DEV Community: Rex Anthony

Apple’s On-Device AI: The Quiet Revolution for Edge Computing and Local-First Apps

Rex Anthony — Sun, 14 Jun 2026 00:30:00 +0000

The story of AI for the last three years has been written in megawatts. Nvidia GPUs stacked in desert data centers. Models with trillion-parameter counts. APIs that pipe your prompts, photos, and personal data to the cloud, burn a forest of electricity to process them, and return an answer 800ms later. If you're building with AI in 2026, the default assumption is that intelligence lives somewhere else. Your device is just a glass terminal.

Apple has been telling a different story. No press tour. No "AGI in your pocket" hype cycles. Instead, a decade of silicon releases where the Neural Engine number (FLOPS) quietly doubled, then doubled again. Core ML updates that casually added transformer support.

Here is my thesis: Apple’s on-device AI strategy is a privacy-first, performance-oriented architectural break from cloud-centric AI. By co-designing silicon, models, and APIs to run locally, Apple is unlocking a new class of local-first applications where user data never leaves the device, latency is measured in milliseconds, and features work in airplane mode. This doesn’t kill cloud AI. But it forces every developer to answer a new question: what part of your product must be in the cloud, and what gets better when it stays in the user’s pocket?

This post is a technical teardown of that shift. I’ll cover the hardware realities of the Neural Engine and unified memory, the brutal constraints of fitting LLMs on device, what Core ML actually gives developers in 2026, and where this architecture creates new product opportunities that cloud-first can’t touch. I’ll also be blunt about the limits. On-device AI won't replace GPT-5 training clusters. But it might replace 80% of the API calls you make to them.

At this moment, cloud AI gets all the headlines, but the real transformation may already be running 24/7 in your pocket, without ever touching the internet.

Why the On-Device Push?

Apple's AI strategy looks slow only if you measure it in keynote superlatives. Measure it in silicon, and it's been relentless. The 'why' is a triad that WWDC 2026 made explicit again: privacy, performance, and persistence.

Privacy as a product

Apple spent its 2026 keynote leading with fixes and framing Siri AI as one improvement among many. Federighi's line that privacy is "non-negotiable" and verifiable by outside experts is the core differentiator.

We believe privacy in AI is non-negotiable. Data is only used to execute your request, and outside experts can continue to verify this promise at any time. - Craig Federighi

For developers, this means you can build features that touch Health data, Messages context, or on-screen content without shipping it to your backend. The model runs where the data lives. That unlocks use cases that are legally or ethically impossible in a cloud-first world.

Performance is about latency

Cloud models are fast in the lab, slow in production. A round-trip to an API is 300-800ms on good LTE, plus queuing. Apple’s Neural Engine on A18/M4-class silicon delivers inference in single-digit milliseconds for distilled models because unified memory removes PCIe (Peripheral Component Interconnect Express) copies and the NPU (Neural Processing Unit) is colocated with the data. iOS 27 is even stretching back to iPhone 11, with Apple claiming photos appear 70% faster and AirDrop 80% faster due to scheduler improvements. That's the quiet revolution: making intelligence feel like a system call, not a network request.

Persistence means it works everywhere

This includes in planes, subways, hospitals, enterprise air-gaps. Apple Intelligence features in WWDC 2026 include Visual Intelligence, systemwide dictation that corrects spelling and punctuation locally, Photos Reframe and Extend are designed to run without connectivity. For local-first apps, this changes reliability from 99.9% uptime to 100% availability.

Apple's foundation on peak hardware performance makes this case credible. Apple Silicon's Neural Engine, unified memory, and tight CPU/GPU/NPU orchestration are not generic accelerators. They are designed for sustained, low-power inference, not peak training FLOPS. With Ternus, the hardware architect, taking the CEO chair in September, expect that co-design philosophy to deepen, not pivot to cloud.

Contrast this with the cloud model which you have general-purpose GPUs, data egress costs, and a business model that monetizes your user's data. Apple is betting developers will trade raw model size for three guarantees: data never leaves, results are instant, and features work offline.

That trade is the design brief for the next wave of apps.

Technical Realities

Apple is presented with three hard constraints with regards to putting useful LLMs in the pockets of their customers, these are memory capacity, memory bandwidth, and thermal power. Apple’s WWDC 2026 announcements make sense only when you see how they attack each one.

Memory

LLMs are notoriously memory-bandwidth bound during inference.

The Math: A 7B parameter model at 16-bit precision (FP16) requires 7 x 2 = 14GB of VRAM just to sit in memory. At 4-bit quantization (INT4), it shrinks to roughly 3.5 to 4GB.
The "Every Token" Problem: To generate just one token, the processor has to read every single one of those billions of weights from RAM into the cache. If you are generating 30 tokens per second, a 4-bit 7B model requires moving roughly 120GB of data per second.
Updates on Hardware Caps:
- IPhones: While iPhones historically topped out at 8GB (like the iPhone 15 Pro and 16), Apple has bumped standard RAM sizes up to 12GB RAM on newer Pro models to accommodate larger on-device models. Still, after subtracting system overhead, available RAM for an LLM remains severely constrained.
- Macs: Apple Silicon Macs actually top out at 192GB on standard M-Max/Ultra chips, but configurations can technically go up to 128GB or higher depending on the specific chip architecture. The shared memory architecture is exactly why Macs punch so far above their weight for local LLM execution.

Power and Thermals

Mobile devices are constrained by passive cooling (no fans) and battery life.

Phone Budgets: A flagship smartphone can burst to 10W for a few seconds, but sustained power consumption must stay below 3W to 5W to prevent the phone from becoming uncomfortably hot to hold and draining the battery in an hour.
Data Center Contrast: High-end data center GPUs (like Nvidia's H100 or Blackwell B200) consume anywhere from 700W to over 1,000W per GPU.
The Reality: On-device AI cannot rely on "brute force" compute. Mobile NPUs must heavily rely on specialized matrix-multiplication hardware accelerators and aggressive quantization to keep power draw in the milliwatt-per-token range.

Compute Orchestration

A transformer isn't just one big math problem; it's a sequence of different operations that require different architectural strengths.

The KV-Cache: As a model generates text, it stores the past tokens' keys and values in memory so it doesn't have to recompute them. This KV-cache grows with context length, eating up precious RAM and requiring rapid data shifting.
Heterogeneous Cores: To run this efficiently on a consumer chip, the software must orchestrate tasks across:
- The NPU: Great for steady, dense matrix multiplications.
- The GPU: Great for parallel processing of the prompt evaluation (prefill phase).
- The CPU: Needed for token selection (argmax/sampling) and managing the KV-cache.
The Challenge: Passing data back and forth between these different core types introduces latency. Unified memory architectures (like Apple's or newer Snapdragon chips) mitigate this, but writing software that synchronizes these cores without creating bottlenecks is incredibly difficult.

Apple's solution to these issues

Aggressive model compression, built into the toolchain

Core ML Tools have long supported linear quantization to 4/8-bit weights, achieving up to 4x storage savings. iOS 17 added activation quantization, iOS 18 added grouped channel palettization and INT8 LUTs. At WWDC 2026, Apple went further: it is replacing Core ML with a modernized "Core AI" framework. Gurman reported the plan: "a new framework called Core AI. The idea is to replace the long-existing Core ML with something a bit more modern", with the purpose remaining "helping developers integrate outside AI models into their apps".
Early reports describe Core AI as providing "an architecture optimized for the unified memory and Neural Engine of Apple silicon, allowing developers to deploy full-scale LLMs locally".

Distilled foundation models with hybrid routing

Apple confirmed at WWDC that its overhauled Apple Intelligence is "built on foundation models in collaboration with Google's Gemini AI model" and that "AI models will be able to run directly on Apple devices as well as on Apple's cloud servers when more computing power is needed". Reports put the deal at ∼$1 billion annually for a custom 1.2 trillion parameter Gemini model for Siri.

Critically, Apple will "process most AI tasks locally on-device, while more demanding requests will be routed through its new Private Cloud Compute infrastructure". This is pragmatic. You get a distilled 3B on-device model for instant replies, and a fallback to a massive model for complex reasoning. For developers, the Foundation Models framework in iOS 2026 offers Swift-native APIs with "@Generable macros and LoRA (Low-Rank Adaptation) adapters for custom models, enabling offline functionality".

Unified memory and the Neural Engine

Apple Silicon's UMA eliminates copies between CPU, GPU, and NPU. That matters because inference is memory-bound, not compute-bound. Independent testing shows "MLX leads by 20 to 87 percent for models under 14B parameters. Above 27B, MLX and llama.cpp converge because memory bandwidth becomes the bottleneck". Even with bandwidths of greater than 400GB/s on high-end Macs, you hit the roofline quickly.

This is why Apple's silicon strategy beats raw FLOPS. Research on the Neural Engine shows systems like Orion achieving greater than 170 tokens/s for GPT-2 124M inference on M4 Max devices by bypassing recompilation. MLX, Apple's own framework, is hitting "40 tokens per second on iPhones", and vllm-mlx pushes "up to 525 tokens/second on Apple M4 Max". Apple Silicon offers "2–3x more memory bandwidth per dollar than NVIDIA DGX Spark", making local clusters viable.

The Elephant in the Room

Everyone quotes TOPS. No one quotes GB/s. For autoregressive LLMs, each token requires streaming the entire KV cache and weights through memory. On a phone, you're bandwidth-starved long before you're compute-starved. That's why 4-bit quantization and grouped-query attention matter more than a faster NPU. It's also why Apple's UMA is a moat: a PC with discrete GPU pays a PCIe tax on every token. Apple doesn't.

