Safari on iOS has always had some pretty strict limits on what extensions can do. For example, content blockers don’t get to run code on the pages you browse, it’s more like they give the browser a list of what type of thing to block when you install and configure it, then when you’re browsing, the extension isn’t even doing anything, it’s just the browser using the list. Obviously that’s more limiting, there might be ads that are best dealt with by running a bit of code, so it makes sense that they’d consider it “lite”. (The benefit of those limits is that ad blocking extensions can’t run amok and kill your phone’s battery since the browser’s handling it by itself.)

There isn’t a simple evolutionary definition of “fish”, not the same way there is for, say, mammals. If you found the common ancestor of everything we call a mammal and said “everything descended from this one is also a mammal”, you’d be correct. If you did that for everything we call fish, every animal in the world would be a fish. Also, we decided which animals were fish mostly on vibes, so without a clear definition you can pedantically argue that everything is a fish including mammals.

I would guess it’s web requests made during a chat session, e.g. the user asks about kayaks and the AI searches and fetches some pages to put into its context before answering. That’s not really scraping, it’s data being used in the moment in response to a user request, closer to what a “user agent” has always meant in the web world. A crawler would be crawling the site, systematically trying to follow every link and collect what’s there with little to no human involvement.

If nothing else, it seems reasonable to assume that a computer could run a program that does all the important things 1 neuron does, so from there it’s “just” a matter of scaling to human brain quantities of neurons and recreating enough of the training that evolution and childhood provide to humans. But that’s the insanely inefficient way to do it, like trying to invent a clockwork ox to pull your plow instead of a tractor. It’d be pretty surprising if we couldn’t find the tractor version of artificial intelligence before getting an actual digital brain patterned directly off of biology to do it.

It’s making mistakes and failing to think abstractly at levels previously only achieved by humans, so it’s only rational to expect it to take over the world in 5 years

That’s a decently rational response you’ve described, though. If you were really at immediate risk, you’d probably know it, especially with an alarm going off to get you looking for signs of danger. And it’s usually better to have a lazy, skeptical evacuation than a panicked stampede. Schools do fire drills to check the alarms, sure, but it’s also important to make them a routine thing that all the kids know how to handle.

I’m sure shuffled pieces have been a thing, there’s also Really Bad Chess which gives you different piece quantities entirely.

The tokenization is a low-level implementation detail, it shouldn’t affect an LLM’s ability to do basic reasoning. We don’t do arithmetic by counting how many neurons we can feel firing in our brain, we have higher level concepts of numbers, and LLMs are supposed to have something similar. Plus, in the “”“thinking”“” models, you’ll see them break up words into individual letters or even write them out in a numbered list, which should break the tokens up into individual letters as well.

Ah, yes, I see that you know your judo well. Good one.

I am deeply troubled by the way that AIs slip right past peoples’ defenses and convince them of things that are absolutely not true. Not just things like the AI psychosis that some people have been driven into, not just the hallucinations or overly fawning over terrible ideas, it goes so much further than our monkey brains can understand. These things do not think, they do not have feelings, they don’t have motivations, they don’t have morals or values, no sense of right or wrong, they are, quite literally, word prediction machines that are selecting their responses semi-randomly. And yet, even people who know this to be the case cannot stop themselves from anthropomorphizing the AI. All of our human instincts scream “this is a person!” when we interact with them, and with that comes an assumption of values, morals, thoughts, and feelings, none of which are there. There is a fundamental mismatch between the human mental model and the reality of the software, and that is harmful, even dangerous. We will give benefit of the doubt to them, we will believe their explanations about “why” they “chose” to say or do what they did, and we will repeatedly make the same fundamental mistakes because we want to believe a total lie.

And that’s not even getting in to the harm when they are working properly, encouraging us to outsource our thinking and creativity to a service that charges monthly. I’m seriously worried that kids in school now are going to have their learning stunted terribly, it’s just so easy to have any and all homework done in a matter of minutes without learning a single thing.

Older Unix systems used to only do the first 8 bytes for passwords. Sometimes for my own amusement when logging into one of the Sun machines at school, I’d type in enough of my password to count and then just mash the keyboard.

There is basically zero actual evidence here. The argument basically goes “this could’ve happened, then this other thing could’ve happened, then a third thing could’ve happened, someone said something vaguely ominous in a group chat, and then something we all know is impossible happened: Donald Trump was elected President despite being obviously bad. There’s only one conclusion: the election was stolen and now we just need to track it down.” Read the article again and try to pick out the things that are shown to have actually happened and weren’t just suspicions or possibilities.

