So I'm hearing that you find optimism, that you feel the government is acting in this direction.

I feel we're starting to think about it.

So there was a tender out in June for about almost a billion dollars to start building essentially a supercomputer.

And the language is a little bit vague.

It leaves things open.

You know, you have to be a Canadian-controlled company.

I don't know if Canadians are going to be excited if, say, hypothetically, Microsoft Canada, the wholly-owned foreign subsidiary of Microsoft, gets our supercomputer deal.

What's at stake if we don't make these changes or keep this momentum building?

Even that has changed how Google operates, right?

Yeah, I mean, the incentives changed and the bargain changed.

So, you know, Google started to really shift when they layered in advertising and suddenly people could, firms could bid, you know, and pay money to be elevated in search because that elevated their credibility and the likelihood that we clicked.

And now, you know, yes, there's that challenge of when you do search something there's just so much clutter and so much noise it's very difficult to find kind of what you want but the point i wanted to write about is that still that kind of approach still presumes that the information we want or theorize Two things are happening.

One, that's increasingly not the case, right? Nothing on the internet is forever.

And two, you know, just like people, you know, again, firms spending advertising dollars to be privileged in search or kind of seem maybe more credible than they otherwise are, people are now gaming the AI that's you know, scrapes the systems that scrape the internet for that information because search is mostly an AI summary.

So people are putting at times like, right, fake information, missing disinformation, inflated information on places like Reddit because Reddit gets scraped for an AI summary.

What else would be a sort of a platform plank that we need to have with this thing?

Even that has changed how Google operates, right?

yeah i mean the the incentives changed and the bargain changed so you know google started to really shift when they layered in advertising and suddenly people could firms could bid you know and and pay money to to be elevated in search because that elevated their their credibility and the likelihood that we clicked and now you know yes there's that challenge of when you do search something There's just so much clutter and so much noise.

It's very difficult to find kind of what you want.

But the point I wanted to write about is that still that kind of approach still presumes that the information we want or theorize is And two things are happening.

One, that's increasingly not the case, right? Nothing on the internet is forever.

And two, you know, just like people, you know, again, firms spending advertising dollars to be privileged in search or kind of seem maybe more credible than they otherwise are, people are now gaming the AI that's you know, scrapes, the systems that scrape the internet for that information, because search is mostly an AI summary.

So people are putting at times like, right, fake information, mis- and disinformation, inflated information on places like Reddit, because Reddit gets scraped for an AI summary.

What else would be a sort of a platform plank that we need to have with this thing?

Vas, do you wanna start with the pros and then we can move to the cons? Or where do you land on it?

Sure.

Sure.

I mean, I'm, I'm similarly mixed.

I would say in the scheme of things, it is another historic moment for Canada's attempts to bring forward more accountability in a digital economy, right? And I kind of focus on whether and what we can actually do here.

You know, the focus on safety is notable.

It's important.

... too basic.

When the government's doing it, sovereign for Canada, not very sovereign for the sovereigns.

I see what you're saying.

In, in pushing adoption, we're maybe not, we're not respecting people's choice and autonomy enough when it comes to AI.

Um, you know, there is this idea that all post-secondary students will have more access to AI.

One wonders if that's just signing a really big check to, uh, foreign, you know, American hyperscalers.

Um, I, I don't see it as completely [laughs] disrespectful.

What I actually see as somewhat, a somewhat disrespectful premise is that trust is low, and in order to build trust, we just need to build people's literacy, right? So saying that low trust is a function of like not understanding AI enough, I think is materially incorrect.

Right? And so people are g- have every right to be scared, but at the same time, there was a reason why people thought it was voodoo.

Am I dreaming, Vas? Yes.

No, you're not dreaming.

I mean, we had all these pre-existing vulnerabilities right before the Trump administration, and now we're more kind of awake and attuned to them because the Trump administration is more willing and open to exploiting them.

Maybe we can think of digital sovereignty in kind of two ways, right? One is just like sovereignty writ large, which is can you govern the economy that you have? That's a pretty simple yes or no.

And a lot of the time with digital sovereignty, we're on the no side because of a range of provisions we signed away in Kuzma last time we engaged on that trade agreement.

But there's also the yes side, right? Can you govern the economy you have? Yes, we have the tools, we have the ability.

Well, then the question becomes, okay, cool, are you? And there are elements, you know, digital elements of the economy where we're slow or we've chosen not to regulate.

I mean, why wouldn't we?