Even reusing existing office space generally results in construction. Every one I've been a part of tears down every bit of drywall and most non-load bearing walls.

Of course, the price associated with that is way lower than building from scratch, and the requirements are less stringent than datacenter space.

I mean sure there are going to be examples, but examples of things largely intact.

If you look at, say, the new Scrubs episodes, it's pretty much a 'by the book' continuation, preserving pretty much the original tone and everything.

Picard tried to 'change it up', but by season 3 settled about as close as they were going to get to 'ok, fine, here's some TNG the later years', though they couldn't bring themselves to completely ignore all the stuff that would obviously be different 30 years later. Hence why I think it's very good decision to use animation to let them continue without some big time gap to overcome which makes things necessarily require some sort of reworking that could break it.

Though a lot of the Ghibli fair is family friendly, which may not assuage his 'cartoons aren't adulty' enough, so they'd probably need to be steered toward, say, Grave of the Fireflies rather than My Neighbor Totoro.

Of course, plenty of examples, e.g. 'The Liberator' is clearly not a fun happy colorful cartoon, and maybe meets his standards by being gritty enough, certainly shows like Invincible and Harley Quinn are very much not for children, though maybe he would find the style too 'bright and fun'. Blade Runner Black Lotus may be a pretty good reference for something tonally consistent with what he'd need to see to believe it can match.

Wash is going to be looking a bit... decomposed by this time as a character.

Frankly, a bunch of people getting the gang back together for space adventures in their 50s sounds like a pretty dubious concept, *especially* in a series that leaned more into physical action on occasion like firefly did.

Maybe all the animation you watched were 80s cartoons or maybe something like Bluey with your kids, but that does not mean that animation is some 'kiddy' thing.

There's some *very* family unfriendly animation out there as a clear counterpoint to your perception that 'animation is for kids'.

What makes you think this? Feel like this is generally some false narrative that some people believe.

Well, I suppose I should have been explicit as to why animated makes sense, their looks age but voices less severely so, so playing 20 years younger is pretty much only an animation thing.

You aren't an LLM, so you're reading and learning is *not* the same as LLM ingest of the material, regardless of what the AI companies want to say. Also, quoting is very specifically laid out in terms of what is 'fair use' or not.

You are free to ignore it if you have some non-negotiable disdain for animation.

It's been over 20 years, they want Alan Tudyk in it, and besides probably an easier pitch to get produced anyway. They can't play the characters at the age they were, Alan Tudyk's character couldn't be part of a 'twenty years later' scenario, and generally speaking it would probably have to be a very different story anyway and I'm not sure whatever 'magic' that combination of folks bring to the table would work in a new formula that would be needed for a '20 years later' approach.

Because realistically, this is the *only* way they can make it work.

They want Wash to be in the show. There's no getting around that it *must* be set before the film. You can't have people credibly play the same character at the same age after 20 years in live action.

Even without the Wash situation, I'd argue that any scenario trying to reconcile 20 years into the future of the characters would be, necessarily, 'f**k'ing with everything. Animated is probably the *least* disruptive approach.

My experience is less about fear of 'rebelliousness' or 'doing the wrong thing', more that they have this employee who manages to make themselves *too* important.

The only 'important' irreplaceable people are 'supposed' to be the executives. So everyone (but them) must be pretty fire-able at a moments notice and keep the business going without fundamentally breaking things. But if someone else can command executive level pay by threatening to leave, well that's not a good possibility. They also do not believe in 'loyalty', since they are so mercenary themselves, they *know* that any important employee will screw them over should that advance their position and wealth.

Pretty much spot on, they have needed those 'nerds' but hate that the 'nerds' have leverage.

They've already leaned hard on incompetent offshoring and waved away resulting business failures, so they get to shift to leaning hard on LLM codegen.

When given the choice of a good solution but they are at the mercy of some skilled employees, or a shitty solution that will likely lose in the market, but the employees are safely fungible, I've seen multiple companies repeatedly go for the 'fungible employee' strategy.

What's amazing is how many of these people think that the circumstances allow them to replace coders with LLMs, but somehow there's room in those circumstances for them. For example, a software sales org that briefed their non-developer salespeople that they could vibecode applications for customers without involving any actual programmers, and still sell it to the customers. Why would they buy a vibe coded solution instead of vibe coding it themselves?

I think this pokes around the issue I find core: Not all "coding" is really the same job.

