Every major CEO AI-first memo and mandate, collected. Shopify, Box, Duolingo, Fiverr, Meta, Klarna, Alibaba, Notion, and more.
Reading these memos side by side, a pattern emerges. They all use the same format (CEO writes to all employees about AI transformation) but contain three fundamentally different philosophies about what "AI-first" means.
AI as gate. Shopify, Duolingo, and Fiverr share a version of this: before you get resources (headcount, budget, tools), demonstrate that AI can't do the work. The human must justify their role relative to AI. This is the most provocative framing, and it generates the most press.
AI as ladder. Box takes this approach. AI doesn't replace people, it makes them more productive, and the productivity gains get reinvested. Teams that adopt AI get more resources, not fewer. Levie specifically contrasted his approach with Duolingo's "prove AI can't do it" framing.
AI as fait accompli. Klarna didn't write a forward-looking memo. It reported what had already happened. This framing is the riskiest. When Klarna's numbers turned out to tell a simpler story than reality allowed, the reversal was public and awkward.
The memo is the strategy
The CEO AI memo isn't a communication about strategy. It is strategy. Writing it and publishing it under your own name does several things at once that no Slack message or quiet policy change could.
It creates accountability. Every manager now has cover to enforce it and no room to ignore it. It sets the narrative externally. Investors, analysts, and potential hires all read these memos. Lutke didn't just tell Shopify employees to use AI. He told the market that Shopify is an AI company. And it creates peer pressure. Not having a memo started to look like not having a strategy.
Nobody defines it
The most revealing thing about these memos is what's absent: a definition. None of them define what "AI-first" actually means. Lutke says it's a "baseline expectation" but doesn't specify what that looks like for a designer vs. a supply chain manager. Von Ahn says Duolingo is "AI-first" but has to walk it back months later because people filled in their own definition.
That's not a flaw. The memo works because it's directional rather than definitional. It says "this is where we're going" without getting bogged down in what it looks like when you arrive.
The investors are further along on definitions than the operators. Bessemer draws a line between companies advancing AI as a science vs. using it as a distribution machine. Intel Capital uses a four-tier spectrum from AI-Enhanced to AI-Native. Sequoia looks at revenue per employee. Hit $1M+ and you're probably the real thing.
But the CEOs don't need taxonomies. They need momentum. And the memo, public, permanent, attached to their name, creates it.
The Klarna lesson
Every CEO writing one of these memos should read Klarna's story. Siemiatkowski went further than anyone: a hiring freeze, a 40% headcount reduction, public celebrations of AI replacing hundreds of agents. Then he reversed course, admitted quality had suffered, and started hiring humans again.
He's been more open about this than most executives would be. He's publicly wrestled with the implications, writing that AI could make his own role unnecessary, which he finds "gloomy." But the arc (from loud AI-first announcements to quietly hiring humans again) exposes the gap between the CEO AI memo as a communications device and what's actually happening inside the company.
Read the original article
Comments
By MontyCarloHall 2026-02-1116:0513 reply Whatever happened to "show, don't tell"? Other productivity boosters certainly didn't need such memos; they were naturally adopted because the benefits were unambiguous. There were no "IDE-first company memos" or "software framework-first company memos"; devs organically picked these up because the productivity gains were immediately self-evident.
By munificent 2026-02-1116:297 reply Think about the Industrial Age transition from individual craftspeople working on small shops using hand tools to make things into working in factories on large-scale assembly lines. The latter is wildly more productive than the former. If you owned a business that employed a bunch of cobblers, then moving them all out of their little shops into one big factory where they can produce 100x as many shoes means you just got yourself 100x richer.
But for an individual cobbler, you basically got fired at one job and hired at another. This may come as a surprise to those who view work as simply an abstract concept that produces value units, but people actually have preferences about how they spend their time. If you're a cobbler, you might enjoy your little workshop, slicing off the edge of leather around the heel, hammering in the pegs, sitting at your workbench.
The nature of the work and your enjoyment of it is a fundamental part of the compensation package of a job.
You might not want to quit that job and get a different job running a shoe assembly line in a factory. Now, if the boss said "hey, since you're all going to be so much more productive working in the factory, we'll give you all 10x raises" then perhaps you might be more excited about putting down your hammer. But the boss isn't saying that. He's saying "all of the cobblers at the other companies are doing this to, so where are you gonna go?".
Of course AI is a top-down mandate. For people who enjoy reading and writing code themselves and find spending their day corralling AI agents to be a less enjoyable job, then the CEO has basically given them a giant benefits cut with zero compensation in return.
