[srveale]

I don't necessarily disagree with the main argument, but

> did your boss ever have to send you a memo demanding that you use a smartphone

Yes, there were tons of jobs that required you to have a smartphone, and still do. I remember my second job, they'd give out Blackberries - debatably not smartphones, but still - to the managers and require work communication on them. I know this was true for many companies.

This isn't the perfect analogy anyway, since one major reason companies did this was to increase security, while forcing AI onto begrudging workers feels like it could have the opposite effect. The commonality is efficiency, or at least the perception of it by upper management.

One example I can think of where there was worker pushback but it makes total sense is the use of electronic medical records. Doctors/nurses originally didn't want to, and there are certainly a lot of problems with the tech, but I don't think anyone is suggesting now that we should go back to paper.

You can make the argument that an "AI first" mandate will backfire, but the notion that workers will collectively gravitate towards new tech is not true in general.

[Uehreka]

> Yes, there were tons of jobs that required you to have a smartphone, and still do. I remember my second job, they'd give out Blackberries - debatably not smartphones, but still - to the managers and require work communication on them.

Anil is referring specifically to the way that people who were told to use a Blackberry would bring an iPhone to work anyway and demand that IT support it because it was so much better. In the late 2000s Blackberries were a top-down mandate that failed because iPhones were a bottom-up revolution that was too successful to ban.

So look for situations where employees are using their personal AI subscriptions for work and are starting to demand that IT budget for it so they don’t have to pay out of pocket. I’m seeing this right now at my job with GitHub Copilot.

[pxx]

I don't think your example is really a counterexample. Work-provided Blackberries allowed you to be more responsive to work messages while communicating over an ostensibly secure medium.

on the other hand, making sure that people use AI for performance reviews would be akin to measuring the percentage of work days that you used your blackberry. that's not something that anyone sane ever did.

somewhat in the same vein, nobody ever sent a directive saying that all interoffice memoranda must be typed in via blackberry.

[ryandrake]

Yea, the point is, if a product or technology is useful, people will want to use it. They'll bang down your door to be allowed to use it. They'll even surreptitiously use it if you don't allow it. If you have to mandate that they use it, what does that really say about the tool?

A better example is probably source control. It might not have been true in the past, but these days, nobody has to mandate that you use source control. We all know the benefits, and if we're starting a new software business, we're going to use source control by default from day one.

[anildash]

Yeah, I think that's fair, but those bosses that made us get Blackberries were mostly doing that because they wanted to be able to call us and make us work, not because we had to be convinced that smartphones had value, right? We all ended up buying smartphones on our own as well.

[srveale]

You may underestimate how many people do not need to be convinced. Again, I'll refrain from making a value judgment, but the hard numbers show that LLMs have been one of the most quickly adopted technologies in the history of mankind, including the time before anyone was forced to use them.

Not sure if these are the best stats to illustrate the point, but ChatGPT was released November 2022, 2.5 years ago, and they currently claim ~1 billion users [1]

By comparison, iPhone sales were something like 30 million over the same time period, June 2007 through 2009. [2]

In other words, what took ChatGPT several months took smartphones several years.

Of course there are problems with the comparison (iPhones are expensive, but many people bought each version of the iPhone making the raw user count go down, Sam Altman is exaggerating, people use LLMs other than ChatGPT, blah blah blah), so maybe let's not concentrate on this particular analogy. The point is: even a very skeptical view of how many people use LLMs day-to-day has to acknowledge they are relatively popular, for better or worse.

I think we're better served trying to keep the cat from scratching us rather than trying to put it back in the bag. Ham-fisted megalomaniac CEOs forcing a dangerous technology on workers before we all understand the danger is a big problem, that's for sure. To the original point, "AI-first is the new RTO", there's definitely some juice there, but it's not because the general public is anti-AI.

[1] https://www.forbes.com/sites/martineparis/2025/04/12/chatgpt...

[2] https://www.globaldata.com/data-insights/technology--media-a...

[bluefirebrand]

> In other words, what took ChatGPT several months took smartphones several years

You are comparing a cheap subscription service to an expensive piece of hardware that would replace hardware that most people already owned

Of course smartphones were slower to adopt. Everyone had phones already, and they were expensive!

ChatGPT is *free

[srveale]

Do you have any thoughts on the second half of my comment?

[remich]

Well, we all ended up buying smartphones eventually. But the delta between when Blackberries first were adopted in corporate environments and when iPhones/Androids were can't-miss technologies wasn't small.