Hey — It’s Nico.

Welcome to another edition of Failory.

Our newsletter has had the same structure for a while, and we’d love your help shaping its future. In just one minute you can tell us what you enjoy, what could be better, and what you’d like to see next. Take the quick survey here.

Now, let’s dive in. This issue takes 5 minutes to read. If you only have one, here are the 5 most important things:

• MyPickup, a ride-hailing startup with a fleet of electric rickshaws has shut down — learn why below

• The 15 AI distribution plays that build real moats.

• OpenAI launches ChatGPT Pulse to proactively write you morning briefs.

• Oura ring maker raises $875M Series E.

• A Stanford research suggests that AI slop in the workplace (workslop) is costing companies a lot of money — learn more below

Getting compliant doesn’t have to take months.^AD

With Delve, it takes 15 hours.

SOC 2, HIPAA, ISO, and more, automated with AI.

That’s why 500+ of the fastest-growing AI companies (Lovable, Cluely, Wispr Flow) already run compliance on Delve.

But compliance is only half the battle. The other half is enterprise reviews.

At Delve, they’ve sat through 100s of hours of security questionnaires.

They’ve seen every pointless checkbox, every SaaS-era question, every deal stalled because of them.

So today, they’re publishing something different: a manifesto with Bland AI on how security questionnaires should work in the age of AI.

📖 Read the manifesto here.

⚡️ Book a demo here with code FAILORYL1KOFF for $1,000 off + white-glove onboarding.

This Week In Startups

🔗 Resources

• Become the #1 thought leader in your industry on LinkedIn with Playbookz. 20k reach in 28 days or you don’t pay. * 

• The 15 AI distribution plays that build real moats.

• The rising consumer to enterprise AI playbook.

• Building pipeline: pre-PMF and post-PMF

📰 News

• OpenAI launches ChatGPT Pulse to proactively write you morning briefs.

• Y Combinator launches ‘Early Decision’ for students who want to graduate first and build later.

• Neon, the No. 2 social app on the App Store, pays users to record their phone calls and sells data to AI firms.

• Silicon Valley bets big on ‘environments’ to train AI agents.

💸 Fundraising

• Doorstep raises $8M seed to help find missing food deliveries.

• Oura ring maker raises $875M Series E.

• Chip cooling startup Corintis raises $24 million.

• WeTravel raises $92M to scale its multi-day travel operations with AI.

* sponsored

Fail(St)ory

Uber, But Pre-Booked

Ride-hailing is a brutal game. Margins are thin, demand is erratic, and customer loyalty evaporates the second a cheaper ride appears. Into this jungle, MyPickup thought it could tame chaos with subscriptions, electric rickshaws, and salaried drivers.

It lasted less than two years.

What Was MyPickup:

MyPickup launched in 2022 with a promise: zero-stress commuting. Instead of booking a rickshaw every morning, office goers and school kids could buy weekly or monthly subscriptions.

Riders had to commit to at least five pre-scheduled trips per week. For example, a daily 9 a.m. ride to the office or school. No surge fees. The same driver every day.

The fleet was all-electric and fully owned by the company. Drivers were on salary, not commissions. It was a deliberate rejection of the Uber/Ola “gig worker” model. Customers didn’t deal with cancellations or bargaining. Parents liked the idea of a familiar “auto uncle” taking kids to school.

At its peak, the company had 19 vehicles in Bengaluru. In May 2024, those rickshaws delivered about 4000 rides to more than 100 subscribers, with retention above 80%. Customers seemed happy. Investors weren’t.

The startup raised about $179k in mid-2023. The money went to build the app, upgrade scheduling, and expand the EV fleet. Before that, operations were scrappy: bookings handled on WhatsApp or phone-calls.

In a post-mortem post, founder Abhijeet Jagtap admited that  the company tried several pivots, but never cracked product-market fit outside peak hours. Most of the fleet sat idle during off-peak periods, and the thin demand wasn’t enough to cover fixed costs like salaried drivers and owned vehicles.

By September 2024, the ride count wasn’t scaling, the burn didn’t shrink, and the capital dried up. No institutional investor was willing to extend the runway so MyPickup was forced to shut down operations.