It doesn’t hold up for the same reasons the 2020 doubts didn’t hold up. Did they do this in every state? Because the results were pretty uniform across the country, it was a big swing right. It’d be the biggest and most successful conspiracy in history, getting away with rigging a wide variety of completely separate voting systems, many of which are heavily or entirely paper-based, many of which are run by blue states or weren’t even competitive, with no leaks and no discrepancies in any of the public records.

Or, maybe, just maybe, Biden was incredibly unpopular and Kamala didn’t run a good campaign, while Donald “I’ll fix everything and everyone will be rich” Trump promised to take action and not just continue the same policies for another four years, so people gave him another shot. “Oh, but he had felonies! Surely the electorate would never!” Yeah, they would. We elect terrible people all the time. He won. This isn’t productive.

I think that joke’s been around for a while, but there is the Terry Pratchett line about how if you had a button with a sign next to it saying “pressing this button will end the world, do not touch,” the ink wouldn’t even have time to dry.

If you ask an LLM to help you with a legal brief, it’ll come up with a bunch of stuff for you, and some of it might even be right. But it’ll very likely do things like make up a case that doesn’t exist, or misrepresent a real case, and as has happened multiple times now, if you submit that work to a judge without a real lawyer checking it first, you’re going to have a bad time.

There’s a reason LLMs make stuff up like that, and it’s because they have been very, very narrowly trained when compared to a human. The training process is almost entirely getting good at predicting what words follow what other words, but humans get that and so much more. Babies aren’t just associating the sounds they hear, they’re also associating the things they see, the things they feel, and the signals their body is sending them. Babies are highly motivated to learn and predict the behavior of the humans around them, and as they get older and more advanced, they get rewarded for creating accurate models of the mental state of others, mastering abstract concepts, and doing things like make art or sing songs. Their brains are many times bigger than even the biggest LLM, their initial state has been primed for success by millions of years of evolution, and the training set is every moment of human life.

LLMs aren’t nearly at that level. That’s not to say what they do isn’t impressive, because it really is. They can also synthesize unrelated concepts together in a stunningly human way, even things that they’ve never been trained on specifically. They’ve picked up a lot of surprising nuance just from the text they’ve been fed, and it’s convincing enough to think that something magical is going on. But ultimately, they’ve been optimized to predict words, and that’s what they’re good at, and although they’ve clearly developed some impressive skills to accomplish that task, it’s not even close to human level. They spit out a bunch of nonsense when what they should be saying is “I have no idea how to write a legal document, you need a lawyer for that”, but that would require them to have a sense of their own capabilities, a sense of what they know and why they know it and where it all came from, knowledge of the consequences of their actions and a desire to avoid causing harm, and they don’t have that. And how could they? Their training didn’t include any of that, it was mostly about words.

One of the reasons LLMs seem so impressive is that human words are a reflection of the rich inner life of the person you’re talking to. You say something to a person, and your ideas are broken down and manipulated in an abstract manner in their head, then turned back into words forming a response which they say back to you. LLMs are piggybacking off of that a bit, by getting good at mimicking language they are able to hide that their heads are relatively empty. Spitting out a statistically likely answer to the question “as an AI, do you want to take over the world?” is very different from considering the ideas, forming an opinion about them, and responding with that opinion. LLMs aren’t just doing statistics, but you don’t have to go too far down that spectrum before the answers start seeming thoughtful.

In its complaint, The New York Times alleges that because the AI tools have been trained on its content, they sometimes provide verbatim copies of sections of Times reports.

OpenAI said in its response Monday that so-called “regurgitation” is a “rare bug,” the occurrence of which it is working to reduce.

“We also expect our users to act responsibly; intentionally manipulating our models to regurgitate is not an appropriate use of our technology and is against our terms of use,” OpenAI said.

The tech company also accused The Times of “intentionally” manipulating ChatGPT or cherry-picking the copycat examples it detailed in its complaint.

https://www.cnn.com/2024/01/08/tech/openai-responds-new-york-times-copyright-lawsuit/index.html

The thing is, it doesn’t really matter if you have to “manipulate” ChatGPT into spitting out training material word-for-word, the fact that it’s possible at all is proof that, intentionally or not, that material has been encoded into the model itself. That might still be fair use, but it’s a lot weaker than the original argument, which was that nothing of the original material really remains after training, it’s all synthesized and blended with everything else to create something entirely new that doesn’t replicate the original.

