Back to all posts
November 16, 2026

Better the Devil You Know

Better the Devil You Know
Listen to this post
0:00 / 7:23

Every "sovereign AI" startup pitches the big model companies as villains. I'm not defending them, but they don't train on API data, you can opt out on the consumer side, and there are audits and regulators. The startup promising "security" usually has a vibe. Pick the devil with the paperwork.

Every week a new AI startup lands in my inbox promising "sovereign," "private," "secure" AI, and the pitch always has a villain in it. OpenAI is reading your data. Google is training on your customers. Anthropic is a black box. Come to us instead, we built our own model, your data never leaves our hands.

The hooded devil leaning out of a laptop inbox with a flyer that paints the red devil as the villain, while the red devil watches unimpressed
Every new pitch in my inbox has a villain. Guess who they cast.

I run an AI company and I'm about to say something that'll annoy half my industry: I'd rather hand my data to the devil I can audit than the one I can't.

A business owner holding a folder of data turning toward the red devil being inspected under a spotlight, away from the hooded devil in the shadows
Two devils. One lets you check his work. I'll take that one.

I'm not defending the big guys

Let me get this out of the way. OpenAI, Anthropic, Google, the rest of them have real problems. There are lawsuits that are going to take years. The terms of service read like they were written to be skimmed. Their incentives are not your incentives. I'm not carrying water for any of them.

The red devil buried at his desk under piles of lawsuit papers and gavels while a business owner stands with arms crossed
The known devil has real problems. Lawsuits, fine print, his own agenda. I'm not his lawyer.

But here's what they also have, and what the "secure AI" startup pitching you almost never has: a published data policy you can actually read. A zero-data-retention option you can turn on. HIPAA-eligible configurations with a signed agreement behind them. Third-party audits with reports you can request. Opt-outs from training that are documented, not promised. Regulators who know their names and have subpoena power. Journalists who would love to catch them lying.

The red devil mobbed by reporters with cameras and inspectors with clipboards while the hooded devil slips away down an empty alley
Reporters with cameras. Inspectors with clipboards. The known devil can't sneeze without someone writing it down.

Those aren't just words on a sales page. OpenAI lists SOC 2 Type 2 and ISO 27001 audits for its API and business products, and signs a HIPAA business associate agreement for eligible health care customers. Anthropic lists SOC 2 Type I and Type II, ISO 27001, and ISO 42001, the international standard for how a company manages AI. Google Cloud is certified under ISO 42001 too. Anthropic will send you its compliance reports through its trust portal if you ask.

The red devil carrying a towering stack of stamped audit reports while the hooded devil holds open a glossy brochure that is empty inside
One of these is annoying to read. The other one has nothing in it to read.

Is the legal language annoying? Yes. Is it a pain to find the right page? Also yes. But it exists, it's specific, and if they violate it there are people whose job is to make that expensive for them. That's what "the devil you know" means. Not that he's good. That you know what he does.

The red devil lounging in an armchair wearing an ankle monitor with cameras on him, while the hooded devil lurks unwatched in a dark corner
Not a good guy. A tracked guy. That's what the devil you know means.

Here's what that looks like in real life. In 2025, in the New York Times copyright lawsuit, a federal court ordered OpenAI to save ChatGPT logs it would normally delete, even chats people had deleted themselves. That's one of the real problems I was talking about. But look at everything else you know. According to OpenAI, business customers on ChatGPT Enterprise, Edu and zero data retention were outside the order, and the broad hold ended that September. All of it played out in public, in court, where anybody could read it. If a startup got the same order, would you ever hear about it?

The red devil stuck in a glass box under spotlights, surrounded by reporters, lawyers and a judge, while the faceless hooded devil stands alone and unwatched in a dark alley
One lives in a fishbowl with a crowd of lawyers. The other lives in the dark.

Two things most people don't know, and should

These are the two facts that quietly demolish most of the fear-based pitches, and almost nobody selling against the big labs brings them up.

The red devil punching two glowing cards numbered 1 and 2 through a scary monster brain poster held up by the hooded devil
Two boring facts from the fine print. They wreck most of the scary pitches.

One: they don't train on API data. There are two different doors into these companies. There's the consumer app, the chat window you type into on your phone. And there's the API, which is what a developer plugs software into. The major providers state plainly in their business and API terms that anything sent through the API is not used to train their models, not by default and not without you specifically asking for it. So when your software calls a model to answer a customer's question, that conversation isn't going into the next version of the model. It's processed, it comes back, and that's the end of it. If you want the retention window gone too, most of them will give you zero-data-retention on request, which means they don't even keep a copy for the usual abuse-monitoring period.

