Hey guys! A month ago I've committed to build an AI product The idea is simple: use AI to build personal brands for a business owners on LinkedIn and X (so they get more clients and all other benefits of having a personal brand) It originally happened as a client's project: a $100M founder from the US came to me and ordered this AI system I've built it for him, and then I realized that it can be scaled 🚀 Now every day I'm working on this product, improving it and trying to get more views for myself and my clients I'm dropping a vlog showing how I'm building this startup LIVE soon You can also follow along the journey on my LinkedIn and Twitter: - https://www.linkedin.com/in/melnikoff-oleg/ (2.6k subs) - https://x.com/olegrows (5 subs - would HIGHLY appreciate your support here 😁)

Google just dropped (quietly 👀) its biggest technological leap to date: 𝗚𝗲𝗺𝗶𝗻𝗶 𝟯.𝟬, the most powerful evolution in its entire multimodal model lineup. And yes — it’s packed with game-changing upgrades. 🔥 W𝗵𝗮𝘁 𝗺𝗮𝗸𝗲𝘀 𝗚𝗲𝗺𝗶𝗻𝗶 𝟯.𝟬 𝘀𝗽𝗲𝗰𝗶𝗮𝗹? Gemini 3.0 isn’t an update. It’s a full rebuild of Google’s multimodal capabilities — now capable of understanding and reasoning across: ✔️ Text ✔️ Images ✔️ Code ✔️ Audio & video ✔️ Complex combinations of all of them But this time with next-gen reasoning. It can solve long, multi-step or complex tasks in a single pass — what Google calls one-shot performance. ⚙️🚀 𝗦𝗲𝗿𝗶𝗼𝘂𝘀 𝗽𝗼𝘄𝗲𝗿 𝘂𝗻𝗱𝗲𝗿 𝘁𝗵𝗲 𝗵𝗼𝗼𝗱 The Pro version of Gemini 3.0 may exceed one trillion parameters, resulting in: ⚡ Deeper contextual understanding ⚡ Much faster inference ⚡ Ultra-low latency for real-time apps ⚡ Higher accuracy on multimodal tasks 𝗢𝗻 𝘁𝗼𝗽 𝗼𝗳 𝘁𝗵𝗮𝘁, 𝗶𝘁𝘀 𝗻𝗲𝘄 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 𝘀𝗵𝗶𝗻𝗲𝘀 𝗶𝗻 𝗰𝗿𝗲𝗮𝘁𝗶𝘃𝗲 𝘄𝗼𝗿𝗸, 𝗻𝗼𝘄 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗻𝗴: 🎨 Complex HTML/CSS 🎬 Embedded animations 🧩 Clean, structured code 📷 Coherent visual edits from text All thanks to improved tokenization and attention mechanisms. 🕵️‍♂️💡 𝗦𝗵𝗮𝗱𝗼𝘄 𝗹𝗮𝘂𝗻𝗰𝗵 (𝘆𝗲𝘀 — 𝗚𝗼𝗼𝗴𝗹𝗲 𝗱𝗶𝗱 𝗶𝘁 𝗮𝗴𝗮𝗶𝗻) Google didn’t announce it publicly… They hid it inside the mobile model selector. If you choose 👉 “Gemini 2.5 Pro” in Canvas mode, you’re actually triggering Gemini 3.0 Pro. Developers already confirmed it by inspecting backend calls and comparing outputs. Classic Google shadow launch. 😅 📢 W𝗵𝗮𝘁’𝘀 𝗰𝗼𝗺𝗶𝗻𝗴 𝗻𝗲𝘅𝘁? A formal announcement is expected any moment. ✨ Full API access ✨ Deep Vertex AI integration ✨ New tools for video + autonomous agents ✨ Better production-grade performance Gemini 3.0 isn’t coming — 👉 it’s already moving. 🧩 𝗥𝗲𝗮𝗹-𝘄𝗼𝗿𝗹𝗱 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲𝘀 𝘆𝗼𝘂 𝗰𝗮𝗻 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗶𝗺𝗮𝗴𝗶𝗻𝗲 𝘄𝗶𝘁𝗵 𝗚𝗲𝗺𝗶𝗻𝗶 𝟯.𝟬 🎥 𝟭. 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗲𝗱 𝘃𝗶𝗱𝗲𝗼 𝗲𝗱𝗶𝘁𝗶𝗻𝗴 Describe the edit → Gemini analyzes the footage → trims, sequences scenes, adds transitions. Perfect for creators who want to publish 10x faster. 💻 𝟮. 𝗙𝘂𝗹𝗹𝘆 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱 𝗹𝗮𝗻𝗱𝗶𝗻𝗴 𝗽𝗮𝗴𝗲𝘀 Prompt + visual reference → Gemini → outputs HTML + CSS + animations ready to ship.

As my self-hosting, both Locally & Remotely gets more complex I am constantly looking for suitable tools to spin up VMs for n8n DEV & Testing environments. A good friend of mine suggested ProxMox which seems capable but it is squarely aimed at the Enterprise IT folk. However Google News just served me an article from the XDA Forums on another open source tool made for self-hosted virtualizations, called Incus 👏🏻 Top 5 Benefits of Incus: 1). Superior Security - Runs all containers as "unprivileged" by default, meaning they have no root power on the host system. - Features deep kernel integration with AppArmor, Seccomp, and SELinux for a fortified, secure foundation. 2). Efficient Image-Based Workflow - Launch instances from pre-built images for fast, consistent, and reproducible environments. - Supports a massive library of Linux distributions, simplifying setup and boosting productivity. 3). Unified Containers & VMs - Manage both system containers and full virtual machines with the same simple set of commands. - Choose the right isolation for the job—lightweight containers for density or VMs for kernel-level security. 4). Simple Scaling & Clustering - Use an intuitive command-line interface that's easy to learn but powerful enough for production. - Scale from a single node to a full cluster with live-migration, all managed as a single entity. 5). Massive Scale & Organization - Engineered for high performance with minimal overhead, capable of running thousands of instances. - Organize everything using "projects" to cleanly separate teams, clients, or environments within one cluster. So it's possible and I will have to try it out that I could run 10 x lightweight VMs on each of my 8 Nodes from my cluster. If so, the next test would be n8n on each of those Incus-based VMs. Interesting 🤔

Hi all: Just a quick update about my cluster, it is up and running, but I am still configuring filesystem monitoring, tracking and performance metrics plus a visualization dashboard. Rough specifications: - 8 x HP EliteDesk G3 Nodes - Ubuntu Server 24.04 LTS - Docker Engine Latest - n8n Production release v1.121.2 - Fail2Ban/Tripwire - UFW - Prometheus & VictoriaMetrics - Extensive system tools Note: I do not need to make it as `bulletproof` as usual as it will be running 100% within my home network, behing the router FW, and am considering limiting it's ability to do anything else but source `APT` updates from the relevant Repository Pool & Cache. All interactions will be via an n8n workflow located on my hosted VPS. The plan is to convert a Python-based Business Intelligence tool into a Docker Image, then connect it and my Google GCP Workspace using some n8n modules then Do Cool Stuff™️ Am also considering testing an n8n Benchmarking & Security analysis tool using the cluster 🤔 But early days yet. For the time being, here is minimal proof ... It's alive 👾🌩️👨🏻‍🔬 Cost so far? About $2250 😱 ... getting a custom case built, plus a custom transformer so 🤦🏻‍♂️

For those managing multiple environments (Dev/Staging/Prod), how are you handling version control and deployment of your workflows? Are you using the CLI, a custom sync script, or a different method? I'm keen to learn about systems that allow for safe iteration and rollback without manual export/import.