The problem is the jokes aren’t funny. Or even really jokes. It’s just the same hateful garbage that you’ll find in any right wing comment section with no clever twist or respect for the humanity of the people being made fun of. It’s all variations on “haw haw, these people are pretending to be something they’re not, ew gross”. It’s not true, it’s not “keeping it real”, it’s not insightful, and anyone who actually knows or cares about the trans community knows that hearing that all the time will drive some people to kill themselves. Maybe even worse than that, it’ll foster that attitude in people even less compassionate that Dave Chappelle, who I don’t think has any particular malice toward individual trans people, but he’s telling those who do that they’re right.

There’s definitely humor to be had about the trans community, just visit any trans meme board and you’ll find it. There are stereotypes and self-deprecation and tons of really dark humor going on. What’s coming out of Chappelle’s mouth isn’t that, it’s just undercooked right wing bigotry.

These models aren’t great at tasks that require precision and analytical thinking. They’re trained on a fairly simple task, “if I give you some text, guess what the next bit of text is.” Sounds simple, but it’s incredibly powerful. Imagine if you could correctly guess the next bit of text for the sentence “The answer to the ultimate question of life, the universe, and everything is” or “The solution to the problems in the Middle East is”.

Recently, we’ve been seeing shockingly good results from models that do this task. They can synthesize unrelated subjects, and hold coherent conversations that sound very human. However, despite doing some things that up until recently only humans could do, they still aren’t at human-level intelligence. Humans read and write by taking in words, converting them into rich mental concepts, applying thoughts, feelings, and reasoning to them, and then converting the resulting concepts back into words to communicate with others. LLMs arguably might be doing some of this too, but they’re evaluated solely on words and therefore much more of their “thought process” is based on “what words are likely to come next” and not “is this concept being applied correctly” or “is this factual information”. Humans have much, much greater capacity than these models, and we live complex lives that act as an incredibly comprehensive training process. These models are small and trained very narrowly in comparison. Their excellent mimicry gives the illusion of a similarly rich inner life, but it’s mostly imitation.

All that comes down to the fact that these models aren’t great at complex reasoning and precise details. They’re just not trained for it. They got through “life” by picking plausible words and that’s mostly what they’ll continue to do. For writing a novel or poem, that’s good enough, but math and physics are more rigorous than that. They do seem to be able to handle code snippets now, mostly, which is progress, but in general this isn’t something that you can be completely confident in them doing correctly. They make silly mistakes because they aren’t really thinking it through. To them, there isn’t really much difference between answers like “that date is 7 days after Christmas” and “that date is 12 days after Christmas.” Which one it thinks is more correct is based on things it has seen, not necessarily an explicit counting process. You can also see this in things like that case where someone tried to use it to write a legal brief, where it came up with citations that seemed plausible but were in fact completely made up. It wasn’t trained on accurate citations, it was trained on words.

They also have a bad habit of sounding confident no matter what they’re saying, which makes it hard to use them for things you can’t check yourself. Anything they say could be right/accurate/good/not plagiarized, but the model won’t have a good sense of that, and if you don’t know either, you’re opening yourself up to risk of being misled.

There just isn’t much use for an approach like this, unfortunately. TypeScript doesn’t stand alone enough for it. If you want to know how functions work, you need to learn how JavaScript functions work, because TypeScript doesn’t change that. It adds some error checking on top of what’s already there, but that’s it.

An integrated approach would just be a JavaScript book with all the code samples edited slightly to include type annotations, a heavily revised chapter on types (which would be the only place where all those type annotations make any difference at all, in the rest of the book they’d just be there, unremarked upon), and a new chapter on interoperating with vanilla JavaScript. Seeing as the TypeScript documentation is already focused on those exact topics (adding type annotations to existing code, describing how types work, and how to work with other people’s JavaScript libraries that you want to use too), you can get almost exactly the same results by taking a JavaScript book and stapling the TypeScript documentation to the end of it, and it’d have the advantage of keeping the two separate so that you can easily tell what things belong to which side.

The Fairness Doctrine only survived the 1st Amendment because the airwaves are a public resource: each area only has one electromagnetic spectrum, and the sections of it that are useful for broadcasting are limited enough that not everyone can have a useful slice of the pie. As such, if you’re lucky enough to get a slice, the government gets to have a lot more control than they normally do over how you use it. You’re using something that belongs to all of us but only a few people get permission to use, so you have to do your part to serve the public good in addition to the programming you want to broadcast.

Cable has none of that scarcity, since we can have effectively as many cables in an area as we want, and each cable can be stuffed with more signal than the airwaves can, since you don’t have to worry about whether any given frequency can pass through walls or buildings, just copper. Without that, the government can no longer justify dictating content.

When 66 to 72 million years old you reach, look as good you will not, hmm?