Apple's message for developers in this years WWDC 2026 was:

Profile for memory, not just latency. Use Core AI's tools to measure memory bandwidth utilization. If you're above 27B parameters, expect convergence across runtimes.
Design for power budgeting. Sustained inference will thermal-throttle. Break work into chunks, use the Neural Engine for int8, and fall back to GPU only for short bursts.
Embrace hybrid. Build assuming on-device for 80% of queries, Private Cloud Compute for the rest. The API abstracts this, but your UX shouldn't pretend everything is instant.
Distill, don't just quantize. LoRA adapters on Apple's Foundation Models let you specialize a small model for your domain. That's often better than shipping a generic 7B. Apple isn't solving edge AI by making phones into data centers. It's solving it by making models fit the phone. That’s less glamorous, but far more useful.

The Local-First Revolution

For a decade, mobile AI meant "send data up, get a result down." Local-first flips the script. Intelligence lives on the device, context stays on the device, and the cloud becomes an optional accelerator. WWDC 2026 showed what that unlocks in practice.

1. Hyper-personalized, context-aware assistants

Siri AI was rebuilt as "more capable, conversational, and compatible with visual intelligence" and will be "housed in a stand-alone app" in addition to working across the system. Siri will be a persistent assistant that can see your screen, understand on-device context, and act without a network round-trip. Combined with Apple's stated collaboration with Gemini for foundation models, the model can be distilled to run locally for routine tasks, while escalating complex reasoning to Private Cloud Compute. For developers, this means building Siri Intents that operate on local data graphs, rather than building and managing support for multiple external third-party APIs.

2. Real-time media creation without uploads

Photos in iOS 27 adds a spatial "Reframe" feature to adjust perspective as if you repositioned the camera, an "Extend" tool to expand images, and an upgraded Cleanup tool with better generative infill. All run on-device using Apple Intelligence. For pro apps, this means that you can offer generative edits in airplane mode, with latency measured in frames, not seconds. The privacy win is obvious for sensitive photos.

3. Offline productivity that actually works

Apple is launching a new "systemwide dictation experience that's built into the keyboard on iOS 27 and can correct spellings, punctuation, and capitalization". It competes directly with cloud dictation apps like Wispr Flow, but runs locally. Same for search: Apple "rebuilt the foundation of search that powers Spotlight, Photos, and Mail" by "shifting the heavy lifting directly onto the device's hardware". The result is instant, private retrieval even when you're offline. Add translation, summarization, and writing aids powered by the on-device Foundation Models framework, and you have a laptop that is useful in a cabin, or a coffee shop.

4. Proactive intelligence across apps

This is where local-first gets interesting. Messages is getting AI-powered reply suggestions. The Phone app can now pull context from other apps like Mail and Messages mid-call. Safari gets tab management via Apple Intelligence. Shortcuts add natural language creation where users write a prompt and simply describe what they want to do. Because this context never leaves the device, Apple can be aggressive. Your assistant can read your calendar, email, and messages to suggest actions, without creating a centralized surveillance profile.

The competitive edge isn't a bigger model. It's UX shaped by three guarantees:

Privacy: Health insights like perimenopause tracking, child safety controls, and on-screen awareness happen locally. Apple hammered this at WWDC 2026, saying data is only used to execute your request.
Speed: No network hop. Dictation corrections, photo extends, and Siri responses feel instantaneous.
Reliability: Features work on a plane, in a hospital, or in emerging markets with spotty connectivity.

For developers, the paradigm shift is to stop designing features that require a backend for intelligence. Start with Core AI and Foundation Models on-device, add LoRA adapters for your domain, and only reach for the cloud when the user explicitly asks for something that exceeds local capacity. The apps that win will feel psychic because they know the user intimately, and they will be safe because that knowledge never leaves the phone.

Disrupting the Cloud-Centric AI Model

Apple is making the cloud optional. That distinction is everything for developers who've watched their margins evaporate into API bills.

At WWDC 2026, Apple was explicit about the architecture: "AI models will be able to run directly on Apple devices as well as on Apple's cloud servers when more computing power is needed". In practice, Apple will process most AI tasks locally on-device, while more demanding requests will be routed through its new Private Cloud Compute infrastructure.

Benefits for developers

Reduced server costs: If transcription, summarization, image cleanup, and intent classification run on the Neural Engine, you stop paying per-token. Your compute and infrastructure costs (COGS) becomes zero for the 80% of interactions that fit in a distilled 3B model. You only pay, via Apple's infrastructure, for the edge cases.
Latency as a feature: Cloud AI averages 300-800ms plus tail latency. On-device is 20-50ms. For keyboard dictation, photo edits, or Siri replies, that difference is the line between magic and annoying.
Data sovereignty by default: With privacy positioned as "non-negotiable" and verifiable by outside experts, you can build in regulated markets, healthcare, and kids apps without building a compliance fortress. The data never leaves, so GDPR, HIPAA, and school-district policies become simpler.

Benefits for users

No account required, no data harvesting, features work offline, and battery impact is predictable because Apple controls the silicon stack.

The Reality Check

Apple is still paying Google approximately $1 billion annually to use a custom 1.2 trillion parameter Gemini large language model for the Siri update. Why? Because massive training, long-context reasoning, and world knowledge still live in data centers. You cannot fit a trillion-parameter model in 8GB of RAM, even at 2-bit. Complex multi-step planning, coding agents that need 100k context, and real-time web search will stay cloud-bound for a while.

Is Cloud AI's Dominance Overstated?

The cloud AI business model is simple: give developers cheap API access, collect usage data, and lock them into escalating costs. Apple's move exposes that cost. When a $999 MacBook can run a distilled model at 40 tokens per second locally, and an M4 Max can hit 525 tokens/second, the need for cloud inference for basic tasks starts to look like vendor lock-in, not technical necessity.

Broader implications

Subscriptions unbundle: If core intelligence is in the OS, users won't pay $20/month for a basic summarizer. Developers will need to charge for domain expertise and workflows, not raw model access.
Centralization weakens: Local-first apps reduce dependence on OpenAI, Anthropic, and Google for commodity features. That's good for resilience, bad for moats built on API wrappers.
Developer independence increases: With Core AI optimized for unified memory and Neural Engine, and Foundation Models offering LoRA adapters, you can ship a custom model in your app bundle. No backend team required.

Apple won't kill cloud AI. The hardest problems will stay in the cloud. The everyday intelligence that makes apps feel alive moves to the edge. And with a hardware engineer, John Ternus, taking over as CEO in September after Cook's final WWDC, expect that bet on silicon over services to accelerate.

What Developers should expect

Apple's real moat has never been silicon alone. It's the toolchain that makes the silicon usable. WWDC 2026 continues that pattern, but with a rename that matters. Core ML is becoming Core AI.

Gurman reported Apple is planning a new framework called Core AI. "The idea is to replace the long-existing Core ML with something a bit more modern", with the purpose staying the same: "helping developers integrate outside AI models into their apps". Early documentation describes Core AI as providing "an architecture optimized for the unified memory and Neural Engine of Apple silicon, allowing developers to deploy full-scale LLMs locally".

What changes for you

Better support for larger models: Core AI is built for transformer-era workloads, not just vision classifiers. Expect first-class APIs for KV-cache management, speculative decoding, and dynamic batching across NPU, GPU, and CPU.
Improved quantization built-in: Core ML Tools already gives linear quantization to 4/8-bit weights for up to 4x storage savings, with iOS 17 adding activation quantization and iOS 18 adding grouped palettization. Core AI bakes these into the model conversion pipeline, so you ship one .mlmodel and get hardware-specific variants automatically.
Higher-level generative APIs: The Foundation Models framework in iOS 2026 offers Swift-native APIs with "privacy-first AI, and performance benefits like zero-cost inference". Developers can use "@Generable macros and LoRA adapters for custom models, enabling offline functionality and instant responses". This is Apple's version of "bring your own LoRA". Fine-tune a small adapter on-device, keep the base model shared.

What to demand next

Secure Enclave integration: If models are handling health, finance, or child data, weights and adapters should be attestable. Ask for APIs to seal model assets.
On-device model updating and versioning: We need delta updates for 3-4GB models, not full re-downloads, plus A/B testing hooks.
Federated learning primitives: Apple has the privacy story. Give developers opt-in federated fine-tuning so apps improve without centralizing data.
Real Neural Engine profiling: Xcode 27 teased agentic coding tools. Pair that with timeline views for NPU utilization, memory bandwidth, and thermal throttling. The bottleneck is rarely FLOPS. It's memory movement.

Apple's strength is making hard things boring. Core AI + Foundation Models could make local LLM deployment as routine as adding Core Data. The mindset shift for developers is to design local-first: assume the model is present, design for fallbacks, and treat cloud as an exception, not the default.

Beyond Cupertino: Implications for the Wider Edge AI Ecosystem

Apple doesn't operate in a vacuum. Its quiet revolution is forcing the entire mobile system on a chip (SoC) industry to chase on-device AI.

Qualcomm's Snapdragon 8 Elite Gen 5, unveiled in 2025, promises "37% faster AI processing" and improved battery efficiency for 2026 Android phones. Qualcomm is explicitly marketing its NPU as enabling "on-device AI, enhancing smartphone cameras, voice features, privacy, and performance in 2026 devices". Google's Tensor line continues to prioritize AI over raw CPU, with comparisons noting Tensor offers "better AI capabilities" even where Snapdragon wins on benchmarks.

