On my third startup. I keep this Ghalib as an amulet, but haven't forgotten any incidents.
My most interesting one was where I was asked,
"which IIT?" (I'm none).
In 2001 I intercepted a partner at a VC who was trying to escape his office before our meeting was supposed to start. I ended up pitching him in his parked Lexus from the passenger seat.
At one point he grabbed my laptop placed on his large belly which was pressed against the steering wheel and rapidly flipped through the slides himself.
2001 fundraising hit different
@utsengar thank you.
More such myths have to be de-mythโd.
When top tier investors write such things, it hurts each and every founder that is not in the confirmed mould.
The below is over-generalized and certainly not true for selling to the Top 100K enterprises in the world.
Also, the death is not of the 3-act playbook. It is the gap between the acts.
Plus thereโs one more thing. Distribution.
Let me explain and start with buyers.
1. Very few buyers especially with CIO/CISO title are going to say, โcome, son, replace my chessboard.โ
Instead, they would want you to play with their pieces.
And thatโs where the wedge comes handy.
2. Big platform budgets do not exist for new products until there is a positive category signaling from analysts. This takes a few years. 99% of enterprise buyers wait for categories to mature and a few vendors to come in before pushing the RFP.
Early adopters are only 1000 or so for any category. So you canโt create a billion $ business even with large ACVs such as 250K.
3. Much of the large ACV sales plays are driven by channel & other indirect relationships. Take pao alto for example. 70% is indirect and that take years to build.
Building an indirect channel takes maturity of product and starts with a wedge that everyone is asking for.
4. Most founders are ambitious. Calling them under ambitious is disservice. Ambition is not one thing but is layered with access to capital. It is like saying that if you are currently attempting the climb to a smaller hill you are under ambitious. But given even capital the same founder will get trained with a professional, and then hire a Sherpa, ferry supplies ahead of time.
5. Last point, it is the gap between the acts that is reducing. You will still build it in acts but go out to the customer and expand the deal size with new features and adjacent products.
Aside, much of whatโs getting written as wisdom requires nuance. Who does this apply to and thatโs missing in the public discourse.
Aside, aside, no tokens were harmed writing this.
I get that business insurance is similar Nobel level type of pursuit as ground breaking physics and the Manhattan project. Hopefully the blast radius will be contained.
I donโt think the disagreement is whether hard problems require intensity.
The disagreement is whether intensity has to become a permanent operating model, and whether working seven days a week is the thing that compounds.
My argument is that for most startups, the real compounding advantage is not raw hours. It is clearer thinking, better judgment, learning, and a team that can sustain high-quality work for a long time. You can always spend a lot of time working, but the PMF might never arrive.
There are moments where extraordinary effort is necessary. Launches, incidents, existential deadlines, customer commitments. Those moments matter, and great teams rise to them.
But if the company requires heroics every day of the eek, that usually points to a system problem. It means the operating model depends on burning reserve capacity instead of building it. Company that is constantly on fire is company that is not operating well.
Whenever you put something out there, people will argue and people can argue the way I run Linear. The reason I comment on these things to offer some counter point.
There is a growing clichรฉ in startup culture where founders and startups feel the need to perform intensity publicly. How hard they work, how little they sleep, how many tokens they spend, how busy they are, how much personal sacrifice they make.
You almost never see this from the most successful companies or people. Even if they work that way, they usually donโt make it the story, because they have more important things to talk about, like the product, the customers, the insight, the strategy, the quality of the work.
Thatโs my issue with the narrative and why I think startups shouldn't blindly follow it. Not that is bad to work hard but grindmaxxing narrative can become the greater goal and become counterproductive. The performative intensity becomes the thing, and loosing sight of what actually matters.
Lets check back in 7 years.
Quick summary of what is happening with LLM model companies in China. 1) There is more VC $ available for open-weights than you think, 2) they are generating real revenue (as did open-source sw/saas companies in the West).
https://t.co/vluRvWJsId
@martin_casado Agree.
From a technical pov โ fine tuning etc are still expensive compared to RAG. So this is going to play out over the next 3-4 years.
The open models chose to play the long game coz they missed out in the early innings on capital and GPU.
SITUATION EXPLAINED: How much are frontier labs actually spending on training data?
.@SeanZCai: "Frontier labs are spending about $10 to $15 billion per lab on data."
"Really good long horizon tasks go up to $20,000 each. A complete browser-use version of SAP was rumored at $500,000."
"Despite everybody thinking the market is super crowded, we still don't have enough good quality data vendors that actually understand how to deliver product plus services in a way researchers are looking for."
"I have not seen a contract for genuinely good data gets turned down because of budgetary concerns yet."
As an AI builder, you'll get asked this question again: "Can frontier models not build what you already have?"
Show this Opus 4.8 screenshot and then show your own architecture containing:
- guardrails
- task graph
- deterministic execution
- validation
- state comparison
- harness
- idempotency
- ... not counting observability, audit trail & evidence
Using a model is different than giving control to a model and hoping it would do the right thing.
This was a good week answering these questions.
@adityarohit9 the only way is to adapt -- we are seeing a flood of CVEs in packages that we use.
this is going to get worse and then we will be in the land of milk & honey without security bugs ๐