Izkata

Any interface is a possible entrypoint to hack in and gain access to the machine (the government made several attempts over the course of the series, we even meet one of the hackers). Losing that chat interface was part of locking it down before handing it over to the government, so they couldn't weaponize it, and it comes back in the 5th season (not as a retcon, as circumstances within the series).

I tend to place it when Detective Carter finally deals with H.R.

There was definitely a lot building up to that point, but that felt like a major shift where the writers weren't holding back anymore.

Also had multiple failures. There's an extended flashback in one episode where we learn why the "good" AI's creator was so nervous about the whole situation: something like two dozen previous attempts all did various bad things, everything from lying to him about changing its own code to trying to kill him.

> but the beauty for me is in the nuance.

Towards the end there was one scene where it became obvious to me the writers were trusting the audience to understand more than most shows without having to be spoonfed what's going on: they referenced the Great Filter, and gave no explanation at all.

Firefox gives me a bunch of errors like these:

  Content-Security-Policy: The page’s settings blocked an inline script (script-src-elem) from being executed because it violates the following directive: “script-src 'self' 'sha256-0bE0Rzxw6+Rr0Vb1I3B7EN5JZ+NZ9jT+5FiVc5111kE=' 'sha256-5nr3sDO484a+ptnrxPqZmGq9mlgPZRDXnzoSNeOMyJg=' 'sha256-D3lWSzWAu/enrsttp78qOMc3fjgP0Dy4HxDkiZURtVE=' 'sha256-esHsHg/IaqocU5CUjFnT9EU/fiPT0+n92knv+GpsddU=' 'sha256-mKenitUjAI3FLQd3kfGh5iwQOrIytahc0/R5FoT6JKw=' 'wasm-unsafe-eval' https://static.cloudflareinsights.com”

  Content-Security-Policy: The page’s settings blocked a JavaScript eval (script-src) from being executed because it violates the following directive: “script-src 'self' 'sha256-0bE0Rzxw6+Rr0Vb1I3B7EN5JZ+NZ9jT+5FiVc5111kE=' 'sha256-5nr3sDO484a+ptnrxPqZmGq9mlgPZRDXnzoSNeOMyJg=' 'sha256-D3lWSzWAu/enrsttp78qOMc3fjgP0Dy4HxDkiZURtVE=' 'sha256-esHsHg/IaqocU5CUjFnT9EU/fiPT0+n92knv+GpsddU=' 'sha256-mKenitUjAI3FLQd3kfGh5iwQOrIytahc0/R5FoT6JKw=' 'wasm-unsafe-eval' https://static.cloudflareinsights.com” (Missing 'unsafe-eval')
  Source: HTMLElement.prototype.focus = ((realFocu… textjs.org

> This comment section blows my mind! Are we that insulated from the world?

Very much yes, and not on just this topic. I've said it many times over the years in contentious topics (and probably many more I didn't feel like jumping in on), HN is absolutely a filter bubble where most members can't even imagine they're in one.

Not really. One of my coworkers has been using it more recently, and just pushed a merge request last week that mixes a bunch of different styles with little abstraction. And I don't mean styles like linting, it looks like it was done by three different devs who couldn't agree on functional or object oriented style, or whether "spooky action at a distance" was good or not (extra effort to avoid it in some places, relying on it in others). Plenty of things aren't named well either so you have to look at the implementation to figure out what they're doing. And this wasn't even very big, around a thousand lines greenfield.

At this moment on the front page (early morning in the US), this post is #1 and the other one is #2, so it's working like a warning about the other one being misleading before clicking on it.

Sure didn't cause any problems when usage was rapidly rising in the early 2010s. That was not a new policy.

> But it only took 3,000 reputation to cast a close vote, and 15 to flag to close a question (prioritizing it for consideration by the first group).

Quick context for anyone who isn't aware, rep there doesn't work like karma here: your question getting upvoted is apparently now +10 (used to be +5), your answer getting upvoted is +10, either getting downvoted is -2 (and costs the downvoter, they get -1), plus a few others.

I was a SO user from 2011 or so and have not experienced it. I'm with GP and GGP where whenever someone does post an example, it almost always shows them having misunderstood the site or misrepresented the situation, and should have been closed even in the good early days.

For example, one of the uses for closing a question is to give the asker time to improve it based on feedback before people start piling on irrelevant answers (such as including more detail showing why it's not a duplicate of another question). IIRC doing so automatically makes it eligible for reopening, without even having to ask.

Another is that duplicates are meant to point at the best version of the question/answer. Usually this is caught when the new question is posted and it's closed as a duplicate of the older one before answers are posted, so you don't see GGGP's situation, but if it's not caught by someone who knows about the older question the new one can become better and then, by design, the old one is closed and pointed at the better new question.

