Ranga Sampath | Network Expert | AI Things Builder

I went dark for three years. The podcast downloads didn't stop. Post January 2022, I stopped the You+AI podcast — it had reached Top 8 globally in its category. Full guest pipeline. Sent them all declines. Later, took an enterprise engineering role and went heads-down. Didn't delete anything. 💡 Eight months after stopping: still getting monthly downloads. Twelve months. Twenty months. Requests for new episodes and guest recommendations arriving in 2023, 2024, even into 2026. In 2020 I had written about four things I believed made innovation work: *Be Curious. Embrace Change. Ride the Wave. Partner.* 🔍 What I found when I came back in 2026 — to build open-source agentic AI tools for cloud network troubleshooting — was that all four had kept running without me. My website archive did the work I was not doing. The timestamped work log outlasted any portfolio I could have built. The name You+AI turned out to describe a thesis, not a topic: human expertise and AI capability working together, whoever the human is, whatever the domain. The tools I am building now are the most honest expression of that thesis yet. Dormant is not dead. The foundation is real, or it isn't. Stopping is not the same as failing. Original 2020 recipe, built to last, read the full recipe @ youplusai.com/dormant-is-not…↗ #PersonalBrand #Innovation

@saybwala Seriously valuable cheat sheet

A friend who is relocating to India after an adult life spent in the US, today casually asked which is a good 2 wheeler to buy. 🙄 He ended up getting a lecture from me and a survival cheat sheet for when he wakes up in the morning.

@captsanthoshkc How may we source this?

My mother, aged 70, meticulously tends to our farm, cultivating these exquisite Miyazaki mangoes. Her dedication to nature is evident in her sustained efforts. 🥭👩‍🌾🌿

@ParveenKaswan So beautiful!

#Diversity is the insurance for the future. Seeds of different species often fall beneath trees where hornbills nest. This is how these magnificent birds help regenerate forests; and why they are rightly called the “farmers of the forest.” Today is #BiodiversityDay. But how do they do it? India is home to 9 species of hornbills. During the nesting season, the female seals herself inside the nest cavity to incubate and raise the chick. For nearly 3–4 months, the male takes complete responsibility for feeding the family. During this period, the male continuously collects fruits and figs from across the forest and brings them to the nest. Hornbills swallow fruits whole, with the seeds remaining intact. The transfer, however, is not always smooth. Many seeds fall during flight, feeding, or transportation. As a result, seeds get dispersed far and wide across the forest, helping new plants and trees grow in different areas. Protecting biodiversity therefore also means protecting these incredible species that silently work every day to keep forests alive.

@supriyasahuias @MegamalaiS @tnforestdept @UNBiodiversity @UNEP @WWFINDIA @WWF @IUCN @dhanu_paran @shanmughanandam @pnsenthilkumar @sriram_p_murali To state "once existed" breaks my heart 💔 We must do better.

Show this to your children. They deserve to know that there once existed magical forests lit by fireflies, purple frogs that emerged with the rains, slender lorises watching silently from the night canopy, vultures circling wild skies, dugongs grazing seagrass meadows, striped hyenas walking forgotten scrublands, turtles blessing our shores, Raptors soaring high, Amur Falcons crossing oceans without rest for days, and Nilgiri Tahrs ruling mountain escarpments like a kingdom above the clouds. Perhaps we are the last generation to witness many of these wonders in the wild. The Earth is losing biodiversity at a pace never seen before and with every disappearance, something ancient, irreplaceable and deeply alive fades away forever. Biodiversity is Earth’s heartbeat. The moment it begins to fade, the planet will slowly forget how to breathe. If we want a future, we must help the Earth breathe again. We already know what to do. Let’s do it. Protect. Conserve. Cherish. Happy #BiodiversityDay @UNBiodiversity #LocalAction #KMGBF #IDB2026 #ForNature

I set out to build a Detective. Built a Bureaucrat instead. I wanted to build an AI that could diagnose cloud network faults the way a senior engineer would. Cloud networks are opaque. Faults invisible until something breaks in production. I built an agent. Gave it specific commands for specific scenarios. --- What I had built, without realising it, was a bureaucrat. It followed the procedure. It didn't reason. The moment a new scenario arrived — similar, but not identical — it failed. 💡 The fix: encode intent, not procedure. Not "if you see X, check Y and then run this tool on these targets." BUT the reasoning behind the step — why you'd run it in the first place. When a new scenario arrived, the principle held. --- That got me closer to what I set out to build: a detective that cuts through cloud opacity and finds what's actually broken. Closer. Not there yet. Read up on what closed the remaining gap: youplusai.com/detective-vs-b… #AIBuilders #SystemPrompts