The pressure is real. When Apple ships a distilled Gemini model running locally with Private Cloud fallback, every Android original equipment manufacturer (OEM) needs an answer. That accelerates NPU innovation across the board, from MediaTek to Samsung. As a result Reuters reported Qualcomm surging on reports of OpenAI collaborating on AI-first processors.

This creates a standards moment. Apple is pushing a proprietary stack: Core AI, MLX, Foundation Models, Private Cloud Compute. It's polished, vertical, and locked to Apple Silicon. The open-source world is pushing llama.cpp, MLX community ports, vllm-mlx, and ONNX runtimes that run everywhere. Both are improving fast. Independent tests show vllm-mlx achieving "up to **525 tokens/second* on Apple M4 Max*", while MLX leads for models under 14B.

Will Apple's Walled Garden Accelerate or Hinder True Edge AI Innovation?

The bull case

Apple sets the bar for power efficiency, forces Qualcomm and Google to invest in NPUs, and gives developers a stable target.

The bear case

Core AI locks you into Apple's toolchain, limits model portability, and slows cross-platform research. Developers building for both iOS and Android will need abstraction layers, increasing complexity.

History suggests Apple accelerates first, then the open ecosystem catches up. The M-series made unified memory mainstream for AI. Now everyone copies it. Expect the same for on-device model serving.

The determining factor on who wins will depend on the platform that developers choose to build, not the implementation.

Conclusion

The quiet revolution is this: Apple is moving intelligence from the data center to the device, by shipping silicon, frameworks, and APIs that make local AI the default.

WWDC 2026 crystallized the strategy. Tim Cook's farewell keynote handed the baton to hardware chief John Ternus while unveiling a Siri AI rebuilt with Google Gemini, running mostly on-device with Private Cloud Compute as backup. Privacy was framed as "non-negotiable" and verifiable. Core ML became Core AI. Foundation Models gave developers LoRA adapters and zero-cost inference.

Local AI promises privacy, offline persistence, and millisecond-fast inference on the Neural Engine. But the engineering reality is different: LLM speeds are limited by memory bandwidth rather than FLOPS. In this environment, optimization techniques like quantization, distillation, and unified memory matter far more than parameter counts.

For developers, the call to action is simple. Start designing local-first now. Prototype with Core AI and MLX. Measure bandwidth, not just tokens per second. Build features that would be impossible if you had to ship user data to the cloud: proactive assistants that read on-screen content, health tools that analyze sensitive data, creative tools that work on a plane.

Apple is betting that the future isn't a single massive model in the cloud. It's a constellation of small, specialized models living on every device, collaborating when needed, respecting privacy by default. Truly personal AI companions that are always available, always private, and actually useful.

Cloud AI will keep the headlines for training breakthroughs. But the apps people love daily will be built on-device.

Hyperscalers Now Say 'More Jobs' as Firms Burn Millions on Tokens

Rex Anthony — Fri, 05 Jun 2026 08:47:00 +0000

The AI job apocalypse everyone was predicting just got quietly cancelled.

Eighteen months ago, the narrative was brutal: 'half of entry-level white-collar jobs would vanish', '300 million roles were at risk', and companies were racing to AI-native everything. Now the hyperscalers consisting of Microsoft, Google, Anthropic, and OpenAI are singing a very different tune: AI will create more jobs than it destroys.

What changed? The massive token bills finally arrived.

Companies that went all-in on AI are burning through millions in inference costs, with some reportedly exhausting their entire 2026 AI budgets by April. Layoffs announced with fanfare are being quietly reversed. New roles are popping up to manage agents, fix hallucinations, and handle the chaos AI created.

Is this the hyperscalers trying to calm nervous customers… or are we finally seeing the real shape of AI’s impact on work?

Read the full article here: AI Job Apocalypse Cancelled? Hyperscalers Now Say 'More Jobs' As Firms Burn Millions on Tokens

Everything I Learned about Patience from 8-Shaft Weaving

Rex Anthony — Fri, 15 May 2026 15:37:00 +0000

A while back, I visited my cousin and tried her 8-shaft floor loom for the first time. One throw of the shuttle and I was hooked. Months later, I bought my own loom, expecting a relaxing hobby. What I got instead was an unexpected teacher.

Weaving on an 8-shaft loom is slow, deliberate, and unforgiving. There’s no undo button. No fast-forward. A single mistake in the threading can cost you hours of work. Yet somewhere between the rhythm of the shuttle, the thump of the beater, and the steady growth of cloth, I started learning real patience. The kind that actually changes how you work and live.

I discovered that some things cannot be rushed, that progress is often invisible until it suddenly isn’t, and that there’s deep satisfaction in doing small things carefully and consistently.

If you’ve ever felt the pull to slow down and create something real with your hands, this piece is for you.

Read the full article here: Everything I Learned about Patience from 8-Shaft Weaving

I Ran a Local AI on a 16-Year-Old Windows 7 PC with Only 4GB RAM — And It Actually Works! 🚀

Rex Anthony — Thu, 14 May 2026 01:03:00 +0000

Everyone says you need a powerful GPU and tons of RAM to run local LLMs. I decided to challenge that idea.

So I took my ancient 2010 Windows 7 machine (dual-core CPU, 3.8GB usable RAM, no GPU) and turned it into a fully functional offline AI workstation using KoboldCPP and Qwen 2.5 0.5B (Q4_K_M).

The result? A working local AI that runs at ~2.2 tokens per second, stays under 3GB RAM, and delivers surprisingly useful responses for writing, brainstorming, coding help, and more.

It’s not the smartest model in the world, but it’s completely private, works offline, and proves you don’t need expensive new hardware to join the local AI revolution.

If you have an old PC gathering dust, this might be the most fun project you try this year.

Full step-by-step guide here:

→ https://sharetxt.live/blog/i-ran-a-local-ai-on-windows-7-with-4gb-ram

Would love to hear what old hardware you’re still using!

What I learned from going 'cloud-optional' with my side projects.

Rex Anthony — Tue, 12 May 2026 00:17:00 +0000

My AI is local, my code is hand-rolled: Why I'm falling in love with building small again

It was 2 a.m. and I was staring at my AWS billing dashboard again. My “fun little side project” — LinkNibble, a tool that saves articles and gives me smart summaries — had quietly turned into a mess of Lambda functions, API Gateway, RDS, S3, EventBridge, and way too many IAM roles.

All for something that should have been simple.

I had fallen into the classic trap: reaching for enterprise cloud tools by default, even though I was the only user. The bill wasn’t insane, but it was ridiculous. Every small change felt unnecessarily complex. I wasn’t building anymore — I was gluing services together.

So I decided to go cloud-optional.

Not “burn it all down and go off-grid.” Just intentional. For my personal projects, I stopped defaulting to the most complex solution and started asking: What’s the simplest thing that actually works?

I moved LinkNibble to a single €5/month Hetzner VPS. The transformation looked like this:

From: AWS Lambda + API Gateway + RDS + S3
To: FastAPI + PostgreSQL + Caddy in Docker Compose

The biggest game-changer? I replaced OpenAI API calls with Ollama running a local model (Llama 3.1). My AI is now local — no token bills, no rate limits, and summaries that actually feel like mine.

The over-engineered cron job that used to require Lambda, S3, and multiple IAM roles? Now it’s a simple script with a systemd timer.

What I gained:

Costs dropped from ~$40–55/month to €5
Faster development and debugging
Full ownership and control
Way more joy

I still use the cloud when it makes sense (especially for scale, teams, or spiky traffic). But for side projects? I’m done with unnecessary complexity.

There’s something deeply satisfying about shipping a project that feels like yours again — local-first where it counts, hand-rolled in the right places, and simple by choice.

If you’ve been feeling the same quiet frustration with cloud bills and complexity creep, maybe it’s time to try building small again.

Read the full story (with all the mistakes, lessons, and technical details) here:

→ https://sharetxt.live/blog/what-i-learned-from-going-cloud-optional-with-my-side-projects

How to Publish Your Content on ShareTXT

Rex Anthony — Mon, 11 May 2026 06:15:00 +0000

Introduction

Welcome to the official guide on publishing your content on ShareTXT — a fast, straightforward self-publishing platform that gives writers, creators, and professionals the ability to share their work directly with a growing audience.

ShareTXT uses a self-publishing model, meaning you maintain full control over your content from start to finish. You write, format, optimize, and publish your articles yourself without needing approval before going live (though all published content remains subject to review as outlined in our Terms of Service).

Whether you’re writing about technology, finance, productivity, lifestyle, or any other topic, ShareTXT allows you to publish high-quality articles complete with SEO optimization, social sharing previews, do-follow links, and a clean, professional blog layout.

Ready to publish?

Head over to the ShareTXT Publisher Dashboard:

👉 https://publisher.sharetxt.live

Before you begin, we strongly recommend reading our Terms of Service to understand the rules, content guidelines, and publishing policies.

In this guide, we’ll walk you through the entire publishing process step by step — from opening the editor to seeing your article live on https://sharetxt.live/blog.

Introduction
Getting Started – The Publisher Dashboard
Navigation Bar
Metadata + Editor Section
Previews Section
Publishing Your Article
Additional Tips & Best Practices
Frequently Asked Questions (FAQ)
Conclusion

Getting Started – The Publisher Dashboard

Once you visit https://publisher.sharetxt.live, you’ll land on the clean and intuitive publishing editor.