> SO had been flatlined since 2014 and declining since 2017.

During this peak popularity period, they were actively trying to stop people from flooding the site with low-quality "do my homework"-style questions.

> Why doesn't it move the question to the other site and then make a stub link? More awful, user-hostile experience.

It's to teach the user where it belongs, so they don't make the same mistake again, by making them actually go through the action of going to the other site. Otherwise by auto-migrating it they'll learn to just keep posting on the wrong site and someone else will do the work for them.

> Some later Trek shows (e.g. Voyager) turned captains into geniuses who knew more about everything than all of their underlings.

Not everything. Picard was command track his whole career, I think the same was true for Sisko. But Janeway started as a science officer before switching to command, and Archer was a test pilot. Their experience in those fields would sometimes be useful. It's similar to how Picard's interest in archeology was sometimes useful.

In the US, I also only say summer time/winter time because I can never keep straight which is which.

...so go to sleep while I'm still at the office at 4:30pm?

I see that example as the opposite, people just like to sleep in in general, and later sunrise/sunset gives them more light in the times that matter. They don't cancel out, doing both gives the best result.

You have to know ahead of time whether people value their sleep over sunlight, or sunlight over sleep, to know which version makes the most sense.

Go further back and it just meant happy or cheerful.

I can't help but read it as "la-tinks". And then I think of Tinkerbelle, who has the nickname Tink.

Approval voting has all the upsides and none of the downsides, with the added bonus that it's simple to count and easy to explain to the average person.

> > Maybe new models training on the previous generation’s synthetic output is compounding the effect? LLM Ouroboros.

> This is an interesting failure-mode to me, and I've been interested in how accurate it is since the inception of these models.

The term you want to look for more information on this is "model collapse".

The square bracket version tells the reader it's required, the .get version tells the reader it's optional. Kind of the reason both exist in the language (the .get version has an optional second argument to return instead of None when the key is missing).

> but it's added some fragile and/or unintended consequences

In python, in a co-worker's code, it seems to significantly prefer "some_dict.get('key')" over "some_dict['key']". For the non-python-users, the first one will return None (equivalent to null) when the key is missing, while the second raises a KeyError. It was being used in a whole bunch of places that would cause problems (bad data instead of errors) if it did receive None. Some of these keys were actually optional and it does makes sense for those, but it looks like whichever model he was using just repeated the pattern everywhere instead of just where it was appropriate.

A bit of a tangent, this was written by an Italian singer to sound like English without being English: https://vimeo.com/1071644859

Unfortunately the Youtube upload of that performance seems to have been copyright struck in the past couple of months after being up for 13 years...

You only get that pronunciation by breaking at least two English pronunciation rules: Leading "g" is always pronounced like a "g" (and in this position the "h" prevents it from being pronounced like a "j"), and "ti" is only pronounced like "sh" when followed by a vowel. Can't remember if there was a rule it broke about the "o".

> Pronouncing never heard before words is practically impossible today.

Pronunciation rules do exist. They're not perfect but if you know them you'll get the right pronunciation most of the time. The problem seems to be that they're rarely taught and usually just absorbed through exposure. Most people aren't even aware of them and just do it automatically because of that.

For example, a common gotcha word is "'ghoti is pronounced 'fish'" - except it's not. Ask a native speaker and it's nearly guaranteed they'll say something like "go-tee" or "go-tie", because to get the "fish" pronunciation you have to break multiple English pronunciation rules: The "f" sound for "gh" never occurs at the start of a word, and the "sh" sound for "ti" requires a vowel after it. The "h" also prevents the "g" from being pronounced like a "j" when it starts the word. I can't remember if there was a rule violation for the "o".

Here's someone's attempt at organizing and formalizing them: https://www.zompist.com/spell.html

And from what I've seen the AI tools love creating these, so the overall state is just going to get worse over time as everything drifts.

As someone mentioned above, it probably would have been better if they weren't called functions. Macros, maybe.

Also the blog really doesn't help by describing it like this:

> If you’re familiar with JavaScript, this should feel pretty familiar. We define new functions with @function and specify the returned value with result.

> but it's common for the legal definition of assault to require both willfulness and the knowledge that application of force will arise from your actions

Based on their other comments, they apparently do actually mean it like that: dog owners are intentionally using their dogs to attack runners, then claiming "self-defense" out of fear of runners.

I have never heard of this before, either.

The first time I heard of this effect like two decades ago, it was using the words "takete" and "maluma".

Aside from this bouba question, Kiki is a name. I've also wondered why these two words are used so often when talking about the effect, when they're both much more easily associated with something than takete/maluma.