Assume the "coding" is feeding ego of business leaders to make a "bespoke" implementation of very common stuff where you could have taken an off-the-shelf offering, maybe choosing from a few open source implementations. Then in that scenario, I could believe someone could 'vibe code' their way to a credibly bespoke "in house" solution. I see certain software at threat here, e.g. "Jira" isn't exactly rocket science until some bureaucrats get their bespoke workflows in there. So Jira is a "fine" issue tracking tool that through it's customization can be made into a rough ill fitting bespoke "solution" for some companies workflows, but a vibe coded solution might feel more tailored, since ultimately the workflows aren't really *that* different but they like putting pretty mundane twists on it.

The common pattern of "some GUI frontend to a database" development would be in the same boat.

But some don't deal as much with those. For that person, a lot of the popular frameworks have been similarly annoying, because while they are good at helping with common sorts of applications that are popular to make but not really that different except in style, they don't fit all purposes but some less-than-qualified person is pretty insistent that it should. LLM is just orders of magnitude even more annoying, and less technical leaders are true believers.

Based on my career experience, so many people have been "coding" that both just absolutely hate it and are generally speaking lost and kind of bang away at whatever they are faced with until by some miracle it works. They frequently have no idea why it works even when it does, sometimes only because someone came and reworked it for them. Basically at least back to the dot com rush, "coding" was seen as something to get money even if you don't like it, aren't any good at it. The opportunity is that many of the people controlling the purse also don't know enough to evaluate the people getting paid. At work I get to see a lot of closed source work and for many folks I get a sense of who actually understands their job and who are just kind of mucking with things almost randomly until they work.

It's those folks that have been absolutely the big fan of LLMs. The result is kind of like what they do on their own, but much faster. It's also those folks whose job are ultimately at the biggest risk, unless they can out-politic the other people, which a fair amount of those people are already capable at, since it's how some have stayed employed despite no vaguely sane reason they should have been retained as more useful people get let go. These people are declaring that AI frees the executives from those 'nerds' and now the 'cool' guys can be "10x developers".

Meanwhile, others are frustrated because they know how to precisely get the behavior they want, and LLMs are terrible at getting there. Worse, they are terrible at indicating what they can't do, and pretend to do it anyway in a way that *looks* credible but is wrong. So instead of dealing with '// TODO: Get help implementing this function because I didn't understand the api documentation', you instead get a series of made up function names that sound believable that don't actually exist.

There are some domains where this amount of imprecision is 'good enough', there are some domains that are so very well trodden that LLMs can credibly get to end game without much amending, or amending with vague human input from a person that doesn't quite understand what's going wrong. But there are many domains that LLMs haven't really gotten that good at, including the oh-so-coveted Opus 4.6, and the nature of the limitations seem inherent to the approach. They can get more and more tokens to stay somewhat coherent, but the "whoopsies" persist.

So as presented it *sounds* like the typical respondent thought they spent about as much time as it saved and the numbers were fairly big.

I'm skeptical that a survey of self-reported experience would manifest that way. I wager that some said 9 hours saved with barely anything needed and some said 9 hours of fixing the mistakes and no time saved, and that instead of '16 minutes per week', you just have slightly more people annoyed by it than enthusiastic about it.

My guess is that it wildly varies by the job and situation. People for whom it works badly for are looking around and wondering what the hell is wrong with the people advocating, and the people for whom it works are amazed at the time savings, and given the current hype the former people are forced to do it anyway, even as it is terrible.

For example, a discussion arose where developer mentioned they could make a quick tool to take care of something that was likely to be an issue for users if folks wanted it. An executive declared "no need, I just had Claude make the tool for me in like 2 minutes, and I'll share with everyone". That executive felt *incredibly* empowered, they didn't know how to code and they used the tool and the result of the tool in the AI generated sandbox looked right. Then folks tried the tool and it failed horribly because it didn't have any actual clue about the technology to manage, it managed only to make a tool and sandbox for it to demo in that was consistent with the desired narrative, with no whiff that the results had no relationship to the documentation site included in the executive's prompt. Further, if it *had* worked, it would have broken a number of security mechanisms based on how it *tried* to do things. However, the executive would never feel the consequences, they unleashed their 'awesome tool' and the grunts had to clean up the mess and bringing that back up to executives to ask them to stop doing that stuff is a very risky move employment wise, and they'll not believe you anyway, it's just a conspiracy to keep our jobs after all...

Even when it is being reasonably useful, it's annoyingly prone to mistakes, randomly screwing up even if it's being asked to do the same easy little thing it has done successfully the last 12 times in a row in other contexts. Leaving you feeling a bit gaslight as you see the internet gushing about how unbelievably impossibly good Claude Opus 4.6 is after you just spent a week using your work's budget on that premium model to see if at least *they* are on to something and still finding a pretty annoying experience while Anthropic is crowing that they have all but finished off the job of software developer.

Oh look, working as designed, I can't fix my own thing, but I shouldn't have to care as long as I give Apple money to pay for any work.