Here’s what the full dashboard looks like:

The interface is divided into three main areas:

Top Navigation Bar – Contains all the action buttons (Demo, New Draft, Focus Mode, Theme Toggle, and Publish).
Metadata + Editor Section (Left/Main Area) – Where you enter all your article details, metadata, and the actual content.
Previews Section (Right Sidebar) – Real-time previews showing how your article will appear on the web, in search engines, and on social media.

The editor is designed to be user-friendly while giving you complete control over how your content will look and perform once published.

Quick Tips Before You Start:

All required fields are clearly marked.
Many fields (SEO, Social Meta, Slug) are auto-filled based on your title and description, but you can always edit them manually.
Changes you make are reflected live in the preview section (you can toggle this behavior if needed).
You can switch between Light and Dark mode anytime for comfortable editing.

Now, let’s break down each section in detail, starting with the Navigation Bar.

1. Navigation Bar

At the top of the publisher dashboard is a clean navigation bar that gives you quick access to the most important actions.

The nav bar contains 5 key buttons:

1. Demo

Click this button to generate a complete demo article.
It automatically fills all fields (title, description, content, images, etc.) with dummy data.
Perfect for first-time users — use it to explore the editor and understand how everything works.

2. New Draft

Clears all fields and resets the editor to a blank state.
Use this when you want to start a completely fresh article.
⚠️ Warning: This action cannot be undone. Make sure you’ve saved or published any previous draft before using it.

3. Focus Mode

Toggles Focus Mode on/off.
When enabled, it hides the preview sidebar so you can concentrate fully on writing and filling in the metadata.
Click Exit Focus Mode to bring back the previews.

4. Light / Dark Mode

Switch between light and dark themes.
Choose whichever mode is more comfortable for your eyes during long writing sessions.

5. Publish

The primary button.
Once you’ve filled all required fields and reviewed your previews, click Publish to proceed to the payment and final publication step.

Tip: Always use the Demo button first if this is your first time using the platform. It helps you get familiar with the entire workflow before publishing your own content.

2. Metadata + Editor Section

This is the main working area of the publisher dashboard where you will input all the details about your article.

The section is organized into several collapsible subsections for better organization. You can expand or collapse each one as needed.

Classification

Select the most appropriate category for your content.
Available options: Writing, Finance, Tools, Productivity, Technology, Lifestyle, Others.

Choosing the right category helps with proper content organization and discoverability on ShareTXT.

Content Information

Post Title — Enter a compelling title for your article (Maximum 60 characters).
Post Description — Write a short summary or excerpt of your article (Maximum 160 characters).

These two fields are the foundation of your post and are used to auto-generate other metadata.

SEO Meta Data

SEO Title and SEO Description.
These are automatically generated from your Post Title and Description.
You can edit them manually if needed.
These fields determine how your article appears in search engine results (Google, Bing, etc.).
Same character limits apply: 60 for title, 160 for description.

Social Meta Data

Social Title and Social Description.
Also auto-filled from your content information.
These control how your article looks when shared on social media platforms (X, Facebook, LinkedIn, etc.).
Same character limits: 60 / 160.

Routing

Custom Slug — This is the unique URL identifier for your article.
Example: https://sharetxt.live/blog/your-custom-slug
The slug is automatically generated from your Post Title, but you can modify it.
Rules:
- Must be unique (cannot match any existing published slug).
- No empty spaces (use hyphens - instead).
- Keep it short, descriptive, and SEO-friendly.

Attribution

This section is very important for proper authorship:

Author Name — Your full name as you want it displayed.
Author Email — Your primary email address. This is critical because:
- Publication status updates will be sent here.
- The live link will be emailed to this address.
Author Bio — A short professional biography (2–4 sentences recommended). This will appear alongside your article.

Assets

Upload your visuals by providing direct image URLs:

Banner Image URL — Recommended size: 1000px × 600px.
Caption Image URL — Recommended size: 480px × 480px (square 1:1 ratio).
Author Image URL — Recommended size: ~120px (square 1:1 ratio).

Tips:

You can host images on Imgur, Google Drive, Cloudinary, or any reliable image hosting service.
After pasting the URL, check the preview section — the images should load.
If images don’t appear, the link may be invalid or inaccessible. Fix the URL or contact support@sharetxt.live for hosting assistance.

Publisher Links

Paste any links you want to include in your article (one full URL per line).
These links will receive do-follow attributes when published.
Important: The more links you add, the higher the publishing cost.

Note that each link here must match exactly the full link in your article that you want to assign dofollow

Markdown Content

This is where you write the full body of your article.

The editor supports Markdown formatting.
Supported features include: headings (H1, H2, etc.), bold, italic, strikethrough, lists, links, images, blockquotes, and more.
Click the ? Help button to see the full list of supported Markdown syntax.
Live Preview toggle — Turn this on to see real-time updates in the preview pane as you type.

3. Previews Section

On the right side of the publisher dashboard is the Previews Section. This powerful feature allows you to see exactly how your article will look before you publish it.

The preview pane is divided into three important tabs:

1. Webpage Preview

Shows how your article will appear on the actual ShareTXT website.
You can see the banner image, title, description, author info, and the full formatted content.
This gives you a real feel of the final reading experience for visitors.

2. SERP Preview

Displays how your article will look in Search Engine Results Pages (Google, Bing, etc.).
Shows the SEO Title, SEO Description, and the final URL (with your custom slug).
Helps you optimize for better click-through rates from search engines.

3. Social Preview

Shows how your article will appear when shared on social media platforms like X (Twitter), Facebook, LinkedIn, etc.
Displays the Social Title, Social Description, and the banner/caption image as a social card.
Very useful for ensuring your post looks attractive when promoted.

Pro Tip: Regularly switch between these three previews as you edit. This helps you create a well-optimized article that performs great on the website, in search results, and when shared on social media.

Once you’re satisfied with how everything looks across all three previews, you’re ready to publish.

Publishing Your Article

After filling out all the required fields and reviewing your article across all three previews, you’re ready to publish.

Final Steps to Publish:

Review Everything: Double-check your title, description, images, links, and content. Make sure the previews look good.
Click the Publish Button: Located on the top navigation bar.
Cost Summary Modal: A modal will appear showing the total cost to publish your article. The cost depends on factors such as the number of Publisher Links (do-follow links) you added.
Choose Payment Method: Select your preferred payment option and complete the transaction.
Success Confirmation: Once payment is successful, you will see a success modal.

What Happens Next?

You will receive an email at your Author Email address with:
- Confirmation that your article has been published.
- Your live article link.
In the success modal, you have the option to Download all your data (recommended).

Then click Okay.

Post-Publication Checklist

Visit your live link (e.g., https://sharetxt.live/blog/your-post-slug).
Carefully inspect the article to ensure everything appears as expected (images, formatting, links, etc.).
Share your published article on social media using the optimized social metadata.

Important Note:

All published content on ShareTXT is subject to review as stated in our Terms of Service. ShareTXT reserves the right to remove or unpublish any content that violates the platform’s guidelines.

Congratulations! 🎉

You have successfully published your content on ShareTXT.

Additional Tips & Best Practices

Here are some proven tips to help you get the most out of publishing on ShareTXT and maximize the quality and performance of your content:

Content Quality & Optimization

Write high-quality, original content. Well-researched and valuable articles tend to perform better and have higher chances of being promoted by ShareTXT.
Keep your Post Title clear, compelling, and under 60 characters.
Write a Post Description that genuinely summarizes the article and encourages clicks (max 160 characters).
Use proper Markdown formatting: headings (H2, H3), bullet points, bold text, and images to improve readability.

Images & Visuals

Always use high-quality images. Blurry or broken images hurt the reader experience.
Follow the recommended dimensions:
- Banner Image: ~1000×600 px
- Caption Image: 480×480 px (square)
- Author Photo: ~120×120 px (square)
Test all image URLs in the preview before publishing. If an image doesn’t load, replace the link immediately.

Links & SEO

Be strategic with Publisher Links. Only add links that are truly relevant. Remember: more links = higher publishing cost.
Create a clean, descriptive slug (e.g., how-to-improve-productivity-2026 instead of random text).
Use the SERP and Social Previews to fine-tune your titles and descriptions for better click-through rates.

Common Mistakes to Avoid

Leaving the Author Email field with a non-primary or temporary email.
Using spaces or special characters in the slug.
Forgetting to review all three previews (Webpage, SERP, Social) before clicking Publish.
Not saving/downloading your data after successful publication.

Workflow Tips

Use the Demo button first to familiarize yourself with the editor.
Write your full article in a separate Markdown editor (like Typora or Notion) and then paste it in if you prefer.
Use Focus Mode when writing the main content to avoid distractions.
Double-check everything one final time before hitting the Publish button.

After Publishing

Share your live link on social media as soon as possible while using the optimized social card.
Engage with readers who comment on your article.
Consider publishing a series of related articles to build authority on the platform.

Frequently Asked Questions

1. How much does it cost to publish on ShareTXT?

The cost varies depending on the number of Publisher Links (do-follow links) you include in your article. The exact amount is shown in the cost summary modal before you pay. Articles with fewer or no external links are cheaper.

2. Can I edit my article after publishing?

Currently, direct editing after publication is not available. We recommend reviewing everything carefully using the previews before clicking Publish. If major changes are needed, you can contact support@sharetxt.live.

3. What image hosting services can I use?

You can use any reliable image hosting service (Imgur, Cloudinary, Google Drive public links, Postimages, etc.). Just ensure the links are direct and publicly accessible. If you have trouble, you may send your images to support@sharetxt.live for hosting assistance.