The red devil as a doorman between a chat app door and an API door, with data going in the API door and coming right back out
One door is the chat app. The other is the API. Your business uses the second one.

And it's written down, by name. OpenAI stopped training on API data by default on March 1, 2023, and without zero data retention it keeps API data for up to 30 days to check for abuse. Anthropic says the same about its API, even when you reach Claude through Amazon Bedrock or Google Cloud. And Google says that when you pay for the Gemini API, it doesn't use your prompts or responses to improve its products.

The red devil unrolling a giant scroll of sealed fine print down a staircase while the hooded devil holds up one tiny blank napkin
One hands you a scroll of fine print with signatures. The other hands you a napkin.

Google's terms prove my point from the other side, too. On the free tier, Google says it does use what you send to improve its products, and human reviewers may read it. The terms say, in bold, do not submit sensitive, confidential, or personal information. That's not a scandal. That's the paperwork telling you exactly which door is which. Pay for the door that doesn't train on you.

The red devil holding a rulebook and pointing to a locked paid door instead of a free door that sucks data inside
Free door, paid door. The terms tell you which one trains on you.

That matters because the scary version of this story, "your customer data is becoming part of a giant AI brain," is specifically the thing the terms say doesn't happen on the channel that businesses actually use. And you don't have to take my word for it or theirs. It's written down. Go read it.

The red devil handing a business owner a thick contract while paper envelopes float past a locked glass case and back to the owner
The giant AI brain eating your customer data? That's the exact thing the contract says doesn't happen.

Two: on the consumer side, you can opt out. The chat app is a different deal. There, your conversations may be used to improve the model, and historically that's been on by default. But there's a setting. In every major one of these products you can turn off "improve the model for everyone" or whatever they're calling it this quarter, and once it's off your chats stop feeding training. Business and enterprise plans generally have it off from the start.

The red devil flipping a giant wall switch labeled training down to off
Buried in settings? Sure. But the switch is real, and it works.

Anthropic is a good example of how this plays out. Until 2025 it didn't train on consumer chats at all. Then it changed its consumer terms and asked every Free, Pro and Max user to pick. Say yes, and your chats can be used for training and kept for up to five years. Say no, and they stay on the old 30-day retention. Its business, government and education plans and its API weren't part of the change. Plenty of people were mad, and fair enough. But notice how you found out: a public announcement, a pop-up, a deadline, and a setting you can change any time.

The red devil presenting a giant phone with a pop-up that has yes and no buttons, next to a calendar with a deadline circled
Yes or no, a deadline, and a setting you can change. That's how the known devil changes the rules.

Is "it's on by default and buried in settings" an annoying design choice? Absolutely. I'd rather it were opt-in. But an annoying toggle that exists and works is still infinitely more than a startup telling you they take security seriously and offering you nothing to switch, nothing to read, and nobody to complain to.

The red devil beside a wall of switches with a manual on a chain, and the hooded devil beside a totally blank wall
One gives you switches and a manual. The other gives you a blank wall.

The devil you don't

Now look at the startup that says "we built our own model, so your data is safe." Ask them the questions you'd ask the big guys. Where does the data live? Who can access it? Is it used to train anything? Is there an audit? Who did it? Can I see it? What happens to my data when you get acquired or run out of runway?

A business owner with a checklist of questions so long it rolls across the floor, facing the red devil, who gestures openly, and the faceless hooded devil, who holds up a blank glowing sign
Ask them both where your data lives. One starts talking. The other holds up a blank sign.

I've asked. The answers are usually some version of "we take security very seriously." That's not a policy. That's a vibe. A twelve-person company with a model they trained on a rented cluster has fewer controls, fewer eyes on it, and less to lose from getting it wrong than a company with a hundred lawyers and a stock price. The small company might be honest. I'd bet most of them are. But honest and accountable aren't the same thing, and I can only build a business on the second one.

The red devil leaning on a scuffed steel door with stickers and a window, beside the hooded devil at a glossy door with no window, handle or lock
The ugly door has a window. That's the whole argument.

Most "AI companies" are application layers, and that's fine

Here's the part the sovereign-AI pitch doesn't want you to think about. Most of the companies telling you they're an AI company are calling one of the big providers under the hood. Including a lot of the ones telling you not to trust the big providers. They've got a product on top, a prompt layer, some retrieval, a nice interface, and then an API key to the same model everyone else uses. Scratch the paint and it's OpenAI or Anthropic underneath.

The faceless hooded devil giving a keynote on stage while a power cable runs from under his jacket to a giant data center, where the red devil holds the plug
Look behind the stage and the cord runs straight back to the big guys.

