Shubham Khetan

i set out to replace myself with AI 🤖
so does it really matter who I am today? if you still seem to care, find me on LinkedIn // building ++ ex-VC x Founder

[Shubham Khetan @shubhamknitt]

https://t.co/rqbvFcP51x

[mursaleen. @dev_mursaleen]

I run my company from my watch now

This morning, in one sentence spoken into my watch,
I moved a meeting,
messaged my colleague on WhatsApp,
merged a pull request,
and checked that one of my backend was healthy.

My laptop never opened. @steipete

[Osborne Saldanha @os7borne]

It's been a year since I started my solo investing platform, investing $50 - 100K cheques into fintech and consumption-led startups. Five investments in so far, including @SimetrikHQ, @Lets_BytePe, @flenthomes, Agara, and @shoffr_in. Indian and global fintech leaders invest alongside me in these companies.

This post is partly a reintroduction, and partly me sharing what I actually look for - in case it helps you decide if I'm the right person to talk to.

I evaluate roughly 100 companies a month (yes there are that many fintech and adjacent startups!). Somehow, the past two months, the quality of business ideas and founders has skyrocketed and that's got me excited about fintech again.

Two things separate the founders I want to back from the rest: ambition, and bias for action.

On ambition: recently most founders seem to temper their ambition. Maybe it's VC pressure, maybe it's the fear of sounding naive. So I get pitches that are either too conservative or too vague.

I don't want the market size. I want you to paint me a picture of the world you're building. Dream a little. I want to dream with you. If it doesn't work, we all know the capital goes to zero either way. It's my job anyway to evaluate the risk.

The only thing I'm optimising for is the upside. So show me the the grand ambition. Quantify that ambition as much as possible.

On bias for action: there's no clean way to measure execution speed from a short 30-45 min conversation. So I use a proxy: customer behaviour insights.

It's usually the second question I ask founders. And within the first ten minutes, I have a hunch about whether I want to go deeper.

I want founders throwing customer insights at me like they're throwing darts. High insights-per-minute tells me you've been in the field, not just in the building.

You cannot build in a vacuum and hope customers grab it with both hands, a lot goes wrong when rubber hits the road. Insight only comes from time spent with customers.

I'll admit, my bar here is high. I've spent years in this industry and spoken to thousands of customers - enterprise and retail. It's hard to surprise me with insights. But here's the thing: that's also true for someone with just a year or two in the trenches.

Everyone knows the standard pain points by now.

The differentiation isn't knowing the problem. It's knowing something about the customer that incumbents have stopped noticing, because as businesses scale, their CXOs spend less time with customers, and the ground truth gets stale. An early-stage founder still in the field can out-see them.

Action produces information. Talk to customers, form a sharp insight, build, test, iterate, repeat: faster than everyone else. That loop, shortened, is bias for action.

If you're building in fintech or consumption-led sectors and think you fit this mold, reach out. I'd love to hear the world you want to build, and the insights you're collecting along the way.

[Shubham Khetan @shubhamknitt]

is there still a debate?

after the last few days of Fable maxxing, the debate has settled in my head.

I have not seen myself setting alarm for 2am for anything else.

would love to know what did you really ‘crack’ this weekend with Fable?

for non-starters, you still have a day left before you could use high API costs as an excuse. go for it and trust me you’ll sleep better tonight.

[Shubham Khetan @shubhamknitt]

@hrutu_ way to go!!

[Shubham Khetan @shubhamknitt]

is there still a debate?

after the last few days of Fable maxxing, the debate has settled in my head.

I have not seen myself setting alarm for 2am for anything else.

would love to know what did you really ‘crack’ this weekend with Fable?

for non-starters, you still have a day left before you could use high API costs as an excuse. go for it and trust me you’ll sleep better tonight.

[Shubham Khetan @shubhamknitt]

404: frontier model not found

[Ojasvika Sahu (Oz) @Ozacle23]

Dropping our side project Minimi today.

Minimi is your ambient memory layer that you can connect to Claude.

Claude can now answer your most insane questions -

[Shubham Khetan @shubhamknitt]

@vaibhavbetter most businesses need a managed harness-layer for agent outcomes.

we are living in these parallel realities where AI coworker is inevitable but at the same time most of AI has failed to deliver. I see this gap as the opportunity

[Shubham Khetan @shubhamknitt]

@ponnappa better model ≠ silver bullet
in most cases a more expensive higher-tier model won't solve the problem. Work has to happen around the harness layer so that actually can move the needle.

harness-layer for agent outcomes is the way to go IMO

[Shubham Khetan @shubhamknitt]

it's a marathon, what matters most is finishing it.
Finishing is pre-requisite to winning.
or maybe there is not winning, only finishing.

[Shubham Khetan @shubhamknitt]

@businessbarista AI-native

[arnav @arnav_kumar]

Ambition is a double edged sword.

Ambition without courage to do the hard things leads to a frustrated, depressed, angry person.
Who is perpetually feeling trapped and stuck and takes it out on others, especially their loved ones.

Great ambition blooms when it is coupled with great courage.

Courage to do the hard things, courage to walk a different path, courage to fail, courage to be misunderstood, courage to be mocked.

[Shubham Khetan @shubhamknitt]

@sama few will get elevated, most will get replaced

[Shubham Khetan @shubhamknitt]

@Ayush_cg yeah so true, I love the 'hiring a junior xyz' framing
only when AI is part of the team (and context), it can actually move up the ladder (and add more value).
else the intern in team would remain just that

[Shubham Khetan @shubhamknitt]

every agent being built right now is built to serve one person.
but a world where everyone has an agent is a world where those agents need to find each other. your agent and mine have no way to talk. or to figure out if we should even be talking
who all are solving on this layer?

[Shubham Khetan @shubhamknitt]

@paraschopra I step out of an energetic discussion about agent-to-agent interactions and turns out I cannot really step out

Bangalore: village pretending to be a city. Full of urbanite like me.

[Shubham Khetan @shubhamknitt]

@viraj_sheth I think it's a coping mechanism at core. as long as it is curiosity, it makes sense.

don't you think folks who are inclining towards these lectures are anyways not going to go and do what you suggested? if not the lectures, something else to fill the gap?

[Varuni Sarwal @sarwal_varuni]

Startup idea that I will pay for today: Make an agent that connects to your linkedin, searches through thousands of connections and gives you the top ~100 based on what your're looking for (eg: investment, hiring, customers). @Trifetchai

[Shubham Khetan @shubhamknitt]

@nikitabier I think very much, "this person posted recently" is more useful than their follower count esp for almost any outreach goal.
timing matters as much as targeting. probably more.