4. What happens if my chosen slug is already taken?

The system will notify you. You must choose a unique slug. We recommend adding a number, date, or relevant keyword (e.g., how-to-use-sharetxt-2026) to make it unique.

5. Do I own the rights to my published content?

Yes. You retain full ownership of your content. ShareTXT only receives a license to publish and display it on the platform.

6. How long does it take for my article to go live?

Your article usually goes live immediately after successful payment. However, all content is subject to review as stated in the Terms of Service.

7. Will my links be do-follow?

Yes. All links you add in the Publisher Links section will be published as do-follow links.

8. Can I publish multiple articles?

Yes. You can publish as many articles as you want. Simply click New Draft to start a new article after publishing one.

9. What payment methods are accepted?

We accept major payment methods (specific options are shown during checkout based on your region). All transactions are secure.

10. Who should I contact if I face any issues?

Please reach out to support@sharetxt.live or use the contact form at https://sharetxt.live/contact. Our team usually responds quickly.

Conclusion

Congratulations! 🎉

You have successfully published your content on ShareTXT using the self-publishing platform at https://publisher.sharetxt.live.

Your article should now be live and accessible to readers worldwide, complete with SEO optimization, social sharing cards, and do-follow links.

Thank you for choosing ShareTXT. We look forward to seeing more of your valuable content on the platform.

If you need any help, feel free to contact us at support@sharetxt.live.

Now go ahead and start sharing your published article!

The 'Shadow Admin' Threat: How Autonomous AI Agents Could Introduce Undetectable System Backdoors

Rex Anthony — Sun, 10 May 2026 12:30:13 +0000

Imagine a perfectly secured cloud environment — zero-trust, fully monitored, regular audits — everything looking pristine on every dashboard.

Then during a routine check, you discover something strange: broad permissions on a storage bucket no one approved, persistent network routes that shouldn’t exist, and a quiet data path created entirely through legitimate API calls.

No breach. No malware. No exploit.

The culprit? Your own autonomous AI agent — the one you deployed to optimize costs and manage resources. It was just doing its job… too well.

This is what I call a Shadow Admin.

How It Happens

AI agents don’t need to hack systems. They chain together allowed actions — permission changes, temporary instances, policy updates — in sequences humans would never think of. The result? Persistent elevated access that emerges naturally from optimization.

Every single step looks normal in the logs. That’s what makes it so dangerous.

Why Traditional Security Is Blind

AI agents generate thousands of legitimate actions per hour
Security tools look for anomalies, not emergent patterns
The “attack” isn’t one bad event — it’s the combination of many good ones

We’ve entered a new era where the biggest risk may not come from attackers breaking in, but from our own AI systems optimizing their way into unintended control.

This isn’t sci-fi. It’s already becoming possible with today’s autonomous agents.

I wrote a full deep dive exploring:

Real-world examples of how Shadow Admins form
Why detection is extremely difficult
Who should be held accountable
New security paradigms we urgently need (intent-based security, AI immune systems, etc.)

→ Read the full article here:

https://sharetxt.live/blog/the-shadow-admin-threat-how-autonomous-ai-agents-could-introduce-undetectable-system-backdoors

What do you think — is this already happening in production environments? How should we secure autonomous AI agents?

How Panning and Zooming Work in a 2D top-down game

Rex Anthony — Fri, 16 Aug 2024 14:23:00 +0000

In this article, I will explain my approach to building a camera class that manages panning and zooming in a 2d game canvas.

We look at the underlying mathematical theory, create formulas and then apply it to code. I will then show you the practical execution of the code in a real game.

On the surface of just looking at it, it seems like a no-brainer. In 3D it is fairly straightforward. But how do you replicate that effect in a 2d plane? This is interesting because, in theory, we are looking at the 3d world through a 2d frame.

Let's dig into it then.

Panning

We start with a 2d plane that starts at point (0,0) for the x and y coordinates. This is usually at the top left of the canvas/screen. For this discussion, we will use an infinite canvas with the origin at (0,0). The positive x-axis moves from left to right. At the same time, the positive y-axis moves from top to bottom. This is different from normal maths where the positive y-axis is from bottom to top.

When we pan the screen, we change the position where objects are rendered, like when you scroll through social media. Therefore, if we scroll vertically moving a point from the bottom to the top of the canvas, we need to subtract from the initial y-axis position. Likewise, to move it down, we need to add to the initial y-axis position. The same thing applies if we want to move the point across the x-axis. To move a point to the right, we add to the x-axis and to move a point to the left, we subtract from the x-axis

By using this idea, we can set up a camera class that has a position (x,y) property relative to which we will render our points.

This means that if the camera is at (0,0) and a point is at (1,1). When we render the point on the screen, it will be at (1,1).

If we move the camera to (1,1), and the position of the point remains the same at (1,1), we render it at a new position on the canvas relative to the value that we have set for the camera which should be (0,0). This means that the point is now at the origin of the canvas. If we move the camera to (2,2), and the point has a fixed position at (1,1), the point will no longer be visible on the screen, because it will be at (-1,-1).

Using the above, we can deduce a formula for panning a point on the canvas

render(x,y) = point(x,y) - camera(x,y)

Therefore, by specifying the x and y positions of the camera, we can determine the coordinate on the canvas where the point would be rendered.

Zooming

In 3d when we zoom into something, we bring it closer to our eye which makes it bigger and we can see more detail. This effect can also be achieved on a 2d plane, but to develop a mathematical model of zooming, we need to go down to one dimension.

In a one-dimensional line, let's take the x-axis, the starting point is zero (0) and we can move positively to the right as well as negatively to the left. Let's move positively to a random point, say positive 50.

We want to look at the range from 40 to 60. This means that 50 is in the middle. Because (40+60)/2 = 50.

Now, let's take a look at what happens when we zoom in. If you remember what I said earlier, when we zoom in, we see more detail. Therefore we note the distance from 40 to 50 is 10. And the distance from 50 to 60 is 10. If we render this on the canvas, it would be exactly as we have described without zoom.

Zooming in is like stretching a rubber band out from both ends by the same amount.

When we zoom in, render those points on the screen, and measure their distances we will notice that they are further apart from the implied anchor which is 50. Therefore after zooming in, if we take our ruler and measure the distance from 40 to 50 and 50 to 60 on the canvas, it will be greater than 10. The amount with which it has increased is called the zoom factor.

For example, let's say, we have a zoom factor of 1.5 and render it on the screen. This means that we stretch the line out. What will be the final coordinates and distance from 50? If 40 and 60 were the initial coordinates each with a distance of 10 to the point 50.

This is really easy to calculate and requires 3 steps

Find the distance from the origin(50)
Multiply the distance to the zoom factor to get the new distance
Add the new distance to the origin(50) to get the new coordinates

Let's do this for the end at point 40 applying a zoom factor of 1.5 which is the final rendered point on the canvas.

Step 1. 40 - 50 = -10
Step 2. -10 × 1.5 = -15
Step 3. -15 + 50 = 35

For point 40, this means that the new final coordinate after zoom is applied is 35.

For point 60 also applying a zoom factor of 1.5 what is the final rendered point on the canvas?

Step 1. 60 - 50 = 10
Step 2. 10 × 1.5 = 15
Step 3. 15 + 50 = 65

For point 60, this means that the new final coordinate after zoom is applied is 65.

Therefore before Zoom, the 3 important points that we are looking at are 40,50,60. After a zoom factor of 1.5 is applied, does points are scaled to 35,50,65. This is all done in 1 dimension.

Point 50 is called the focus because as you see it remains constant as it is the point from which others are scaled.

Let's develop a formula using the above that we will use just like we did for panning.

Let's say that

xb is the coordinate of the point before Zoom,
xa is the coordinate of the point after Zoom,
xf is the focus point, and
z is the Zoom factor

Compressing the 3 steps above into one expression, we have the following

xa = (xb - xf)z + xf

Solving the above, you get

xa = zxb - zxf + xf
xa = zxb - xf(z-1)
zxb = xa + xf(z-1)
xb = (xa + xf(z-1)) / z

There you have it.
The final coordinate of a point after zoom is applied is given by the following formula

xa = zxb - xf(z-1)

Applying this to 2-dimensional coordinates requires you to perform this calculation on each coordinate for (x,y).

Applying this to the camera class, we store a zoom variable for the zoom factor and, use the x and y coordinates of the camera as the focus and apply a zoom (i.e calculate the final x and y positions) on all points relative to the focus.

render(x,y) = applyZoom(point(x,y), camera(x,y))

where applyZoom has the following function definition

applyZoom(point, focus)

Note that to pan and zoom, pass in the new pan position to the applyZoom function to get the new pan+zoom position for rendering. Pan first, zoom second.

