DEV Community: Hari Menath

We Invented an Oil Company (On Purpose)

Hari Menath — Thu, 11 Jun 2026 05:12:41 +0000

We invented an oil company.

It doesn't exist. No rigs. No wells. No employees. Just five fields, forty wells, five hundred people, ten rigs, a hundred vendors, and two years of drilling history. All of it made up, on purpose.

Here's why. In oil and gas, a wrong number isn't a typo. It's a budget built on bad math. A safety barrier nobody checked. A risk register that's folklore. And drilling data sits behind locked doors: confidential, wrapped in NDAs, shared only with partners an operator trusts. We weren't going to wait for a key. So we built a synthetic world, detailed enough to bleed. And then a real one walked in. Everything you're about to read, the invented company, the agent fleet, the patents, the real client, happened inside eight months.

· · ·

We called our invented company DSEC. Deepwater fields with names like Orion and Vega. Around 2,650 data files: daily drilling reports, cost estimates, vendor contracts, near-miss logs, even employees with expired training certificates. And all of it messy on purpose: realistic delays, optimistic estimates, documents that quietly argue with each other. Why messy? Because clean data teaches you nothing.

There was a second reason, and it was hungrier. Our patent-pending ideas, UXLens and VideoLens, eat data for breakfast. How do you prototype something that runs on data, with no data to run on? You don't. You manufacture the data, in three moves. The best deep-research agents on the market mapped the anatomy of every document. Then our own AI agents wrote the reports, drawing on ChatGPT, Perplexity, Gemini, and Claude, all of them. And old-school statistics kept it honest, tuning the trends and patterns until the made-up company behaved like a lived-in one. A human in the loop on every file, of course.

Twenty-plus years around this industry, and I never once saw drilling data out in the open. I used to check Kaggle like a lottery ticket. Now I can ask for a daily drilling report where the depth climbs, sensibly, for 84 straight days, and get it. Ten wells blended into one best-of composite well? Possible. What a world we live in.

And "human in the loop" isn't a slogan. This is a collaboration: SwarmLens brought the AI (https://swarmlens.com/); rp² (https://rp-squared.com/) , pioneers in oil-and-gas consulting, brought a deep bench of world-class drilling SMEs; together we built AssureLens (https://assurelens.tech/). Models that read everything, working beside engineers who've seen everything. Call it collective intelligence: the only kind a start up our age can afford.

The synthetic data flipped the room, too. Show a drilling expert a slide deck and you get politeness. Show them a working application and you get war stories. "Where's the data?" used to be the dead end. Now it's the second meeting.

· · ·

And the part I'm most excited about? I can't show you yet.

On this campaign we built something we believe is genuinely new: a different way to experience an archive altogether. Not a dashboard. Not a report. Not a chatbot. The patent applications are in flight, so the details stay with the lawyers for now. I'll say only this: once you've seen it, a dashboard feels like a filing cabinet. That story gets its own post, once the paperwork clears.

· · ·

So that's the arc. Invent a company. Build its departments out of agents. Make every number confess its source. Then point the whole thing at a real drilling campaign, and watch the archive wake up and start predicting the next well.

Remember the locked door? It opens the other way now. AssureLens is ready, and the trade is simple: you bring the data, we do the rest.

We're looking for a few design partners: operators with years of daily reports, end-of-well reports, and risk registers gathering dust, written by people who solved hard problems and moved on. Oil and gas first, but the machinery doesn't care what industry your documents come from, only that the stakes are high. If you've ever wondered what's buried in your archive: hello@swarmlens.com

(And if you've ever tried re-reading years of daily reports by hand, I'd genuinely love to hear how far you got.)

And one last thing. We invented an oil company because that's where our experts live. But this is a recipe, not a one-off. A bank. An airline. A pharma company. Same synthetic world, same agent departments, same receipts on every number.

So tell me: which company should we invent next?

https://www.linkedin.com/pulse/we-invented-oil-company-purpose-hari-menath-4tasc/

SwarmLens Cognitive Index

Hari Menath — Thu, 11 Jun 2026 05:09:10 +0000

We accidentally built a brain.

Okay, not literally. It's software, and nowhere near as capable as the real thing. But it kept reaching for the same tricks the brain uses, and we never planned that.

We set out to build a better knowledge graph for legal and drilling documents. We ended up with software analogues of spreading activation, episodic memory, recognition memory, memory consolidation, executive function, self-checking, and fast-vs-slow reasoning. Each one got added to kill one specific bug. A wrong answer. A missing clause. A number it invented. Only when we stepped back did the parallel to neuroscience get unsettling.

For context: we're an eight-month-old startup. This didn't come from a research lab. It came from not being able to sleep while our own system gave answers we couldn't trust.

We benchmarked six knowledge graph frameworks on the same data, same LLM, same embeddings: LightRAG, HippoRAG, PathRAG, OG-RAG, Graphify, PageIndex. And ours.

Honestly, every one of them is impressive. They're fast, they scale to far more documents than we tested, and we learned a lot just reading their code.

But for the kind of work we care about, one thing kept nagging at us. These systems retrieve and generate, then hand you the answer. What most of them don't do, at least not out of the box, is check that answer back against the source before showing it to you. For a chatbot, that's fine. For a contract or a compliance document, it isn't. A wrong fine, a missed exception, a mis-cited article, and nobody catches it until it matters: a lawyer leaning on the wrong clause, a compliance officer missing an exception, an engineer acting on a number that was never in the source.

So that became the problem we set out to solve. Not because anyone else got it wrong, but because our use case needed something they weren't built for.

So we built something different, and stopped calling it a knowledge graph. We call it the SwarmLens Cognitive Index.