The Numbers:

• 💸 Funding: $179k in July 2023

• 📈 Peak usage: ~4000 rides in May 2024

• 👥 Subscribers: 100+ with >80% retention

• 📅 Lifespan: Feb 2023 – Sep 2024

Reasons for Failure: 

1. No demand density: Fixed assets and salaried drivers need high utilization. Non-peak hours left rickshaws idle, bleeding cash. Ride-hailing depends on spikiness, not predictability.
   

2. Mismatch with investor appetite: Subscriptions and eco-friendly fleets sound great in pitch decks, but the metrics didn’t scale fast enough. 4000 rides a month is tiny in a market where Ola clocks millions daily.
   

3. Cost structure too heavy: Owning vehicles and employing drivers created stability for riders but killed flexibility for the business. Competitors shift demand risk onto drivers. MyPickup absorbed it.
   

4. Timing against giants: Uber, Ola, and Rapido already subsidize rides and burn capital to hold share. A small EV fleet in one city can’t compete for commuters who just want the cheapest, fastest option.

Why It Matters: 

• Good retention doesn’t mean PMF. Customers can love your service, but if demand isn’t wide or frequent enough, you don’t have a business.

• “Ethical” models cost real money. Salaried drivers and owned EVs solve real pain points, but unless you have patient capital or government backing, the economics won’t sustain.

• Small numbers don’t buy time. Founders often think early adoption plus a little traction equals a bridge round. Investors want growth curves, not loyalty anecdotes.

Trend

Workslop

The hottest new productivity hack is fake productivity. You’ve seen it: perfect-looking slides, tidy summaries, runnable code that quietly breaks the thing that mattered. It looks finished yet moves nothing forward.

Researchers just gave this mess a name: workslop. BetterUp Labs and Stanford surveyed thousands of workers and realized that 40% received slop in the last month. It travels peer to peer and up the chain, wasting everyone’s time.

According to the researchers, AI didn’t invent laziness, it industrialized it. Tools now spit out polished output in seconds, and too many people pass that off as the work. The burden shifts to whoever receives it to fix, reinterpret, or redo.

Why it Matters:

• Slop taxes the receiver, not the sender. Rework is invisible until your sprint slips or a client calls. The study pegs the average cost at $186 per employee per month, which compounds into millions per year at scale.

• Reputation erodes quietly. About half of respondents view slop senders as less creative, capable, and reliable. That label sticks inside teams and with clients, which is how pipelines dry up.

• The perception of AI productivity is shifting. Early hype promised automation and efficiency, but more data points suggest the opposite. If output keeps rising while outcomes stall, the market will demand tools that guarantee quality instead of tools that just generate more.
  

What is Workslop:

Workslop is AI-polished work that lacks the thinking to move a task forward. It’s a slide deck with no point, a summary that misses context, code that compiles but violates the spec. It’s shiny deliverables with empty calories.

The pattern is simple. A teammate pumps a prompt into a model, gets believable text or code, then ships it with minimal inspection. The recipient becomes the real author under deadline pressure.

According to the researchers, one of the main consequences of workslop is that it lowers how colleagues see each other. 

This isn’t “AI bad.” The same tools can accelerate good work, but only when paired with domain judgment and evidence. Without that judgment, you get volume over value.

What this Means for Startups:

First, there are the more obvious lessons. If workslop is pricey in a 10,000-person org, it can be lethal in a team of five. Every hour of rework is runway you never get back.

Startups don’t have buffers. When slop code forces a rewrite, your launch slips and a week vanishes. 

But besides this more obvious point, a pattern is showing up. A while back I shared the MIT result that 95% of corporate AI pilots failed to make a profit. Now add this week’s research: output is up, but outcomes aren’t. The promise of AI “full automation” is not landing in the real world.

If this idea keeps spreading, the winners in AI won’t be the ones promising to automate everything. They’ll be the ones that pick one small slice of a process and nail it every time. Maybe that’s a finance reconciliation tool that never drops a line item, or a code agent that only touches tests and never breaks prod. Narrow scope but great reliability.

There’s a startup play here: build for precision, not spectacle. Design evals, safeguards, and feedback loops so airtight that people trust the output blindly. In a market drowning in half-baked AI drafts, trust is becoming the product.

Help Me Improve Failory

That's all for today’s edition.

Cheers,

Nico