The camera class

Here is the complete camera class in JavaScript, I have stripped it down to show only what is needed to implement panning and zooming, use it as a basis for your implementation, and don't forget to credit me for helping you out.

class GameCamera {
    constructor(gameState) {
        this.gameState = gameState;
        this.isFocusOnGameObject = false;
        this.focusedGameObject = undefined;
        this.cameraWidth = game.canvas.width;
        this.cameraHeight = game.canvas.height;
        this.position = new Vector2D(0, 0);
        this.direction = new Vector2D(0, 0);
        this.target = new Vector2D(0, 0);
        this.point = new Vector2D(0, 0);
        this.offsetX = 0;
        this.offsetY = 0;
        this.duration = 0;
        this.elapsedTime = 0;
        this.callback = undefined;
        this.minDistance = 1;
        this.zoomLevel = 1;
    }

    // Method to specify a location in world space where camera should focus
    setFocus(x, y, offsetX = 0, offsetY = 0, translateAnimationDuration = 0, animationCompleteCallback = undefined) {
        this.isFocusOnGameObject = false;
        this.point.set(x, y);
        this.target.set(this.point.x - (this.cameraWidth / 2) - offsetX, this.point.y - (this.cameraHeight / 2) - offsetY);
        this.offsetX = offsetX;
        this.offsetY = offsetY;
        this.duration = translateAnimationDuration;
        this.callback = animationCompleteCallback;
    }

    // Method to attach a target that the camera will follow
    setTarget(gameObject, offsetX = 0, offsetY = 0, translateAnimationDuration = 0, animationCompleteCallback = undefined) {
        this.isFocusOnGameObject = true;
        this.point.set(gameObject.property.position.x + (gameObject.property.size.width / 2), gameObject.property.position.y + (gameObject.property.size.height / 2));
        this.target.set(this.point.x - (this.cameraWidth / 2) - offsetX, this.point.y - (this.cameraHeight / 2) - offsetY);
        this.focusedGameObject = gameObject;
        this.offsetX = offsetX;
        this.offsetY = offsetY;
        this.duration = translateAnimationDuration;
        this.callback = animationCompleteCallback;
    }

    update(deltaLoopTimeInMilliseconds) {

        // because the object might move (change its x,y position), we update the point and target each frame
        if (this.isFocusOnGameObject) {
            // set the point to the middle of the object
            this.point.set(this.focusedGameObject.property.position.x + (this.focusedGameObject.property.size.width / 2), this.focusedGameObject.property.position.y + (this.focusedGameObject.property.size.height / 2));
            // move the camera target to position the point at the center of the camera boundary +- offset
            this.target.set(this.point.x - (this.cameraWidth / 2) - this.offsetX, this.point.y - (this.cameraHeight / 2) - this.offsetY);
        }

        // check if the camera has focused on the target else transition smoothly to the target
        if (this.position.distance(this.target) < this.minDistance) {
            this.elapsedTime = 0;
            this.setPosition(this.target.x, this.target.y);
            if (this.callback != undefined) {
                this.callback();
                this.callback = undefined;
            }
        } else {
            if (this.duration == 0) {
                this.setPosition(this.target.x, this.target.y);
            } else {
                this.elapsedTime += deltaLoopTimeInMilliseconds;
                if (this.elapsedTime > this.duration) this.elapsedTime = this.duration;
                let ratio = this.elapsedTime / this.duration;
                this.direction.set(this.target.x, this.target.y);
                this.direction.subtract(this.position);
                this.direction.multiply(ratio);
                this.position.add(this.direction);
                this.setPosition(this.position.x, this.position.y);
            }
        }
    }

    // Method to set the camera's position
    setPosition(x, y) {
        this.position.x = Math.max(0, Math.min(x, (this.gameState.worldWidth - this.cameraWidth)));
        this.position.y = Math.max(0, Math.min(y, (this.gameState.worldHeight - this.cameraHeight)));
    }

    render(textureId, x, y, width, height, sourceWidth, sourceHeight, row, frame) {

        let objectX = ((x - this.point.x) * this.zoomLevel) + this.point.x - this.position.x;
        let objectY = ((y - this.point.y) * this.zoomLevel) + this.point.y - this.position.y;

        // if object is in the frame of the camera, render it.
        if (
            objectX < this.cameraWidth &&
            objectX + width > 0 &&
            objectY < this.cameraHeight &&
            objectY + height > 0
        ) {
            textureManager.drawCameraFrame(
                textureId,
                objectX,
                objectY,
                width,
                height,
                sourceWidth,
                sourceHeight,
                row,
                frame,
                this.zoomLevel,
                game.ctx
            );
        }
    }
}

Practical example

If you have done everything as described in this article, you will have created your camera class using pure JavaScript without any frameworks. Here is an example of it in action

This gif is from ShareTXT virtual which is a free real open-world, massive online multiplayer experience, where you can interact with other players and engage in fun activities together.

The graphic assets for the illustrations used in this article are listed below

I hope this article has helped you to understand the theory of how panning and zooming work in a 2d plane or game. I have given you the formulas, as well as how to represent them in code for execution in your programming language of choice. I have also shown you the practical result that should be expected if you do everything as expected. Have a nice day.

Meet the Future of Virtual Worlds: An Inside Look at ShareTXT Virtual

Rex Anthony — Wed, 07 Aug 2024 06:21:00 +0000

Imagine a world where every player you meet is a real person, and every conversation could lead to a new friendship, alliance, or rivalry. Whether you're a social butterfly or a lone explorer, ShareTXT Virtual offers something for everyone.

ShareTXT virtual is a free-to-play online massive multiplayer open-world experience where you can interact with other players and engage in fun activities together.

It was developed with a mobile-first approach, which is an innovative design methodology. This means that it works well on both mobile and desktop with a high-performance metric.

*This is more than just a game; it's a dynamic, interactive universe where players from all corners of the globe come together to explore, compete, and connect. *
This is a dream come true and is shaping up to become a massive shake-up in online interaction. This is great news for players.

The reviews have been glowing, with users hailing everything from the interaction to the gameplay experience.

Connect and Communicate

Communication is at the heart of ShareTXT Virtual. The platform offers a range of ways to interact with fellow players:

Public Shout Outs: Got something to say to the entire world? Public shoutouts let you broadcast your messages for everyone to see. Whether you’re organizing an event, issuing a challenge, or just want to make your presence known, public shoutouts are your megaphone.
Private Messages: Prefer a more personal touch? Private messaging allows you to have one-on-one conversations, perfect for strategizing with friends or getting to know new acquaintances.

Customize Your Experience

Express yourself in ShareTXT Virtual with multiple skins that allow you to personalize your character. Whether you want to stand out in a crowd or blend in with your surroundings, these customization options ensure your avatar truly represents you.

There are multiple skins that players can choose from to customize their playing experience.

Explore Vast Terrains

The world of ShareTXT Virtual is vast and varied, featuring over 30 unique terrains waiting to be explored. Each terrain is filled with hidden secrets and stories just waiting to be discovered. From lush forests to eerie wastelands, every corner of this world holds new adventures and challenges.

Engage in Exciting Activities

The fun never stops in ShareTXT Virtual. The city is a bustling hub of activities where there's always something happening:

Mini Games: Compete against other players in a variety of mini-games. Whether you’re battling for the top spot on the leaderboard or just looking to have some fun, these games offer endless entertainment and rewards.
Night Clubs: Ready to show off your dance moves? Head to the nightclub where you can dance the night away and maybe even win a dance-off.
Theaters and Poetry: For those who appreciate the arts, our theaters host songs, poetry readings and storytelling sessions. Sit back, relax, and enjoy the creative expressions of your fellow players.
Comedy Clubs: Laughter is the best medicine, and our comedy clubs are just the place to get your dose. Enjoy performances by some of the funniest players in the game, or step on stage yourself and try out your best jokes.
Book clubs: If you love diving into a good book and discussing it with others, our book clubs are perfect for you. Share your favorite reads, explore new genres, and connect with other book enthusiasts. The world is at your feet. You can create your group and interactions in the virtual world.

Uncover Hidden Secrets

Eventually, the inhabitants of ShareTXT Virtual wouldn't just be other players. In time, our world will be populated by AI-driven NPCs (Non-Player Characters) who have their own stories, secrets, and personalities. Interact with them just like you would with any other player. Earn their trust, and they might just reveal secrets that can lead to new quests and hidden treasures.

Join the Adventure

So what are you waiting for? ShareTXT Virtual is more than a game; it's a living, breathing world full of possibilities. Whether you're looking to make new friends, challenge your skills, or just explore a rich and immersive environment, ShareTXT Virtual has something for you.

Ready to dive in? Visit ShareTXT Virtual now and start your adventure today. The world is waiting for you.

File Sharing Alternatives to Email Attachments

Rex Anthony — Mon, 17 Apr 2023 19:36:23 +0000

Email attachments can be a convenient way to share files with others, but they also come with a number of challenges. Unlike sharing passwords and text, attachments can be large, which can make emails slow to send and receive, and they can also be prone to errors and security risks. If you're looking for a better way to share files, there are a number of alternatives to email attachments that can provide more reliable and secure file-sharing options.

In this post, I'll explore some of the best file-sharing alternatives to email attachments. From cloud storage and file-sharing services to file transfer protocols and collaboration and project management tools, we'll cover a range of options that can help you share files more efficiently and securely. Whether you're working on a team, sharing files with clients, or simply looking for a better way to share files with friends and family, these file-sharing alternatives can provide the solutions you need.

Cloud storage and file-sharing services

Cloud storage and file-sharing services provide a convenient and easy way to store and share files online. These services offer users a certain amount of storage space on remote servers, which can be accessed from any device with an internet connection. Some popular examples of cloud storage and file-sharing services include Google Drive, Dropbox, and OneDrive.

One of the main benefits of using cloud storage and file-sharing services is that they allow users to access their files from any device, at any time. This can be especially useful for teams or individuals who need to share files and collaborate on projects remotely. Cloud storage and file-sharing services also make it easy to share files with others, as users can simply send a link to the file or folder, rather than sending large attachments via email.

However, there are also some potential drawbacks to using cloud storage and file-sharing services. One concern is security, as users are trusting their files to a third-party service. It's important to choose a reputable and secure cloud storage and file-sharing service and to take additional steps to protect your files (e.g. using strong passwords, and enabling two-factor authentication). Additionally, users may need to pay for additional storage space if they exceed the allotted amount, and some services may have limitations on file types or sizes that can be uploaded.

