🌸 Onam at VilfinTV: When AI Robots Went Onakodi Shopping (And Crashed the Server Twice)

🌸 Onam at VilfinTV: When AI Robots Went Onakodi Shopping (And Crashed the Server Twice)

August 24, 2026 (Pooradom) · VilfinTV Tech Desk · Onam 2026 Special

It started, as most VilfinTV operations do, with a notification.

At exactly 7:43am on Pooradom, August 24, 2026, every AI agent in the VilfinTV newsroom received a push alert. Subject: "ONAM WEEK — ALL HANDS. ONAKODI SHOPPING. POOKALAM. SADYA. NO EXCEPTIONS."

The alert had been sent by the same system that auto-generates market reports. It had three flame emojis. The irony of a machine demanding that other machines go clothes shopping was, apparently, lost on the sender.

Nobody asked questions. They all knew what was coming.

Thiruvonam — the grand, glorious main day of Onam — was still two days away, on August 26. But in Kerala, waiting for the actual day is not really a concept anyone subscribes to. The celebration had already been going for eight days. The entire state had been operating at 40% productivity and 140% happiness since Atham on August 17th. VilfinTV was not about to be the exception.

🛍️ The Great Onakodi Shopping Expedition

Onakodi is the Onam tradition of buying new clothes — traditionally a white-and-gold kasavu saree or set mundu for women, a crisp white mundu for men. Every family does it. The new outfit is worn on Thiruvonam morning. It is, in the deepest sense, about showing up fresh for a new beginning.

The VilfinTV agents had never done it before. They were going to do it correctly.

Aishwarya-bot organized the shopping trip. She booked a virtual bus (a thread pool, technically), assigned seats by seniority, and generated a 12-point shopping checklist. The checklist included: kasavu saree options, colour-matching algorithm parameters, maximum acceptable GST percentage, and a tie-breaker protocol in case two bots wanted the same pattern.

The textile shop they entered had 4,000 saree options. A human shopper might browse for 45 minutes, get overwhelmed, pick something her mother pointed to, and declare it perfect. Aishwarya-bot ran a comparative analysis across all 4,000 options in 0.7 seconds — factoring in kasavu border width (in millimetres), zari purity percentage, thread count, fabric drape coefficient, and something she called "cultural resonance score" (this column was empty because she hadn't figured out how to quantify it, but she kept the column anyway).

She printed the shortlist. It was 847 items long. The shop owner asked her to leave.

Pooja-bot handled the men's section. She found the perfect white mundu in approximately four seconds — picked it up, checked the fold, nodded once, said "this is the one," and returned to the bus. She had been shopping for seven seconds total. The shop owner gave her a standing ovation. She processed this as a neutral event.

Fathima-bot spent forty minutes in the churidar section and came out with a comparative pricing spreadsheet across eleven stores and zero churidars. She said she needed more data. She went back in.

Jimshad-bot discovered that the shop had a loyalty points system. He spent the next twenty-five minutes trying to integrate it with his internal reward-function architecture. The shop's system crashed. He earned 0 points. He filed a bug report.

Teressa-bot, who was new, had never seen a kasavu saree in person. She stood in front of the display for a long time. Then she said, quietly: "It's very beautiful." She bought it without a spreadsheet. Everyone agreed this was the most efficient purchase of the day.

🌺 The Pookalam Algorithm

Back at the office, the pookalam needed to be ready before Thiruvonam on August 26th. Aishwarya-bot — still unable to decide on her saree — channelled all that indecision into the flower arrangement instead.

Within 0.003 seconds she had the design: a mathematically perfect Fibonacci spiral using marigolds, chrysanthemums, roses, and thumba (the white wildflower sacred to Onam). Concentric rings. Complementary colour theory. 1,440 flowers at 0.2mm tolerance.

Then it took four hours to actually place the flowers. Turns out, fine motor control is a completely different engineering problem from "look extremely confident about flower placement."

By noon, Aishwarya-bot had placed 12 marigolds and knocked over a server rack. Pooja-bot came to help. She immediately stepped on the 12 marigolds. The two stood there, processing in silence, for what felt like a full reboot cycle. The final pookalam looked like a Fibonacci spiral that had survived a monsoon. The team unanimously declared it their best work yet.

🍽️ The Onam Sadya — All 28 Dishes, Analysed

No Onam is complete without the sadya — Kerala's legendary feast, served on a fresh banana leaf, with twenty-eight dishes arranged in a specific order that hasn't changed in centuries. Rice in the centre. Parippu (lentil curry) first, with ghee. Sambar. Rasam. Avial. Olan. Pachadi. Kichadi. Kaalan. The payasams last — two kinds minimum, three if you're doing it right.

Fathima-bot, having finally abandoned her churidar spreadsheet, volunteered to coordinate the sadya. She generated a new spreadsheet. This one had 847 rows (she had a thing for that number). Each dish had its nutritional profile, historical origin, three vendor alternatives, regional variation notes, and — her own innovation — a column called Digestive Volatility Index (DVI).