And I want to be clear: there is nothing wrong with that. It's the correct architecture. I don't build my own cloud, I use AWS. Nobody thinks less of a software company for running on AWS. The cabinet shops I work with don't grow their own trees or forge their own saw blades. They buy lumber and hardware and equipment from suppliers who do that at scale, and then they do the part that actually matters to the customer: the design, the fit, the finish, the install.

The red devil delivering crates of lumber and hardware by forklift to a cabinet maker finishing a custom cabinet
He didn't grow the tree. Nobody asks him to.

The dishonest move isn't using somebody else's model. The dishonest move is pretending you didn't, and selling "sovereignty" on top of an API call. If you're an application layer, say so. Then be a great one.

The hooded devil showing off a machine on stage while the curtain slips open to show the red devil cranking the handle that runs it
Using his engine is fine. Hiding it behind a curtain and calling it yours is the problem.

Frontier models are not a startup project

Training a frontier model costs more than most startups will raise in their lifetime, and the compute to run one at scale is rationed. The handful of companies doing it are fighting over chips, power and data centers. That's not a game a seed-stage company plays. It's not even a game most unicorns play.

The red devil and two silver robots fighting over a huge glowing computer chip while the hooded devil watches from a tiny bench with a laptop
The real fight is over chips, power and data centers. Most startups are watching from the cheap seats.

How big is the game? Stanford estimated that the computing power alone to train GPT-4 cost about $78 million, and Google's Gemini Ultra about $191 million. Those were 2023 models, and the bills have only gone up. In 2025, OpenAI, SoftBank and Oracle announced Stargate, a plan to put up to $500 billion into AI data centers in the U.S. Anthropic signed up for as many as one million of Google's AI chips, a deal worth tens of billions of dollars, to bring well over a gigawatt of computing online in 2026. And Amazon, Microsoft, Google and Meta planned about $700 billion in capital spending for 2026, much of it for AI data centers. Your local AI startup is not in that fight.

A cutaway with the hooded devil on a thin app layer at the top and the red devil running a huge server farm deep underground below him
Your app is the top inch. The rest is somebody else's billion dollars, and it should be.

So when a small company says "we built our own model" the honest version is usually "we fine-tuned an open-weight model and we're running it on rented GPUs." Which can be a fine technical choice for some use cases. It is not a security feature. If anything it's the opposite: now your data is on hardware nobody's audited, run by a team that doesn't have a compliance department, with a model whose training data they probably can't fully account for either.

The red devil in a hard hat building a giant data center with cranes, while the faceless hooded devil sits on a folding chair next to one small rented server
One is pouring concrete for gigawatts. The other is renting a box by the hour.

Where startups actually win

The last mile. Every time. The model is a commodity getting cheaper by the quarter. Stanford found that the cost of running a model as good as GPT-3.5 fell more than 280 times between November 2022 and October 2024. What isn't a commodity is knowing exactly how a cabinet shop quotes a job, or what a dental office needs to hear on a missed call, or where a landscaper's leads leak out of the funnel. That knowledge, turned into a product, priced on the outcome it produces, is where the money is for anyone who isn't OpenAI.

The red devil running a conveyor of cheap identical AI brains while a delivery driver carries one to a cabinet shop, a dental office and a landscaper
The brain gets cheaper every quarter. Knowing the customer doesn't.

That's what we do. We own our platform, we control the phone lines and the texts, and the models we call come from providers whose data policies I've actually read, through the API, with retention switched off where it matters. We don't pretend we trained the brain. We built the thing the brain is plugged into, for a specific kind of customer, and we charge for what it produces. A client doesn't care which model answered the phone. They care that the phone got answered and the lead got booked.

The red devil running a giant power plant on the horizon while a small delivery van carries the power up a hill to small businesses
Somebody generates the power. Somebody drives it up the hill. Both get paid.

It's symbiotic, not a war

The pitch that frames big model companies as the enemy is selling you a story, and the story's job is to make you forget to ask for the audit. The real shape of this industry isn't David versus Goliath. It's a power plant and a delivery route. The frontier labs do the thing that costs billions and requires a hundred lawyers. Startups do the thing that requires actually knowing the customer. Startups price on outcomes, build the last-mile logistics, and stand on top of providers who've been forced, kicking and screaming, to write down what they do with your data.

Lawyers dragging the kicking red devil toward a desk to sign a giant contract while a business owner waits with the pen
Not a war. A deal. And the big guys got dragged, kicking and screaming, into writing it down.

That's not a compromise. That's the right split. Pick the devil with the paperwork. Then go build something he can't.

A business owner shaking hands with the red devil holding a stack of signed paperwork, with a busy new shop behind them
Shake hands with the one holding the paperwork. Then build the part he can't.

Want to talk about what you're building?

Get in touch