Shatarupa

Building @ThreddleAi, AI for retail merchandising & inventory. Ex-SAP. Writing about AI, retail, and the hard parts of building a startup.

[Shatarupa @Shatarupasarma]

Today, we launched Threddle on Product Hunt.

Retail teams aren’t short on data.

They’re short on time.

Too much of the day is spent chasing trends, checking competitors, reviewing inventory, and connecting dots across different tools.

We built Threddle to do that work for you.

It spots what matters, helps you decide faster, and gets things moving.

If that sounds interesting, we’d love your support.

🔗 https://t.co/o458GIPD6M

[Shatarupa @Shatarupasarma]

AI won't replace startups.

AI will replace excuses.

"I can't design."
"I can't code."
"I can't write."
"I can't market."

Every one of those is now a prompt away from being good enough.

Execution just became visible.

[Shatarupa @Shatarupasarma]

@adxtyahq Amen to that

[Shatarupa @Shatarupasarma]

@DavidSacks Lmao

[Shatarupa @Shatarupasarma]

@garyvee I like the way you think

[Shatarupa @Shatarupasarma]

@themoviedadsc They want everything in a silver platter presented to them huh

[Shatarupa @Shatarupasarma]

every founder eventually becomes one of these:

"building in stealth"
"building in public"
"taking a break to think"
"thinking about building in public again"
"actually shipping"

the gap between 4 and 5 is where most startups die.

[Shatarupa @Shatarupasarma]

@asparagoid AI companies accidentally invented astrology for software engineers.

[Shatarupa @Shatarupasarma]

@theo nothing ages worse than "enterprise-grade" in a market moving this fast.

[Shatarupa @Shatarupasarma]

@dioscuri LLMs made me realize the bottleneck was never information. it was having enough context to notice when something smells off.

[Shatarupa @Shatarupasarma]

@yacineMTB us

[Shatarupa @Shatarupasarma]

@ArielKwiat ayooo

[Shatarupa @Shatarupasarma]

@2ToneShawn congratulations, you're the richest poor person in your own neighborhood.

[Shatarupa @Shatarupasarma]

@kanavtwt FRRRR😭

[Shatarupa @Shatarupasarma]

@SantiTorAI every chinese demo looks like they skipped the "minimum lovable product" phase and went straight to "physics engine lmaoo

[Shatarupa @Shatarupasarma]

@NickADobos feels like we've gone from "don't trust generated code" to "don't trust yourself to read generated code.

[Shatarupa @Shatarupasarma]

@aakashgupta The scariest part isn't the implosion. It's that the data was already telling the story and everyone kept arguing with it.

[Shatarupa @Shatarupasarma]

@bryan_johnson my favorite genre of advice is "this isn't necessary, but here's the extremely optimized version i personally do."

[Shatarupa @Shatarupasarma]

@sebcrossa Email is the cockroach of software. Every founder thinks their AI is finally the one that'll tame it. Then you realize the real moat is convincing people to change 15 years of muscle memory.

[Shatarupa @Shatarupasarma]

This is exactly why we built Threddle.

I’ve watched this play out too many times a single SKU quietly bleeding margin for weeks and everyone keep staring at the total revenue.

The markdown isn’t the failure but weeks of silence is.

Our inventory agent is designed to end that silence.

[Threddle @ThreddleAi]

Breaking down how a $40 t-shirt turns into a $16 t-shirt, and why nobody caught it in time.

Week 1: SKU launches. Sells fine. Nobody's watching it closely because it's not a top performer or a total dud.

Week 4: Sell-through is 15% behind the same category last season. Nobody notices because the dashboard shows total revenue, not this SKU against its own benchmark.

Week 7: A competitor starts running the same silhouette at 20% off. Your price stays put because nobody's tracking their catalog against yours in real time.

Week 9: Someone finally pulls a report. The SKU is sitting at 60% of forecasted sell-through with 8 weeks of stock left on the floor.

Week 10: The call gets made. Mark it down 40% to move it before it becomes next season's problem.

That's the whole story. Nine weeks between "this is drifting" and "someone looked." The markdown isn't the failure. The nine weeks are.

Threddle's inventory agent watches for exactly this: SKU velocity against category, sell-through against last season, competitor pricing on similar product. It would've flagged this one in week 4, not week 9.

Five weeks earlier. That's the difference between a small price adjustment and a clearance rack.