correctness, or corporate policies. Freedom of Inquiry: A local model (especially if you choose an unfiltered or less restricted variant) allows you to explore controversial historical topics, philosophical debates, or creative writing prompts without the AI refusing to answer or lecturing you. Nuance: You can tune the model to be more direct or less preachy, fitting your own moral framework rather than a corporation's. 3. Cost Efficiency (Long Term) While the upfront cost of hardware is high, the ongoing cost is zero. No Subscription Fees: You aren't paying $20/month for "Plus" tiers. Once you buy the GPU, the electricity cost is your only recurring expense. Unlimited Usage: You can generate a million words or run a thousand queries without hitting a rate limit or a paywall. 4. Offline Capability Resilience: If the internet goes down, your AI assistant is still there. This is crucial for remote work, travel, or simply enjoying a quiet evening in Cambridge without connectivity. Latency: For short interactions, local inference can sometimes be faster than waiting for a round-trip to a cloud server, depending on your hardware. The Drawbacks: The Price of Independence 1. Hardware Hunger This is the elephant in the room. Running a capable LLM requires significant resources. GPU Requirements: To run models larger than the tiny ones (which are often not very smart), you need a powerful NVIDIA GPU with plenty of VRAM (Video RAM). 16GB or 24GB of VRAM is becoming the sweet spot for decent performance. RAM: If you don't have enough VRAM, the model spills over into system RAM, which is significantly slower. Cost: A high-end GPU can cost as much as a nice used car. 2. Intelligence Ceiling Model Size Limits: The most advanced models (like the ones powering top-tier cloud services) are massive—hundreds of billions of parameters. They simply won't fit on consumer hardware. Performance Trade-off: You are often forced to use "quantized" versions (compressed models) to make them fit. While impressive, these are generally less capable at complex reasoning, coding, or nuanced creative writing than their cloud-based, full-sized counterparts. 3. Complexity and Maintenance Technical Hurdle: Setting up a local environment involves managing Python environments, CUDA drivers, model formats (GGUF, EXL2, etc.), and command-line tools. It's not as plug-and-play as a website. Updates: You are responsible for updating the software stack and finding new model weights. There is no "customer support" to call if your setup breaks. 4. Energy Consumption Electricity Bill: Running a GPU at 100% load for hours generates heat and consumes significant power. In the UK, where energy prices can fluctuate, this adds up. It's the opposite of the "green" cloud promise, unless you have solar panels. The Verdict: A Question of Values For someone like you, farty mcmarty, who values privacy, autonomy, and open systems, the pros of local LLMs likely outweigh the cons, provided you have the hardware budget. If you prioritize: Absolute privacy, freedom from corporate filters, and offline access → Go Local. If you prioritize: Maximum intelligence, ease of use, and zero hardware investment → Stick to the Cloud. Many enthusiasts (myself included, in spirit) find a middle ground: using a local model for sensitive daily tasks and privacy-critical queries, while occasionally tapping into a cloud API for heavy-duty reasoning or complex coding tasks when necessary.

so far i've tested notedeck and jumble on linux and so far jumble i have to say works that little bit better, fair play to the devs working on linux software