ForceBru

This is crazy to me as well: half a terabyte (!) of RAM?! How is that even possible lol (I know this is possible, but seems totally overkill, even though I know LLMs eat RAM for breakfast)

Ah, yeah, I guess I meant processes, indeed

I personally don't understand the point of an M4 iPad to begin with, let alone M4 Pro. What kind of CPU-intensive work can you do on an iPad? Video editing, I guess? You can't do any serious programming because there's no proper terminal emulator and you can't compile and run custom code. All you have are interpreted languages like Python and Ruby, but it seems like they still can't use the full power of an M4 because iPadOS doesn't let you spawn threads AFAIK.

Even for photo/video editing, why not use a Mac mini with the same (!) CPU and proper cooling? An iPad doesn't even have a fan, so surely the CPU will throttle.

"A glorified $1000 YouTube machine" is a pretty accurate description of an iPad LMAO

I think this is the entire point, you're supposed to subconsciously believe that you're talking to a human, not a matrix multiplication algorithm

Google trying to break up human relationships: "Are you ready to just walk away? Huh? Are you ready to do it? I have no idea who the guy is, but you better be walking away right now. Do you want to send one final text? Pretty pathetic if you ask me. Has his behavior changed? Just walk away."

That's a nice overview, but it seems to compare iOS 27 to iOS 26. Isn't iOS 26 known to be terrible, though? I'd like a comparison to a "known-good" iOS version. Basically I skipped iOS 26 (so I have zero experience with it) and am now wondering whether iOS 27 is any good

So, is it better (faster, stronger, etc) than iOS 18? Everyone says iOS 26 was not good at all and iOS 27 is much better, but it seems like it's fairly simple to make something that's better than iOS 26. A more interesting question is whether it's better than iOS 18? Does anyone have any experience switching from iOS 18 to, say, some beta of iOS 27?

Well, it's fast. Perhaps "they" want me and everyone else unemployed, but it's not immediate and not particularly certain. However, I can use uv here and now to greatly simplify my workflows, so I use it.

Sure, but these ten characters let you treat columns as values and do math on them, which is super intuitive, in my opinion. I've been using Pandas for quite some time and always kinda sucked at it. One day I decided to give this new library Polars a try. Now I can do things I couldn't even dream of with Pandas! And it's fast, too!

I think of `pl.col` as delayed evaluation: I want to do math on the vector of values of this column. But wait, let me just refer to the name of that column and build the expression that I want to compute. Then I hand this expression to Polars and it retrieves the actual values of the columns my expression refers to and executes the operations.

IMO, it would've been great to just do math on strings, like `"Amount" * "Price" - "Losses"`, but programming languages either don't allow math on strings or that math is actually string concatenation, which is not what we want. So we have to wrap the name of the column into some object. This is just an API thing.

As a side note, it's such a pity that there's basically no Polars for the Julia language! There is some wrapper package, but it seems old and unmaintained. I can't seem to properly learn DataFrames.jl for some reason, I always miss Polars when I use Julia.

How is, say, `age = 5; print(f"Age: {age} years old")` not easy for beginners? IMO it's as easy as it gets: you want the value of `age` printed {HERE}, so you just put it where you want it, surrounded by curly brackets.

I'm a HUGE fan of f-strings and I think they should be added to more or less every language in existence. (This is just to show how much I like f-strings, not to be taken literally) `printf`-style format strings seem outdated: why use format strings when I can put my variables IN THE STRING? I want the result of `x+y` to be put {HERE} in this string. Well, just write `"This is here: {x+y} blah"` — it makes perfect sense and immediately lets me see what the resulting strings generally look like.

Basic usage is a no-brainer: write your string, put variables or short expressions in curly braces, add `printf`-style format specifiers after the colon. This is also great because it's a natural extension of `printf`-style format strings.

Of course you can write complicated and confusing f-strings. But then you can write complicated and confusing... anything, really. Many programming languages have extremely weird quirks and cases where basic syntax can be transformed into an unreadable monstrosity, like C syntax for pointers to functions and arrays.

Sure, this increases the language's complexity, but you don't have to use all of it to reap the benefits.

Are there any real-world time-series that are known to be easy to forecast? Something I could use as an example of "this is what forecasts look like when it's actually possible to predict your time-series and your model is actually doing it"?

Yes, I'm diving a bit too deeply because I don't really know what "thinking" is and therefore I don't understand how we can so confidently say that LLMs don't think, even though they definitely LOOK like they're thinking. They even have a "Thinking" section in their responses! If I say that a rock doesn't think, it's pretty convincing: does a rock look like it's thinking? No — it doesn't even do anything! But an LLM does look like it's thinking, at least while generating a response. When it's "offline" it's just a bunch of "dead" bytes, sure.

So when it's not active, not responding to a prompt, it's of course not thinking. I'm pretty sure nobody actually questions this. Is your computer "thinking" when it's powered off? Can a piece of metal think? Probably not. So there are no thoughts between prompts, this seems obvious.

Thus, this is a question of "discrete time vs continuous time". LLMs "live" from prompt to prompt. Humans are alive continuously. In some sense, we're prompted by a lot of things all the time. As I'm writing this, I'm seeing stuff, I'm hearing stuff, I can feel various parts of my body, I'm thinking about my problems, my goals, other people's problems and goals, etc. When I'm in a sensory deprivation tank, my brain keeps "entertaining" me by "self-prompting", like a recurrent neural network (I guess it literally is a massive RNN).

So it seems like your definition of "thinking" hinges upon the LLMs being discrete-time and single-threaded (can't think about multiple things in parallel).

IMO a more interesting question is whether an LLM is thinking WHILE IT'S GENERATING A RESPONSE, while it's "alive".

LMAO at how the two of you sound authoritative and knowledgeable, but neither linked to ANY studies (or at least personal anecdotes) to support your claims.

Yet here we are, warning each other about the dangers of LLM hallucinations. Humans "hallucinate" (provide random authoritative-looking information without anything to back it up) pretty often too.