🟢NSG says allowed. Azure portal says healthy. 🔴iperf times out. The block lives in the OS firewall — a layer the Azure API can't query. That's the class of problem the expanded Ghost Agent toolkit is built for. The cloud control plane and the OS-level firewall are independent systems. Each can report clean while the other drops traffic. Compound faults — two independent failures at different layers — stay hidden until you look at all layers in a single session. 🔧 Three new tools, three blind spots closed: **Netfilter Inspector** — retrieves iptables/nftables rulesets from Azure VMs via `az vm run-command` or SSH. Snapshots baselines before change windows. Diffs post-change state and sends the diff to an LLM for traffic-impact analysis via `--explain-diff`. 249 tests across iptables, nftables, and IPv6 paths. **Effective Network Inspector** — snapshots the computed effective route table and combined NSG evaluation at the NIC — not the configured resources. BGP route withdrawal leaves zero ARM records. Azure Change Analysis has nothing to log. The diff does. SHA-256 verified artifacts. **Agentic Pipe Meter** — qperf (TCP latency) + iperf2 (8-stream throughput), P90 over configurable iterations, NSG port pre-flight check, baseline comparison, structured JSON artifact to Blob Storage. 194 tests across all pipeline stages. ⚠️ One tested scenario that shows why this matters: Change window closes. Lead engineer emails: "All changes applied. Rollback complete." Port 8080 unreachable 90 minutes later. Effective Network Inspector ➡️ compares the current NIC state against a pre-window snapshot. SHA-256 verification confirms the baseline is authentic. TCP 8080 DENY at priority 150 is in the diff. The rollback was incomplete — the diff proves it. Netfilter Inspector ➡️finds an iptables DROP for TCP 5001 on the same VM. No ARM record. Not in Azure Change Analysis. A field engineer added it during the window and forgot. Two faults. Two layers. Both on the same VM. Two more compound fault scenarios in the full post — youplusai.com/diagnose-compo… ↗including one where a Network Watcher capture on the destination VM is the evidence that motivates checking iptables. Ghost Agent, Netfilter Inspector, Effective Network Inspector, Pipe Meter — all open source at github.com/ranga-sampath/…

@NewsTamilTV24x7 Our namaskaram to such great people who work for the societal good 🙏

பாபநாசம்: தாமிரபரணி ஆற்றில் பக்தர்கள் போட்டுச் செல்லும் துணி கழிவுகளை கடந்த 12 ஆண்டுகளாக அகற்றி வரும் ஓய்வு பெற்ற அரசு அதிகாரி.. #ThamirabaraniRiver #Papanasam #TextileWaste #WasteManagement #EnvironmentalAwareness #CleanRiver #Newstamil24x7

@shavnyuy What sort of furniture should one use in the verandas exposed to the elements given there could be rain, heat, humidity and even mild winds?

Here are a few verandah ideas I’ll share; from rustic social hubs to refined outdoor lounges. A thread 🧵

@RujutaDiwekar Is that a Mysore style painting?

The first mangoes of the season from the tree my great great grandmother planted are offered to the Annapurna 🙏🏼

@gregisenberg Like some companies have Digital Transformation Czars, now we will have AI Transformation Czars but that doesn't change anything until the guys of the org are rewired.

@gregisenberg Excellent breakdown of what does it mean to be truly AI Native. A company that is not AI native will have to do the hard yards of thinking inside out to progress towards becoming AI Native.

Next on my Substack ▶️ 𝗪𝗵𝗲𝗻 𝗘𝘃𝗲𝗿𝘆 𝗟𝗮𝘆𝗲𝗿 𝗥𝗲𝗽𝗼𝗿𝘁𝘀 𝗖𝗹𝗲𝗮𝗻: 𝗗𝗶𝗮𝗴𝗻𝗼𝘀𝗶𝗻𝗴 𝗖𝗼𝗺𝗽𝗼𝘂𝗻𝗱 𝗖𝗹𝗼𝘂𝗱 𝗡𝗲𝘁𝘄𝗼𝗿𝗸 𝗙𝗮𝘂𝗹𝘁𝘀 𝘄𝗶𝘁𝗵 𝗚𝗵𝗼𝘀𝘁 𝗔𝗴𝗲𝗻𝘁 Compound faults aren’t rare. They arise naturally from the way real infrastructure changes: a change window touches an NSG and also runs OS hardening scripts; a field engineer adds a diagnostic iptables rule during an incident and forgets to remove it; Docker or a CNI plugin rewrites the netfilter chain while a separate traffic-shaping rule has been left on the interface. This Substack describes new tools in the Ghost Agent tool universe that help work through these kinds of faults. Here's the Substack -> youplusai.substack.com/p/when-every-l…