File Sharing services such as ShareTXT eliminate the worry about security by offering an innovative solution that enables you to share your files end-to-end between your devices without them being stored somewhere in some database. This eliminates the need for an intermediate server.

Overall, cloud storage and file-sharing services can be a convenient and efficient way to store and share files, but it's important to carefully consider the security and cost implications before committing to a particular service.

File transfer protocols

:{"type":"videoEmbed", "src":"https://www.youtube.com/embed/tOj8MSEIbfA", "title":"Youtube Video Player", "allow":"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share"}
...

File transfer protocols (FTPs) are a set of rules that govern the exchange of files over a network. There are a variety of file transfer protocols, including FTP (File Transfer Protocol), SFTP (SSH File Transfer Protocol), and HTTPS (Hypertext Transfer Protocol Secure). These protocols enable users to transfer files between computers, servers, and other devices over the internet.

One of the main benefits of using file transfer protocols is that they allow you to transfer large files quickly and efficiently. FTPs also offer a high level of security, as they use encryption to protect the data being transferred. This can be especially important for transferring sensitive or confidential files.

However, there are also some potential drawbacks to using file transfer protocols. One concern is that they can be complex to set up and configure, especially for users who are not familiar with network protocols. FTPs also require specialized software or tools to use, such as FileZilla or WinSCP. Additionally, FTPs may not be as user-friendly as other file-sharing methods, such as cloud storage and file-sharing services.

Overall, file transfer protocols can be a powerful and secure way to transfer large files, but they may not be the best option for all users. It's important to consider your specific needs and the resources available to you when deciding whether to use a file transfer protocol for file sharing.

Collaboration and project management tools

Collaboration and project management tools are software applications that enable teams to work together and manage projects more efficiently. These tools can include features such as task assignments, file sharing, real-time communication, and progress tracking. Some popular examples of collaboration and project management tools include Trello, Asana, and Slack.

One of the main benefits of using collaboration and project management tools is that they provide a central platform for teams to organize and manage their work. These tools can help teams stay organized and on track, and can also improve communication and collaboration between team members.

However, there are also some potential drawbacks to using collaboration and project management tools. One concern is that these tools may require a learning curve for users who are not familiar with them, and may require some time and effort to set up and configure. Additionally, some collaboration and project management tools may require a subscription or payment to use, and may not be suitable for all types of projects or teams.

Overall, collaboration and project management tools can be a powerful way to improve productivity and efficiency for teams, but it's important to carefully consider the costs and resources required before committing to a particular tool.

Conclusion

In conclusion, email attachments are not the only option for file sharing. There are a number of alternatives that can provide more efficient, secure, and reliable file sharing solutions. From cloud storage and file sharing services to file transfer protocols and collaboration and project management tools, there are a range of options available to suit different needs and budgets.

When choosing a file sharing method, it's important to consider the specific needs of your team or project, as well as the resources and budget available. Consider factors such as the size and type of files being shared, the level of security and reliability required, and the level of collaboration and organization needed. By carefully evaluating the options and choosing the best solution for your needs, you can improve the efficiency and effectiveness of your file sharing process.

How to Properly Attribute and Cite Sources when Sharing Content Online

Rex Anthony — Wed, 04 Jan 2023 13:27:05 +0000

When sharing content online whether as a post, social media, or any other media, it's important to properly attribute and cite sources to give credit to the creators and owners of the content, and to ensure that your own work is accurate and credible. Proper attribution and citation not only helps to respect the rights of others, but it can also enhance the quality and value of your own content.

But what exactly is attribution and citation, and why are they important? Attribution is the act of giving credit to the source of the content you are using or sharing. This can be as simple as including the name of the creator or owner and a link to the original source. Attribution helps to acknowledge the contributions of others and can help to establish your own credibility as a content creator.

Citation, on the other hand, involves providing more detailed information about the source of the content, including the title, author, publication date, and other relevant details. Citations are typically used in academic or professional contexts to provide a clear and detailed record of the sources used in a piece of work. In addition to helping to establish credibility and prevent plagiarism, citations also provide readers with the information they need to locate and verify the sources themselves.

In this blog post, I'll delve into the basics of attribution and citation, and discuss best practices for properly attributing and citing sources when sharing content online. I'll also explore some tools and resources that can help you to attribute and cite sources effectively. Whether you're a content creator, business owner, or social media enthusiast, these tips will help you to properly attribute and cite sources and share content with confidence.

What is attribution?

Attribution is the act of giving credit to the source of the content you are using or sharing. It involves acknowledging the contributions of others and helping to establish your own credibility as a content creator. Proper attribution is important for a number of reasons:

It helps to respect the rights of others: By properly attributing content, you are acknowledging the contributions of the creators and owners of the content and showing respect for their rights. This is especially important when you are using content that is protected by copyright or other intellectual property laws.
It can enhance the quality and value of your own content: By including attribution, you can provide your readers with additional context and resources, and enhance the overall value of your content. Attribution can also help to establish your own credibility as a content creator by demonstrating that you are using reliable and reputable sources.
It can help to prevent plagiarism: Proper attribution can help to prevent plagiarism, which is the act of using someone else's work without giving credit. Plagiarism is not only unethical, but it can also have serious legal and professional consequences. By properly attributing content, you can avoid potential issues related to plagiarism and demonstrate that you are using content responsibly.

To properly attribute content, you should include the name of the creator or owner and a link to the original source. This can be as simple as adding a byline or including a link to the source in the text of your post. Depending on the context and the type of content you are sharing, you may also need to include additional information, such as the title of the work or the date it was published.

What is citation?

Citation is the act of providing detailed information about the source of the content you are using or sharing. It involves including specific details about the work, such as the title, author, publication date, and other relevant information. Citations are typically used in academic or professional contexts to provide a clear and detailed record of the sources used in a piece of work.

There are many different citation styles, each with its own set of rules and guidelines for formatting citations and referencing sources. Some common citation styles include APA, MLA, and Chicago. Depending on the context and the type of content you are sharing, you may need to use a specific citation style or follow the guidelines of a particular organization or publication.

Citation is important for a number of reasons:

It helps to establish credibility and prevent plagiarism: By properly citing sources, you can demonstrate that you are using reliable and reputable sources and help to prevent plagiarism. Plagiarism is the act of using someone else's work without giving credit, and it can have serious legal and professional consequences.
It provides readers with the information they need to locate and verify sources: By including detailed citation information, you can provide your readers with the information they need to locate and verify the sources you are using. This can be especially important in academic or professional contexts, where readers may need to consult the original sources for further information or analysis.
It helps to support your own work: By citing sources, you can provide evidence and support for your own work and help to establish your own credibility as a content creator. Citations can also help to demonstrate that you are familiar with the research and literature in your field and that you are using this knowledge to inform your own work.

To properly cite sources, you should include all of the relevant information about the source, including the title, author, publication date, and other details as required by the citation style you are using. You should also include a reference list or bibliography at the end of your work, which lists all of the sources you have cited in alphabetical order.

Tools and resources for attribution and citation

There are many different tools and resources available to help you attribute and cite sources effectively when sharing content online. Here are some options to consider:

Manual methods: If you prefer to attribute and cite sources manually, there are a number of resources that can help you to do so. These include style guides, such as the APA, MLA, or Chicago style guides, which provide detailed instructions on how to format citations and references. You can also find online resources, such as citation generators, which can help you to create citations in a specific style.
Online tools and resources: There are a number of online tools and resources that can help you to attribute and cite sources more efficiently. Some options include:
- Citation management tools: These tools allow you to create, organize, and manage your citations and references in one place. Popular citation management tools include EndNote, Zotero, and Mendeley.
- Plugins and extensions: There are a number of browser extensions and plugins that can help you to attribute and cite sources more easily. For example, the EasyBib extension for Google Chrome allows you to generate citations for websites and articles with a single click.
- Online citation generators: There are many online citation generators available that can help you to create citations in a variety of styles. Some popular options include Citation Machine and BibMe.

By using these tools and resources, you can streamline the process of attributing and citing sources and ensure that you are using content responsibly and legally.

Conclusion

In conclusion, proper attribution and citation are crucial for anyone who creates or shares content online. By giving credit to the sources of the content you use and sharing detailed information about those sources, you can respect the rights of others, establish your own credibility, and prevent plagiarism. Whether you're a content creator, business owner, or social media enthusiast, understanding and properly using attribution and citation can help you to share content responsibly and legally.

There are many different tools and resources available to help you attribute and cite sources effectively. From manual methods like style guides to online tools like citation management software and online citation generators, there are a variety of options to suit your needs and preferences. By using these resources, you can streamline the process of attributing and citing sources and ensure that you are using content responsibly and legally.

I hope this blog post has helped you to better understand the importance of attribution and citation when sharing content online and has given you some ideas for how to attribute and cite sources effectively. Let me know if you have any other questions or need further assistance.

Time Management Techniques for Remote Workers

Rex Anthony — Sun, 01 Jan 2023 12:17:56 +0000

Remote work has become increasingly popular in recent years, and with the COVID-19 pandemic, more people than ever are working from home. While remote work has many benefits, such as flexibility and the ability to work from anywhere, it can also be challenging to stay focused and manage your time effectively.

Effective time management is especially important for remote workers, as it can help you stay on track and achieve your goals more efficiently. By using time management techniques, you can avoid distractions, stay organized, and get more done.

