I'm just wondering whether the logic of the ruling is tied to the specific context (search results) or applied more generally to LLM generated text.

In my view, the ruling could mean that Gemini's output is legally seen as first-person speech by Google regardless of where it is published.

They are also gradually introducing advertising to Gemini and they can upsell and cross sell their paid plans (Gemini, Google One, Youtube, Workspace) to a large pool of users across all their apps, platforms and integrations.

They can do none of that with the white label models they to rent to Apple. That's why Apple will always have to pay for it directly.

This is very different from the search deal Apple has with Google. Under that deal, Apple sends users straight to Google along with all the advertising revenue it brings. Google returns some of that revenue to Apple.

You're right that it is orthogonal to the privacy promises Apple makes to its own users.

The moralistic and righteous undertone in their marketing material is questionable though given that these Apple services might not exist if Google didn't exploit Gemini app user data on Android the way it does.

That's fine with me. Users have a choice here. In fact, it's a big improvement over the search deal with Google where Apple sends its own users directly to Google.

a) Engineers don't know and cannot be expected to know whether what they are being asked to implement complies with all regulations. This is completely normal.

b) Engineers know or can be expected to know based on their expertise that they are being asked to cheat. That's when they are on the hook.

VW was a case of (b). It was clear-cut criminal behaviour on a very technical level. But that's not what typically happens in financial services and many other domains.

But if your point is merely that engineers are not automatically in the clear just because someone higher up told them what to do then I agree with you.

I don't see why developers should be in trouble. Developers don't make unilateral decisions on non-trivial compliance matters. A finding of non-compliance at a financial institution would typically be the result of an investigation, a disagreement with the regulator or a court ruling. It would come years after the organisation as a whole decided to adopt the interpretation in question.

What's desperate is announcing a temporary (allegedly) doubling of revenues days before an IPO that has been criticised for being overpriced at 93 times sales.

These data centers were supposed to serve xAI. Now suddenly they get rented out to others. Why the sudden change of plans?

It's either an emergency accounting gimmick or the effective shutdown or repurposing of xAI.

The whole thing looks rather desperate. I wonder what SpaceX's margins are on these contracts.

I think the tech will work well for some tasks where a formal feedback loop exists (such as coding). In other areas it will take many years to adapt business processes and roles to make the best use of this technology. The total productivity boost could be around 1% p.a similar to the industrial revolution of the 19th century.

Stock prices could be at risk not from lack of demand but because the data center buildout is bound to slow dramatically as we come up against some serious bottlenecks like energy, grid, fab capacity, permissions, etc. Not much will have to be written off, but the delays could cause big problems for debt funded projects and companies.

This slowdown will allow the economy and the workforce to evolve away from execution and towards planning, strategy, research and development, idea generation, experimentation, oversight as well as manually handling a million exceptions and gaps left by current AI models.

I don't think there has ever been a tech boom without a tech bust. But that's not the same thing as the tech not working or causing economic collapse. Maybe this time is different. Who knows.

In other words, if you include everything that's required to create useful software then 100x turns out to be a fantasy.

While true on some level, I don't think this is a very useful statement. The importance of mathematical foundations lies in the extent to which they constrain the features of a programming language.

That extent is not the same for all languages. Many programming languages do not appear to be constrained by anything other than some pragmatic hunch of their designers plus the theoretical limits of computability.

SQL is a mess. The author acknowledged that. But the relational model and relational algebra are more serious attempts at creating a small but expressive theory than many of our mainstream programming languages.

If the solution you find is longer and not much faster than the procedural alternative, you throw it away and fall back on procedural code.

Stored procedures are not advanced SQL. Most of them are not SQL at all. There are a few legitimate reasons for using SPs such as reducing roundtrips to the database and writing little pure functions for use in SQL statements.

But many uses of SPs are just laziness or a symptom of organisational dysfunction.

It doesn't have to be a new language. I'm sure some existing language can be used to create a DSL that serves this purpose.

It can obviously never be complete. Some parts of the spec will always have to be natural language if we want to make the best use of LLMs.

Go is too verbose and the type system isn't expressive enough. Rust code is littered with little memory management details and it requires tons of third party libraries.

I think coding agents will eventually be able to get the low level details right on their own. Reviewers should be able to focus on architecture, design and logic mistakes.

I also think we need a high level formal specification language to tell agents what we expect them to do.

To be fair, there will always be something that fails. So the more important question is probably the frequency and severity of those failures.