@levie Also how do you solve multiple teams using the token budget for the exact same thing ie explaining the same code repo or optimising the same big problem -- in large orgs, this is already a problem. With tokens, the cost of such duplicated work just went up a significant notch.

A common trend emerging in larger enterprises is token budgeting as a major topic. As agents can do more and more long running tasks, and thus take vastly more compute, allocation of tokens across teams becomes a very real thing in the enterprise. Companies spend a meaningful amount of time deciding how much to spend on talent, marketing campaigns, events, laptop setups, and even the cost of lunches. Tokens will be no different. Tokens will similarly need to be excruciatingly well-managed because you’ll need to ensure you don’t blow up your budget, and you’ll need to ensure that the tokens are flowing to the highest and most useful parts of work. You don’t want to find out you burned your monthly budget on something relatively low value and then be blocked on the much higher value task later. Doing this at large company scale is extremely hard as you have layers of abstraction on data and visibility into the digital work being done by agents in any central way. This is going to mean that agentic spend will increasingly will expand beyond the confines of the IT budget, and end up in organizational budgets like other expenses. Ultimately team and org leaders will have to be given budgets for this, but even they don’t have adequate visibility and controls in most cases. We’ll need all new software just to solve this problem, and it’s probably an opportunity for startups in its own right. Going to be an all new era of enterprise resource allocation, especially while we compute constrained.

@SagarH62 Craig McDermott

Only 1% cricket fans can recognise this bowler , If you're one of them. Can you guess the bowler's name without the help of Google 👇

@martinvars What kind of evals systems would you recommend? 1-2 examples to learn from?

The honest secret of running AI agents inside a real company is this: the model is not the bottleneck anymore. The bottleneck is what happens when the agent is wrong. I run agents across several of my companies. They sort emails, manage dashboards, block bots on X, draft replies, summarize calls. The first version is always magical. The tenth version is where you learn the real lessons. The model is rarely the problem. The problem is that nothing in the stack tells you, in production, that the agent quietly drifted. It does not crash. It does not error. It just becomes slowly worse at the job, and three weeks later you realize half of its outputs are subtly wrong. What you actually need is unglamorous: evals you trust, logs you can search, the ability to roll a single agent back to last week, and a human review queue for anything that touches money, legal text or a customer. Most teams skip all four because they are not as fun as a new model. The companies that win with agents will not be the ones with the smartest model. They will be the ones whose engineers treat agents like junior employees with bad memory and worse judgment, and build the supervision around them accordingly. Intelligence is cheap now. Accountability is what will be priced.

@ponnappa That's true -- keep it simple without worrying about who wrote it.

I don't see what is wrong with AI written work for most technical topics e.g. where one is expressing how an idea was implemented and what were the gains/pitfalls thereof. Even the LLM writing tells may be mitigated to a large extent by using a /humanize skill in Claude which scrubs most such tells. Creative writing is perhaps where one would want to see the human behind the writing.

yes. 1. it is a poor author using AI to produce ugly tasteless work and/or 2. it is reader's ego, the desire to feel superior to the author by any psychological shortcut necessary rather than by actually judging the content; "oh this is AI written" is now one such shortcut

@clairevo So, back to Claude after the short lived hop to Codex.

claude rate limits just increased, it’s so over we’re so back etc etc

When you say "reusable assets", what are they? Is that Software? The reason I ask is AI allows you to make custom software for every client. So one doesn't have to be a SaaS where the USP is the software built. The USP now with AI is the knowledge of the client's domain. If that can be captured and maintained like a knowledge base, that is the asset to carry forward.

The most obvious, non-sexy business idea right now is to just build an army of Forward Deployed Engineers (FDEs) > be provider agnostic, but opinionated > do bespoke work but build reusable assets across clients > put a temp CAIO in as engagement manager