Nog went up the ranks more quickly because of the war, which provided him with many opportunities. Also, his connections, which allowed him to intern on DS9 and the Defiant, again in the context of war.

Harry was on a ship isolated from Starfleet. If Janeway had promoted everyone who had the merit, she’d have had a ship full of commanders and a handful of captains by the time they got home. But most of those people would have had to keep doing the same jobs, anyway.

But the true reason for the discrepancy noted by OP is just down to the fact that Voyager’s writing team just wasn’t as good as DS9’s (or they were held back from above). The only characters to have a meaningful arc on Voyager were the Doctor, and Seven. The others mostly just had to reset at the end of every episode to preserve the show’s episodic nature.

Harry never got promoted because the writers never figured out how to evolve his function on the show. Now, I assign a small part of the blame to Garrett Wang himself: he rarely gave performances that stood out and made the writers want to write about him more. But most of the blame goes to the writers and show-runners.

For example, if they’d exploited dual crews more throughout the series, with more conflict, etc., Harry might have found a place. Imagine if the Maquis crew had had their own ship for a big part of the series? Harry, eventually frustrated at being passed over for promotion, might have joined them!

If it’s so obvious then due process should be a slam dunk.

Here’s how I understand it: at the time, he thought the penalty was for braking under the safety car, which he thought was bullshit considering how many times Max slowed down just before restarting a race, in the past. He didn’t hear that the penalty was for “erratic driving”.

So, thinking the penalty was bullshit, and that the team were challenging it, he figured he was owed the lead once it would be overturned.

That’s how I figure his reasoning, anyway.

$299

Mic drop.

My own pet peeve is UI components whose associated action is divorced from the components interaction feedback.

For example, a button that seems visibly pressed (even lights up! Maybe there’s even audio or haptic feedback!) but once you release, nothing actually happens because you were supposed to press it or hold it down for slightly longer.

This even happens with physical controls: in some elevators you can press a floor button such that it lights up momentarily, and even beeps, and yet the elevator doesn’t register the command and you have to press again, longer.

“On the other hand”

LoL, good one!

I’ve been “collecting” content for many years now. I learned most of what I needed to know about ripping and transcoding over the years, such that each time I need to deal with a new video format, or a new application, it’s not too hard, because I’m building on everything I’ve already learned.

And each time I was learning new things, it’s not like there was a risk that all my previous content might suddenly become unusable or inaccessible.

Meanwhile, a couple years ago I was finally able to build myself a proper NAS. While I know my way around Linux somewhat, I never kept a Linux-based daily driver because most of the apps I use regularly are on Windows, and I’m not confident about running them stably in Linux, nor am I confident about equivalent native Linux apps. And I’m not confident about setting up and administering my own server. My past experiences have shown me that whenever you need to do anything complex and specific, it involves a lot of work.

So at a coworker’s suggestion, I got a Synology NAS that turned out to be a breeze to setup. And then I figured out how to get Plex server on there (not available in the Synology package manager, but the “manual” process turned out to be simple enough)

And it just WORKS! it’s not perfect, but it’s mostly painless to use. I was happy paying for the lifetime Plex pass at the beginning, because it handles all the routing and discovery that needs to happen to allow me to stream to my phone, or to my parents’ TV when I’m visiting them.

My next NAS might not be by Synology due to their recent announcement about supported hard drives, but I’ll probably be looking for something that “just works” because I can’t be bothered to learn how to be a sysadmin, and risk losing my stuff because I’m making the kinds of mistakes one makes as they’re learning.

Just like, if I owned a car, I wouldn’t be digging under the hood to “tweak the timing” or replacing brake discs. I’d be happy paying someone I trust to do that work, leaving me with a car that “just works”.

Threshold is an attempt at remaking The Fly (mostly the Cronenberg version.) Even though it’s a valiant attempt at body horror, the limitations of network TV really prevent it from reaching its full potential. From the same era, I’d say The X-files did that sort of thing better.

And in fact, the reset button ending is what completely neuters Threshold. The ending to Cronenberg’s The Fly is really powerful because of the final pathos and irreversibility. Instead of leaning on this, Threshold just goes incomprehensibly weird.

I think Star Trek’s most successful body horror episode is probably TNG’s Genesis (though I think Enterprise might have had a couple valiant attempts I should try to remember.). Genesis also suffers a little from its reset button, but since it doesn’t lean on pathos like Threshold and The Fly, the reset isn’t as annoying.

But in general I really feel like Star Trek should now be able to do good episodic TV which still shows real character progression and avoids total reset buttons. Some recent series leaned too far into serialization, but I’m hoping SNW strikes a good balance.

Like tears… in rain.

I don’t like the idea of restricting the model’s corpus further. Rather, I think it would be good if it used a bigger corpus, but added the date of origin for each element as further context.

Separately, I think it could be good to train another LLM to recognize biases in various content, and then use that to add further context for the main LLM when it ingests that content. I’m not sure how to avoid bias in that second LLM, though. Maybe complete lack of bias is an unattainable ideal that you can only approach without ever reaching it.

https://youtu.be/LRq_SAuQDec?si=0bTNdugjosLnu9-r

You can do that today. Just ask Chat-GPT to write you a bash script for something non-obvious, and then debug what it gives you.

The shitty idea was hidden in plain sight: why put chicken on a pizza when you can have pepperoni, bacon, Italian sausage, etc.

The grapes used to make the juice you buy at the store are not the same grapes that are used to make actual wine. The person who wrote that recipe doesn’t know what wine should taste like, and I wouldn’t trust that recipe.

The grape juice you buy for drinking is mainly made from Concord grapes, which won’t make something that tastes in any way like real wine. You need grapes like Cabernet or Pinot Noir (for example). These aren’t just the names of wines, they’re the names of the varieties of grapes used.

Nimoy was a very funny guy! Great comedy instincts! You get another tiny glimpse of it with Spock’s line “I’ve been dead before” in ST:VI. It’s almost a throwaway, and said without emotion, but it’s pretty funny!

Some good examples here. I’d just mention a few other memorable dialogue scenes:

• all the scenes of McCoy with Spock’s Katra in ST:III, including lines like “Yes, Genesis! How can you be deaf with ears like that?” and “Where’s the logic in offering me a ride home, you idiot? If I wanted a ride home, would I be trying to charter a space flight?”
• in Star Trek IV, the scene with Kirk and Spock in Gillian’s pickup, with Nimoy doing his best Gracie Allen impression. Also, the “No dipshit” bit in the restaurant.

I think you have it reversed.

Is that… lilac?

That’s what the Pakled use. The best of them actually use QuickBasic to “make it go faster”.