Shaun

CEO @with_native | AI Partner for the worlds most ambitious enterprises

[Shaun @shaundevan_]

remember when Shopify CEO Tobi Lutke told the whole company "no new hires until you prove AI cant do the job"

everyone laughed

16 months later Sam Altman says Tobi is the only CEO who got AI right, and that "he's 6 to 8 months ahead of every other CEO"

here's how Tobi thinks:

"we're going to look back on 2026 as the year that every business was up for grabs"

someone is going to build the AI native version of Shopify, "and it's going to be me"

so every night he pretends to be Shopify's competitor and works out how he'd rebuild Shopify from zero with AI

someone is doing this to your company right now

[Shaun @shaundevan_]

@verrsane @HamelHusain @sh_reya @danolsen @LeanProdMeetup @raindrop_ai SF's luma makes me green with envy

[Shaun @shaundevan_]

@mattepstein brotha you did not do typesafe's launch

[Shaun @shaundevan_]

@vasuman Is this where Varick came from?

[Shaun @shaundevan_]

OpenAI's models found a flaw in PHP, which runs an estimated 70% of websites.

Then it took two months and a coalition of maintainers and security people to fix it.

Finding the problem is always easier than fixing it

[OpenAI @OpenAI]

OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.

Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors.

Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact.

This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare.
https://t.co/TW3jrSbB3f

[Shaun @shaundevan_]

@verrsane same thing for me but with cc

[Shaun @shaundevan_]

@lazeskid the fix is actually writing docs & memos imo

[Shaun @shaundevan_]

Remember when Shopify CEO Tobi Lutke made every team prove AI couldn't do the job before they could hire

17 months later he says AI made everything worse

Shopify even has a name for what's going on:

"we call those 'slop grenades' that people toss at each other"

"the failure case now of lazy work is not lack of output. it's actually overoutput now"

someone has AI write the long email or make the code change, doesn't read it, and passes it on

"you don't really read it, and now it has to be reviewed by your colleagues, and they are like, 'this doesn't look right'"

If Shopify has slop grenades, so does your company

Most AI rollouts change nothing

Everyone gets a faster way to produce work, the work still moves through the company exactly the way it did before, and it just lands on the next desk with more of it to read

For AI to actually work it has to go into the process itself

Take one workflow in one department, map how the work moves today, and rebuild it so agents do the steps inside the systems the team already uses

Every output gets checked before it moves on, and a person only steps in when the agent isn't sure

So before you hand your team another AI tool, change the work it goes into

Other you're just handing out more slop grenades

[Shaun @shaundevan_]

@LoganTGott If you're having your engineers write code you're genuinely doing something wrong

As long as an engineer can explain what they did w/ AI & why then there's no issues

[Shaun @shaundevan_]

@jacob_posel Spot on - for AI to actually provide value to any business it has to be applied as a service

[Shaun @shaundevan_]

@0xfJuan NYC and there's no debate

[Shaun @shaundevan_]

@Ageless_Ideas haha didn't even notice that till now

[Shaun @shaundevan_]

Sam Altman has access to every model OpenAI makes and he still works the way he did 20 years ago

He still scrolls his inbox hunting for the email that's "least painful" to open. He knows it makes no sense

When David Senra asked him why, he gave the most honest answer I've heard about AI adoption:

In his words: "there's something in my mind that is encoded that doing this kind of stuff is what it means to work"

Look at what "this kind of stuff" is for him: pasting from one messaging app to another, digging through email, working a to-do list

Now look at how work is done in your company. Most of the day is the same stuff. Copying, pasting, chasing updates, moving information from one place to another

And every person doing it has the same AI he does

If the best tools in the world can't break the habit for Altman, handing your team one more tool won't break it either

The habit only breaks when the work itself gets redesigned around AI

The companies that redesign their work around AI are already pulling ahead.

[Shaun @shaundevan_]

@BreezeOg1 haha yep

[Shaun @shaundevan_]

Meta's Muse broke the brain of literally every exec, founder, and operator I've spoken to.

The texts and DMs started the day it launched. And they still haven't stopped.

Every single one of them, without fail, has asked for the exact same thing:

A Jarvis.

AI that reads their email, preps their meetings, books their flights & drafts their replies.

Every time I hear it I cringe.

Their business loses thousands of hours a year to repetitive manual work.

And the thing they want automated first is... their calendar.

I get it. The demos are cool. I'd take the hour a day back too.

So before we build anything, I ask every one of them the same question:

"Will this move the P&L, or do you just think it's cool?"

It's almost always cool.

So we talk most of them out of it and point their AI budget at the work that's eating most of their team's hours.

Hill I will die on: the first AI dollar a $100M company spends should go to the most boring and painful workflow it has.

[Shaun @shaundevan_]

@cobeyluke always!

[Shaun @shaundevan_]

@dibratto can confirm - saw you all over ig somehow

[Shaun @shaundevan_]

@0xfJuan examples? any creators that do it often?