Yup. It’s what I’ve come to realize. My job is probably safe, as long as I will be willing to adapt. I have still not even tried AI once, and don’t care for it, but I know at one point I probably will have to.
I don’t actually think it’ll be a productivity boost the way I work. Code has never been the difficult part, but I’ll definitely have to show I have included AI in my workflow to be left alone.
Oh well…
Why have you never even _tried_ it? It’s very easy to try and surely you are somewhat curious.
By agentultra 2026-02-1119:122 reply I've never had to try lots of things in order to know that I won't like them.
That, and also I specifically loathe the way AI has been created.
In my heart I’m a code “artist,” and like all artists that have been more directly attacked, I also feel personally attacked by all of my stolen “art” that is now monetized by big corporations that do not give a single f*ck about beauty in the code, or whatever.
This may be a strange position, idk. Anyways, that’s the reason why.
By agentultra 2026-02-1214:57 You’re not wrong about it being stolen and monetized without consent. I’m picking up what you’re putting down.
Yeah but the cost of trying it is like 3 minutes of your time. Have you no curiosity?
By agentultra 2026-02-1214:50 Enough curiosity to have read and understood the papers, formalize some of the more grand theorems (they’re not that impressive once restated this way), and listen to the experiences of developers who’ve adopted it.
That’s more than 3 minutes. I’m not being glib when I say I’ve given it serious thought.
By ndespres 2026-02-1122:33 Not for this, no.
By wiseowise 2026-02-1123:22 Agree with every word.
> Of course AI is a top-down mandate. For people who enjoy reading and writing code themselves and find spending their day corralling AI agents to be a less enjoyable job, then the CEO has basically given them a giant benefits cut with zero compensation in return.
This should be a disclaimer every time someone at work forces you to use AI.
Interesting how when WFH became "the norm" during COVID, there were thousands of apologists arguing that employees suddenly received perks for nothing. Where are all of you now? Why aren't you arguing against employer doing a fucking rug pull?
The industrial revolution was extremely hard on individual craftspeople. Jobs became lower paying and lower skilled. People were forced to move into cities. Conditions didn't improve for decades. If AI is anything comparable it's not going to get better in 5-10 years. It will be decades before the new 'jobs' come into place.
Seriously, it took nearly ~150 years before the people actually benefited from the industrial revolution. Saying that we need to condemn two lifetimes worth of suffering to benefit literally a few thousand people out of billions is absolutely ludicrous.
By wiseowise 2026-02-1123:24 But think about corporate aristocracy and their children!
This is basically not true. It's hard to debate this when we don't start from a position of truth.
It pretty much is, unless you think it's totally cool to work in highly dangerous jobs that paid poorly while being treated like chattel slaves. There is a reason why the 1800s had the most violent labor actions in the US, it wasn't because they were treated "well."
Completely disingenuous, learn your labor history.
By munksbeer 2026-02-1223:16 People didn't feel the benefits for 150 years? Just absolute nonsense.
Unfortunately, I would expect the boss to say, "hey, since you're all going to be so much more productive working in the factory, we'll give you all 10x the shoes to repair".
By N_Lens 2026-02-1123:41 The reward for working hard is more hard work.
By MontyCarloHall 2026-02-1119:051 reply >Think about the Industrial Age transition from individual craftspeople working on small shops using hand tools to make things into working in factories on large-scale assembly lines.
I wouldn't analogize the adoption of AI tools to a transition from individual craftspeople to an assembly line, which is a top-down total reorganization of the company (akin to the transition of a factory from steam power to electricity, as a sibling commenter noted [0]). As it currently exists, AI adoption is a bottom-up decision at the individual level, not a total corporate reorganization. Continuing your analogy, it's more akin to letting craftspeople bring whatever tools they want to work, whether those be hand tools or power tools. If the power tools are any good, most will naturally opt for them because they make the job easier.
>The nature of the work and your enjoyment of it is a fundamental part of the compensation package of a job.
That's certainly a part of it, but I also think workers enjoy and strive to be productive. Why else would they naturally adopt things like compilers, IDEs, and frameworks? Many workers enjoyed the respective intellectual puzzles of hand-optimizing assembly, or memorizing esoteric key combinations in their tricked-out text editors, or implementing everything from scratch, yet nonetheless jumped at the opportunity to adopt modern tooling because it increased how much they could accomplish.
By munificent 2026-02-1119:281 reply > As it currently exists, AI adoption is a bottom-up decision at the individual level, not a total corporate reorganization.