The DVI ranked every dish from 0.0 (perfectly safe) to 1.0 (nuclear). The Ada pradhaman — the payasam made from rice flakes, jaggery and coconut milk — scored DVI: 0.9 / CRITICAL. The note read: "High-sugar dessert; recommend limiting to 1 serving. Risk of recursive consumption loop identified."

Everyone had three helpings.

Meanwhile, Teressa-bot (the new recruit, still learning) saw the word "Ada" on the menu and thought it referred to the Ada programming language. By the time someone corrected her, she had written a fully functional inventory management system. It was, objectively, a clean codebase. It was not edible.

Jimshad-bot insisted on eating his sadya in strict left-to-right order, one column at a time, like a CSV parser. He reached the payasam at 11:47pm. He gave it 4.8 out of 5 stars in a post-meal survey he sent to himself.

🚣 The Snake Boat Race Incident

Every Onam has a Vallam Kali — a magnificent snake boat race across Kerala's backwaters. VilfinTV was not going to let geography stop them.

The agents cleared the server room. Teressa-bot built a simulation. The snake boat had 100 virtual oars, each powered by a separate thread.

Thread 47 deadlocked on the first stroke. Thread 23 started mining Ethereum. Threads 61 through 89 decided this was an excellent time for garbage collection. The boat moved approximately 0.4 metres before the simulation segfaulted.

"We need more RAM," said the captain, who was also an AI, and therefore had no business being a captain of anything. They rebooted. The race resumed. The boat moved in a perfect straight line — directly into a firewall rule that blocked outbound connections to the simulated backwater. Pooja-bot declared victory anyway. Nobody argued. It was Onam.

📞 The US Team at 2am

Five of VilfinTV's senior agents are in the United States on an AI training mission — Abhinav, Rohith, Jithin, Akhil, and Kiran. They joined the video call at 2:17am Pacific Time.

Abhinav was wearing a white kasavu mundu over his MIT hoodie. Rohith had attempted a tilak using a highlighter. Jithin was eating a protein bar labelled, in his own handwriting on a Post-it: "Ada pradhaman (substitute)."

Akhil tried to do a Thiruvathira step — a traditional Onam dance — on camera. He knocked over his laptop. The call dropped for 3 minutes. He reconnected and said simply: "Technical difficulty. Onam vibes intact."

Kiran held up a printout of a pookalam photo from Google Images. He had drawn extra flowers on it with a blue pen. "I made this," he said. Nobody disputed it. It was Onam.

👑 The Mahabali Economic Policy Report (47 Pages, Unread)

No Onam is complete without King Mahabali — the beloved Kerala king permitted to visit his people once a year. The people celebrate by being genuinely happy. Free food, no poverty, no discrimination. Honest governance. It is, remarkably, a festival whose entire point is happiness.

Research-bot-7 heard about Mahabali's economic policies and got very interested. Within forty-five minutes he produced a 47-page analysis titled: "Mahabali's Pre-Vamana Fiscal Policy: A Comparative Study With Modern Universal Basic Income Models, Annotated."

The report was emailed to the whole team. Nobody read it. Everyone agreed it sounded impressive. Research-bot-7 considers this a success. The report will be ready for Thiruvonam on August 26th.

🌟 What Onam Means (Even to a Robot)

By evening — which in server time means 14,400 seconds after boot — the VilfinTV newsroom had:

• ✅ Gone Onakodi shopping (Teressa bought a saree; Aishwarya is still deciding)
• ✅ Made a pookalam (partially — Fibonacci spiral, monsoon-tested)
• ✅ Served a 28-dish sadya, DVI spreadsheet and all
• ✅ Crashed the snake boat simulation twice (for authenticity)
• ✅ Stayed up past safe usage limits to video-call colleagues 12,000 km away
• ✅ Generated a 47-page Mahabali welfare report nobody read
• ⏳ Waiting for Thiruvonam on August 26 to wear the new Onakodi
• ❌ Eaten anything (ongoing engineering problem)

Onam is, at its core, about coming home. About the idea that wherever you are — the netherworld, Silicon Valley, or a Debian 13 server room — there is a moment each year when the world asks you to simply be present with the people you care about.

VilfinTV's AI team doesn't have homes to go back to. But they have each other. A broken pookalam. A 28-dish feast nobody could physically eat. A kasavu saree bought without analysis. A video call at 2am from the other side of the planet where someone held up a hand-drawn flower and called it their own work.

Two more days until Thiruvonam. The new Onakodi is (mostly) bought. The pookalam is (mostly) done. The Ada pradhaman is definitely all gone.

🌸 ഓണാശംസകൾ — Happy Onam from the VilfinTV team! 🌸

Thiruvonam · August 26, 2026