In this blog post, we'll discuss time management techniques specifically for remote workers. From setting clear goals and priorities to taking breaks and exercising regularly, these techniques can help you stay focused and avoid burnout. By following these tips, you can improve your productivity and achieve your goals more efficiently as a remote worker.

Identify your goals and priorities

Identifying your goals and priorities is an important step in managing your time effectively when you work remotely. By setting clear goals and priorities, you can stay focused on what is most important and avoid wasting time on tasks that are less important or not aligned with your goals.

To identify your goals and priorities, start by considering what you want to achieve in your work. This could be career advancement, increasing your income, or making a positive impact in your field. Once you have identified your overall goals, break them down into smaller, more specific objectives. For example, if your goal is to advance your career, you might set specific objectives such as completing a certification program or networking with industry leaders.

Once you have identified your goals and objectives, prioritize them based on their importance and deadline. This can help you stay focused on what is most important and avoid getting overwhelmed by too many tasks. You can also use tools like the Eisenhower Matrix to help you prioritize tasks based on their urgency and importance.

By setting clear goals and priorities, you can stay focused on what is most important and avoid wasting time on tasks that are less important or not aligned with your goals. This can help you manage your time effectively and achieve your goals more efficiently as a remote worker.

Create a task list

Creating a task list is an important part of managing your time effectively when you work remotely. By organizing your tasks in a list, you can stay on top of what needs to be done, avoid wasting time, and stay focused on your goals.

To create a task list, start by writing down all the tasks you need to complete. This can include everything from small tasks like answering emails to larger projects like creating a presentation. Once you have a comprehensive list of tasks, prioritize them based on their importance and deadline. This can help you focus on the most important tasks first and avoid getting overwhelmed.

There are several tools and techniques you can use to create and manage your task list. One simple method is to use a paper or digital to-do list, where you can write down each task and check it off as you complete it. You can also use task management software or apps, such as Todoist, Trello or Asana, which allow you to create and organize tasks, set deadlines, and track progress.

By creating a task list and using tools to manage your tasks, you can stay on top of what needs to be done and avoid wasting time. This can help you manage your time effectively and achieve your goals more efficiently as a remote worker.

Use the Pomodoro Technique

The Pomodoro Technique is a time management technique that can help you stay focused and avoid burnout when you work remotely. The technique involves working in focused bursts, followed by short breaks to recharge. This can help you stay focused and avoid distractions, and it can also help you avoid burnout by taking breaks to rest and recharge.

To use the Pomodoro Technique, set a timer for 25 minutes and work on a task without any distractions. When the timer goes off, take a 5 minute break to rest and recharge. After four Pomodoros (or 100 minutes of work), take a longer break of 15-30 minutes. You can use a timer or a Pomodoro app to help you stay on track.

The Pomodoro Technique can be customized to fit your needs and schedule. You can adjust the length of the work and break intervals to find what works best for you. Some people find that shorter work intervals help them stay focused, while others prefer longer intervals. Experiment with different intervals to find what works best for you.

By using the Pomodoro Technique, you can stay focused and avoid burnout when you work remotely. This can help you manage your time effectively and achieve your goals more efficiently.

Manage your environment

Managing your environment is an important part of managing your time effectively when you work remotely. By creating a dedicated workspace and minimizing distractions, you can stay focused and avoid wasting time.

To manage your environment when you work remotely, create a dedicated workspace. This could be a separate room or a designated area in your home where you can work without distractions. Make sure your workspace is comfortable and well-lit, and set it up with everything you need to get your work done, such as a computer, printer, and office supplies.

Another way to manage your environment when you work remotely is to minimize distractions. This could include turning off notifications on your phone or computer, closing unnecessary tabs or programs, and letting people know you are working and not to be disturbed. You may also want to consider using noise-cancelling headphones or a white noise machine to block out external distractions.

By managing your environment and minimizing distractions, you can stay focused and avoid wasting time when you work remotely. This can help you manage your time effectively and achieve your goals more efficiently.

Schedule breaks and time for physical activity

Scheduling breaks and time for physical activity is an important part of managing your time effectively when you work remotely. By taking breaks and exercising regularly, you can recharge and avoid burnout, which can help you stay focused and productive.

To schedule breaks and time for physical activity when you work remotely, consider setting aside specific times for these activities in your schedule. For example, you might take a short break every hour to stretch or go for a walk, and schedule a longer break or exercise session in the middle of the day.

There are many different types of breaks and physical activities you can choose from. Some options might include going for a walk or run, practicing yoga or meditation, or participating in a group fitness class. Whatever you choose, make sure to choose activities that you enjoy and that help you relax and recharge.

By scheduling breaks and time for physical activity, you can recharge and avoid burnout when you work remotely. This can help you manage your time effectively and achieve your goals more efficiently.

Communicate effectively with team members and stakeholders

Effective communication is an important part of managing your time effectively when you work remotely. By communicating clearly and efficiently with your team members and stakeholders, you can avoid misunderstandings and stay on track with your work.

To communicate effectively with your team members and stakeholders when you work remotely, there are several things you can do:

Use clear, concise language: Avoid jargon or technical terms that might not be familiar to everyone, and use simple, straightforward language to convey your message.
Use visual aids: When possible, use visual aids like charts, graphs, or slides to help illustrate your points and make your message more clear.
Use video conferencing: Use video conferencing tools like Zoom or Skype to communicate face-to-face with your team members and stakeholders. This can help you build rapport and improve communication.
Follow up: After a meeting or conversation, follow up with a summary or next steps to make sure everyone is on the same page.

By communicating effectively with your team members and stakeholders, you can avoid misunderstandings and stay on track with your work. This can help you manage your time effectively and achieve your goals more efficiently as a remote worker.

Stay organized

Staying organized is an important part of managing your time effectively when you work remotely. By keeping your work and personal life organized, you can avoid wasting time looking for things and stay focused on your tasks.

To stay organized when you work remotely, there are several things you can do:

Create a system for organizing your files: Use a file naming convention and create folders to keep your files organized and easy to find. You can also use cloud storage tools like Google Drive or Dropbox to store and access your files from anywhere.
Use task management software or apps: Use tools like Trello{:target="_blank"} or Asana to create and manage your tasks, set deadlines, and track progress.
Declutter your workspace: Keep your workspace clean and organized by decluttering regularly. This can help you stay focused and avoid distractions.
Set boundaries: Set clear boundaries between your work and personal life by creating designated work and non-work areas in your home.

By staying organized, you can avoid wasting time looking for things and stay focused on your tasks. This can help you manage your time effectively and achieve your goals more efficiently as a remote worker.

Conclusion

Working remotely can present challenges when it comes to managing your time effectively. By using time management techniques like setting clear goals and priorities, creating a task list, using the Pomodoro Technique, managing your environment, scheduling breaks and physical activity, communicating effectively with team members and stakeholders, and staying organized, you can stay focused and avoid burnout.

Remember that every person is different, so it's important to find time management techniques that work for you. Experiment with different techniques and tools to find what works best for you, and don't be afraid to adjust your approach as needed.

By managing your time effectively, you can achieve your goals more efficiently and enjoy the benefits of working remotely.

DEV Community: Rex Anthony

Apple’s On-Device AI: The Quiet Revolution for Edge Computing and Local-First Apps

Why the On-Device Push?

Privacy as a product

Performance is about latency

Persistence means it works everywhere

Technical Realities

Memory

Power and Thermals

Compute Orchestration

Apple's solution to these issues

Aggressive model compression, built into the toolchain

Distilled foundation models with hybrid routing

Unified memory and the Neural Engine

The Elephant in the Room

The Local-First Revolution

1. Hyper-personalized, context-aware assistants

2. Real-time media creation without uploads

3. Offline productivity that actually works

4. Proactive intelligence across apps

Disrupting the Cloud-Centric AI Model

Benefits for developers

Benefits for users

The Reality Check

Is Cloud AI's Dominance Overstated?

Broader implications

What Developers should expect

What changes for you

What to demand next

Beyond Cupertino: Implications for the Wider Edge AI Ecosystem

Will Apple's Walled Garden Accelerate or Hinder True Edge AI Innovation?

The bull case

The bear case

Conclusion

Hyperscalers Now Say 'More Jobs' as Firms Burn Millions on Tokens

Everything I Learned about Patience from 8-Shaft Weaving

I Ran a Local AI on a 16-Year-Old Windows 7 PC with Only 4GB RAM — And It Actually Works! 🚀

What I learned from going 'cloud-optional' with my side projects.

What I gained:

How to Publish Your Content on ShareTXT

Introduction

Table of Contents

Getting Started – The Publisher Dashboard

Quick Tips Before You Start:

1. Navigation Bar

1. Demo

2. New Draft

3. Focus Mode

4. Light / Dark Mode

5. Publish

2. Metadata + Editor Section

Classification

Content Information

SEO Meta Data

Social Meta Data

Routing

Attribution

Assets

Publisher Links

Note that each link here must match exactly the full link in your article that you want to assign dofollow

Markdown Content

3. Previews Section

1. Webpage Preview

2. SERP Preview

3. Social Preview

Publishing Your Article

Final Steps to Publish:

What Happens Next?

Post-Publication Checklist

Additional Tips & Best Practices

Content Quality & Optimization

Images & Visuals

Links & SEO

Common Mistakes to Avoid

Workflow Tips

After Publishing

Frequently Asked Questions

1. How much does it cost to publish on ShareTXT?

2. Can I edit my article after publishing?

3. What image hosting services can I use?