I'm sorry, but did you forget what page this comment thread is attached to? It's literally about corporate communication from CEOs reorganizing their companies around AI and mandating that employees use it.
> That's certainly a part of it, but I also think workers enjoy and strive to be productive.
Agreed! Feeling productive and getting stuff done is also one of the joys of work and part of the compensation package. You're right that to the degree that AI lets you get more done, it can make the job more rewarding.
For some people, that's a clear net win. They feel good about being more productive, and they maybe never particularly enjoyed the programming part anyway and are happy to delegate that to AI.
For other people, it's not a net win. The job is being replaced with a different job that they enjoy less. Maybe they're getting more done, but they've having so little fun doing it that it's a worse job.
By MontyCarloHall 2026-02-1119:36 >I'm sorry, but did you forget what page this comment thread is attached to? It's literally about corporate communication from CEOs reorganizing their companies around AI and mandating that employees use it.
That’s exactly my point. The fact that management is trying to top-down force adoption of something that operates at the individual level and whose adoption is thus inherently a bottom-up decision says it all. Individual workers naturally pick up tools that make them more productive and don’t need to be forced to use them from the top-down. We never saw CEOs issue memos “reorganizing” the company around IDEs or software frameworks and mandate that the employees use them, because employees naturally saw their productivity gains and adopted them organically. It seems the same is not true for AI.
By abeppu 2026-02-1120:25 > Now, if the boss said "hey, since you're all going to be so much more productive working in the factory, we'll give you all 10x raises" then perhaps you might be more excited about putting down your hammer.
... is now the moment to form worker cooperatives? The companies don't really have privileged access to these tools, and unlike many other things that drive increased productivity, there's not a huge up-front capital investment for the adopter. Why shouldn't ICs capture the value of their increased output?
By withinboredom 2026-02-1119:31 Blacksmiths pretty much existed until the ‘50s and ‘60s for most of the world, making bespoke tools and things. Then they just vanished, for the most part.
We are probably on a similar trajectory.
By paodealho 2026-02-1116:40 Goes to show how infested with disconnected management this industry is.
All the tools that improved productivity for software devs (Docker, K8S/ECS/autoscaling, Telemetry providers) took very long for management to realize they bring value, and in some places with a lot of resistance. Some places where I worked, asking for an IntelliJ license would make your manager look at you like you were asking "hey can I bang your wife?".
Remember when companies all forced us to buy smartphones? Or switch to search engines instead of books? Or when Amazon announced it was "react native first"?
By Mordisquitos 2026-02-1120:27 I agree with the sentiment you're expressing but, to be fair, companies forcing us all to use smartphones (as consumers or as citizens) is, unfortunately, happening implicitly.
People will voluntarily adopt modest productivity boosters that don't threaten their job security. They will rebel against extraordinary productivity boosters that may make some of their skills obsolete or threaten their career.
By nitwit005 2026-02-1118:29 You have to remember that our trade is automating things. We're all enthusiasts about automating things, and there's very clearly a lot of enthusiasm about using AI for that purpose.
If anything, the problem is that management wants to automate poorly. The employees are asked to "figure it out", and if they give feedback that it's probably not the best option, that feedback is rejected.
By MontyCarloHall 2026-02-1116:332 reply That’s simply not true. Developers hand-writing assembly readily adopted compilers, accountants readily adopted spreadsheets, and farmers readily adopted tractors and powered mills.
That's false. Those things were in fact resisted in some cases. For instance, look up the swing riots of the 1830s.
There might be a temporary resistance from violence but eventually competition will take over. The issue in this case is that we're not looking at voluntary adoption due to a competitive advantage - we're seeing adoption by fiat.
AI is a broad category of tools, some of which are highly useful to some people - but mandating wide adoption is going to waste a lot of people's time on inefficient tools.
The competitive advantage belongs to companies, not engineers. That's exactly the conflict. What you're predicting -- voluntary adoption due to advantages -- is precisely what is happening, but it's happening at the company level. It's why companies are mandating it and some engineers are resisting it. Just like in the riots I mentioned -- introduction of agricultural machinery was a unilateral decision made by landowners and tenant farmers, often directly against the wishes of the laborers.
A well run company would provide an incentive to their employees for increasing their productivity. Why would employees enthusiastically respond to a mandate that will provide them with no benefit?
Companies are just groups of employees - and if the companies are failing to provide a clear rationale to increase productivity those companies will fail.
> A well run company would provide an incentive to their employees for increasing their productivity.
Sure, that's tautological.
But in reality, these don't exist.