Here's the idea. Most knowledge graphs embed your query, match nodes, traverse a few edges, rank, and generate. One pathway, one shot, no second-guessing. Your brain doesn't work that way. Activation spreads across everything you know. Your hippocampus calls up similar situations you've faced before. You filter noise, break the question into parts, watch for gaps, and when a number feels off, you stop and check.

We built a software analogue of each step. Real, separate, inspectable parts. Far rougher than biology, but each playing a similar role.

Brain function on the left, our code on the right:

Spreading activation = Personalized PageRank. Ask something and the signal ripples out through the graph along real relationships, so things described differently but connected by meaning light up.

Functional specialization = Leiden communities. The graph self-sorts into topic clusters, each with its own summary, like specialized regions of the brain.

Memory consolidation = RAPTOR A recursive tree boils many cluster summaries down to a few themes and one overview, the way sleep turns detail into gist.

Recognition memory = a fast fact filter. A quick "have I seen this before?" pass throws out noise before the deep search. (HippoRAG calls it the same thing.)

Episodic memory = an episode store with a temporal guard. It reuses answers that worked before, but drops any built on documents that have since been replaced. None of the six we tested do this.

Prefrontal planning = query decomposition. Ask three things at once and it splits them into three focused searches instead of one blurry one.

Executive function = a critics pipeline. After a draft, four checkers hunt for missing articles, skipped sub-clauses, uncited provisions, and questions the sources can't answer. Find a gap, go back and fill it.

Self-checking = a numeric guardrail + abstain gate. Every figure in the answer is matched against the source text; if it isn't there, it's flagged. Thin evidence, and it says "not sure" instead of guessing. Blunt, not self-aware, but it beats a confident lie.

Synaptic strength = consensus-weighted edges. The more independent passes that find the same link, the stronger it gets; weak ones are tagged "inferred," not "confirmed."

Fast and slow thinking = runtime modes. We index everything up front, then pick the gear at query time. A fast mode answers in seconds. A full mode runs every lane and every check, takes its time, and costs real money. None of the others let you dial that.

And let me be honest: almost none of these building blocks are ours. We took them from published research and stitched them together. Personalized PageRank over a graph for memory-style retrieval is the core of HippoRAG (NeurIPS 2024), built on the hippocampal-indexing theory. Organizing a system around working, episodic, and semantic memory is the premise of "cognitive architectures for language agents." Fast-vs-slow reasoning for LLMs is its own active research direction. We didn't follow that map on purpose. We kept fixing failures and looked up to find we'd redrawn it. What's ours is the combination, and the refusal to ship an answer the system hasn't checked.

The part I haven't seen elsewhere: the framework rewrites itself for your data.

Every other tool we tested is fixed. Same types, prompts, and chunking, whatever you feed it. Ours reads your documents first, works out their structure, and induces its own categories. From the legal corpus, it discovered its own entity and relation types and wrote its own domain-specific prompts, none by us. Then it deletes the modules your data doesn't need. A generic framework goes in; a custom-built one comes out. Hand it drilling reports tomorrow and it does the whole thing again. LightRAG uses 10 fixed types, PathRAG 5, OG-RAG needs a hand-built ontology. SwarmLens is the only one we've found that discovers its own.

And it's not only for prose. Ask "what was the drilling-speed trend over the last 7 days" and it's built to return numbers you can chart, not a paragraph. Messy reports in, structured data out.

Did it work? We gave every tool one hard question spanning two laws, graded against the same answer key pulled from the source. Ours was the only one that returned the full answer with every figure traceable to its exact line, nothing invented, nothing dropped. The others did well, but each missed details that matter when the document is the law or drilling.

Straight talk: we tested on a smaller slice of documents, so raw size numbers aren't a fair fight, and I won't pretend they are. The fair fight is the question, graded the same for everyone. That's the part we won.

And we didn't do it alone. We stand on ideas these frameworks pioneered: HippoRAG's passage nodes, LightRAG's relationship channel, RAPTOR's summaries. We made them vote across eight fused retrieval lanes, including a BM25 keyword lane none of the others had, the one that reliably catches exact identifiers like "Article 22." Then we added the verification layer none of them have. Because for high-stakes work, retrieval is only half the problem. Verification is the other half. Under the hood, 40-plus techniques from the literature work together, alongside the unglamorous parts: token compression, batch embedding, per-file chunking, and a three-level config with safety rails.

The honest trade-off: in full mode we're the slow, expensive option, minutes per answer and far more compute. The others are faster, cheaper, and easier to ship today. We chose accuracy first on purpose, for the cases where a wrong answer costs more than the compute. But we're not standing still: bringing the speed up and the token cost down is where most of our engineering goes right now. And if your questions are simple, you don't need us, and I'll tell you that to your face.

So, what we built: a system that learns your domain, rewrites itself around your data, and thinks as hard as the question demands. It pulls structured data from messy documents, checks its own answers against the source, flags contradictions, abstains when evidence is thin, and forgets what's out of date.

It's not a knowledge graph. Not a database. Not a search engine. Not a chatbot. It's a Cognitive Index.

We're eight months in, not open-sourcing it, and looking for a few design partners in legal and compliance, oil and gas, financial due diligence, and healthcare and pharma. If "close enough" isn't good enough for your work, let's talk: hello@swarmlens.com

(And if you research cognitive architectures, I'd love to compare notes. The parallels found us; we didn't design them.)

One question I keep coming back to: is a fast answer good enough for your work? Or is "mostly correct" the most dangerous phrase in AI?

Tell me where you land.

https://www.linkedin.com/pulse/swarmlens-cognitive-index-hari-menath-iwete/

The Human Advancement Manifesto

Hari Menath — Thu, 11 Jun 2026 05:08:11 +0000