By munk-a 2026-02-1214:58 I don't disagree that they're quite rare now - but they used to exist and we could make them exist again.
I'm sorry to say this, but the company does not need employees to respond enthusiastically. They'll just replace the people who resist for too long. Employees who resist indefinitely have absolutely zero leverage unless they're working on a small subset of services or technologies where AI coding agents will never be useful (which rules out the vast majority of employed software developers).
By munk-a 2026-02-1117:56 Oh, they can certainly do that (in part evidenced by companies doing that). It's a large cost to the company, you'll get attrition and lose a lot of employee good-will, and it'll only pay off if you're right. Going with an optional system by making such tools available and incentivizing their use will dodge both of those issues and let you pivot if the technology isn't as beneficial as you thought.
By warkdarrior 2026-02-1116:461 reply Your examples are productivity boosters that don't threaten job security. A human has to provide inputs to the compiler, the spreadsheet, and the tractor.
By mjr00 2026-02-1119:00 The tractor, or more generally farm automation, was maybe the biggest single destruction of jobs in human history. In 1800 about 65% of people worked in agriculture, now it's about 1%. Even if AI eliminated every single computer programmers' job it would be a drop in the bucket compared to how many jobs farm automation destroyed.
By palijer 2026-02-1118:44 There was an Apple memo like this though that said they were word processing first.
https://writingball.blogspot.com/2020/02/the-infamous-apple-...
By SatvikBeri 2026-02-1120:341 reply Bezos's API memo is the biggest example I can think of. It was not individually productive for teams but arguably it was very productive for Amazon/AWS as a whole.
By MontyCarloHall 2026-02-1120:38 That's a top-down organizational change mandating certain capabilities for all Amazon software. Unlike these AI mandates, it's not dictating the exact tools developers use to write that software, but rather what the software itself should do. For reference, here's the API mandate [0]:
This is very different from saying "any developer who doesn't use an IDE and a debugger will be fired," which is analogous to what the AI mandates are prescribing.1. All teams will henceforth expose their data and functionality through service interfaces. 2. Teams must communicate with each other through these interfaces. 3. There will be no other form of interprocess communication allowed: no direct linking, no direct reads of another team’s data store, no shared-memory model, no back-doors whatsoever. The only communication allowed is via service interface calls over the network. 4. It doesn’t matter what technology they use. HTTP, Corba, Pubsub, custom protocols — doesn’t matter. 5. All service interfaces, without exception, must be designed from the ground up to be externalizable. That is to say, the team must plan and design to be able to expose the interface to developers in the outside world. No exceptions. 6. Anyone who doesn’t do this will be fired.[0] https://nordicapis.com/the-bezos-api-mandate-amazons-manifes...
By Noumenon72 2026-02-1116:243 reply On the other hand, there were surely memos like "our facility will be using electric power now. Steam is out". Sometimes execs do set a company's direction.
By MontyCarloHall 2026-02-1116:311 reply AI adoption is a bottom-up decision at the level of the individual worker. Converting an entire factory is a top-down decision. No single worker can individually decide to start using electricity instead of steam power, but individuals can choose whether/how to use AI or any other individual-level tool.
> AI adoption is a bottom-up decision at the level of the individual worker.
Ah, but that is the question, isn't it. Is it bottom up or top down?
By sensanaty 2026-02-120:15 I think they're saying it should logically be a bottoms-uo initiative, since it's the rank and file that use it day in, day out, but instead it's top-down because we're in a massive bubble
By nitwit005 2026-02-1118:14 That's a choice individual employees couldn't make. Or, at least, one management wouldn't let them make. It'd require a huge amount of spending.
By abtinf 2026-02-1118:46 That transition took 40-50 years. Electrical power in manufacturing was infeasible for lot of reasons for a longtime.
Any company issuing such an edict early on would have bankrupted themselves. And by the time it became practical, no such edict was needed.
By SkyPuncher 2026-02-1119:14 That doesn't work in an environment where there are compliance and regulatory controls.
In most companies, you can't just pick up random new tools (especially ones that send data to third parties). The telling part is giving internal safety to use these tools.
By charcircuit 2026-02-1118:441 reply >Other productivity boosters certainly didn't need such memos; they were naturally adopted because the benefits were unambiguous.
This is simply not true. As a counter example consider debuggers. They are a big productivity boost, but it requires the user to change their development practice and learn a new tool. This makes adoption very hard. AI has a similar issue of being a new tool with a learning curve.
By RandallBrown 2026-02-1118:562 reply Did companies actually send out memos saying "We're going to be a company that uses debuggers!"
I would have just thought that people using them would quickly outpace the people that weren't and the people falling behind would adapt or die.
By charcircuit 2026-02-1119:14 >Did companies actually send out memos saying "We're going to be a company that uses debuggers!"
I could believe it. Especially if there are big licensing costs for the debuggers.
>the people falling behind would adapt or die.
It is better to educate people, make them more efficient, and avoid having them die. Having employees die is expensive for the company.
By agentultra 2026-02-1119:14 Do they have to die though? I know some folks that use them and others who don't. They both seem to get along fine.
By rchaud 2026-02-1121:13 Stock valuation has moved further away from Ben Graham-era emphasis on analysis of cash flow.
Productive output is a lagging indicator. Using AI tools is theoretically leading???
By nitwit005 2026-02-1118:17 I had you a power tool, and your productivity goes up immediately. Your IDE highlights problems, same story. Everyone can observe that this has happened.
By badc0ffee 2026-02-1117:34 20 years ago or so, we had an exec ask us about our unit tests.
By apimade 2026-02-1121:57 Contract-first. API-first. Domain-driven. Platform driven. Microservice driven.
Tech loves making something a top priority (and forgetting about it several years later); AI is the first one that is applicable to the masses.
.. Well maybe not User-first. But that was even less clear than AI-first.
By whiplash451 2026-02-1115:591 reply It's so sad to see some of these companies completely fail their AI-first communication [1], when they would just get so much from "We think AI can transform the way we work. We're giving you access to all these tools, please tell us what works and what doesn't". And that would be it.
[1] there was a remote universe where I could see myself working for Shopify, now that company is sitting somewhere between Wipro and Accenture in my ranking.
Unfortunately at this scale, when you are this soft on the message, everyone ignores it and keeps doing what they were doing before. Carrot and stick are both required for performance management at this scale. You can argue whether the bet is worth it or not, but to even take the bet, you need a lot more than some resources and a "please".
If performance was the true goal then we'd just naturally see slow adopters unperform and phase out of that company. If you make good tooling available and it is significantly impactful the results will be extremely obvious - and, just speaking from a point of view of psychology, if the person next to you is able to do their job in half the time because they experimented with new tooling _and sees some personal benefit from it_ then you'll be curious and experiment too!
It might be that these companies don't care about actual performance or it might be that these companies are too cheap/poorly run to reward/incentivize actual performance gains but either way... the fault is on leadership.
By NewsaHackO 2026-02-1121:51 Two things: 1) Employees are not that easy to replace. These employees have already been onboarded, screened, and proven, at least to this point, to be the type of people the company wants. If an employee starts lagging behind solely because they are stubborn about adopting AI, yes, the company can fire them, but then it has to go through the entire hiring process again and risk bringing in someone new, when it could have simply improved performance by helping the existing employee use AI. 2) Companies themselves have performance metrics that are compared to those of other companies. If an employee is not using AI and has reduced output, then the company’s overall output and its profits are affected. No investor cares if the reason is that other companies have a higher rate of AI adoption; investors care that the company was not able to get its employees to use AI effectively to increase profits.
By tikhonj 2026-02-1116:14 It's not inevitable, it's just poor leadership. I've seen changes at large organizations take without being crudely pushed top-down and you'd better believe I've seen top-down initiatives fail, so "performance management" is neither necessary nor sufficient.
The executives pushing AI use everywhere aren’t basing it on actual performance (which is an orthogonal concept). It’s just the latest shiny bauble.
Performance management isn't rating how people are doing. It's transforming the resources of the company into something that you want it to do. If they want to transform the current state of the company into something that has AI use as a core capability, that is performance management.
There are good books on this: e.g. https://www.amazon.ca/Next-Generation-Performance-Management...
By goatlover 2026-02-1117:42 If everyone is ignoring it, it can't be that great. If it's that great, people will adopt it organically based on how it's useful for them.
Fiverr CEO goes on a pretty dark rant about how people who don’t upskill and convert to AI workflows will have to change professions, etc.
Then concludes his email with:
> I have asked Shelly to free up time on my calendar next week so people can have conversations with me about our future.
I assume Shelly is an AI, and not human headcount the CEO is wasting on menial admin tasks??
By azinman2 2026-02-126:05 The irony is if you take this to the limit, ChatGPT replaces fiverr
By yencabulator 2026-02-1716:30 I assume Shelly Paran is Fiverr's Chief of Staff, and a human. You already knew that rules don't apply to CEOs.
