It is surprising for me to see these features finally being added to Node after such a long time. Especially so when I remember reading discussion after discussion about how something like this wasn't possible. I touched on this in a blog post some time ago [1]. Glad Node is catching up.
I don't see in your blogpost any sources cited regarding anyone saying that ES modules were infeasible.
Additionally, io.js actually forked off due to internal drama which started with Ben Noordhuis having changed some pronouns here and there and people wanting to cancel him for that, to which he picked up his toys and left the sandbox.
It so happened that aside from being competent himself, he had competent people on his side, which eventually forced those governing Node.js to concede.
I have been wondering about the use of diffusion techniques for text generation, it is nice to see Google release a model that, seemingly, validates some thoughts I had.
Most folks I have seen experimenting with AI are either using a paid service or running high-grade hardware (even if consumer-level). The best I have in my current repertoire is a 5700XT and am not able to upgrade from that yet. The limitation, though, has at least also given some more significant insights into the shortcomings of current models.
Model sizes have gotten quite large and coherence seems to mostly have scaled with the density of a model, leaving the smaller models useful for only smaller tasks. Context size is also extremely important from my experiments with long-running dialogues and agent sessions, but a smaller GPU simply cannot fit a decent model and enough context at the same time. I do wonder if diffusion techniques will allow for a rebalancing of this density-to-coherence connection, letting smaller models produce chunks of coherent text even if limited by context. From my viewpoint it seems it will. Mixed tool call + response outputs also have the potential to be better.
Speed is also another problem I, and everyone else, has had with modern LLMs. The nature of cycling around the input with a new additional output each time is time consuming. On an older GPU with no AI-specific hardware it is an eternity! Being able to at least track 0-100% progress state would be an improvement from the current solution. At the moment one must simply wait for the LLM to decide to stop (or hit the max number of inference tokens). I am hopeful that, even on lower-end GPUs, a diffusion model will perform slightly better.
This does now beg several questions. If we are processing noise, where does the noise come from? Is there a good source of noise for LLMs/text specifically? Is the entire block sized beforehand or is it possible to have variable length in responses?
Lix is excellent. It is already faster (parsing), safer (better defaults, removed footguns), and easier to use (better errors, etc) than Nix. If anyone wants to get started using Nix then I highlight recommend you install Lix from the link in the parent comment.
Lix is a fork of Nix (the program, "CppNix") and also an alternative community. They provide a few ways to install, like nixpkgs itself, but also a DetSys installer fork.
It's generally a response to corporate interests (including Anduril, who put Nix on autonomous weapons) and elbow-shoving so-called meritocracy (including a very infamous concern troll) becoming dominant inside the Nix community and Nix leadership, by people who didn't want to work from inside the system to reform, but also know they can't maintain all of Nixpkgs by themselves either (which already bled a bunch of maintainers).
For those unfamiliar with the game I would highly recommend it if you are interested in CRPG games with excellent writing. There is a lot of text in this game, but with the most recent version of the game most of it is voice acted. Many lines will stick with you later. It's rare to not be taken by something in the game, as expansive as it is.
For a more general description of the game: you are a detective, you must solve the case, and your fractured psyche will not let you do it alone.
It's also an atypical videogame in that there's no significant combat mechanics (combat is as rare and exceptional as in real life), and the setting is far from your usual videogame setting (it feels like like alternate history mid-to-late 20th century Eastern block, but it's entirely fictional) and themes (ostensibly a detective game, but leans heavily into political commentary, morality, nostalgia). Despite the bleak setting, it's also super funny at times in both subtle and unsubtle ways.
In my opinion that's part of the game, intentionally so. You're a drunk detective who wakes up from a wicked bender and then has to go figure out how to solve a messed up case. Figuring out how do you even do that, trying and failing, is a great simulation of how being a detective actually is. In other words, figuring out how to play the game is a part of the game.
> figuring out how to play the game is a part of the game.
Well said! I'm 5 hours in the game. This is my third time "trying" the game. I'm glad I stumbled on some random walkthrough on YouTube and really got to know what's in store for me.
The game is about talking to people, that's pretty much it. If you treat the game like an interactive novel it makes sense. The skills and stuff really don't matter, just pick something and enjoy the ride.
It took me about a half dozen tries for it to click with me, but I unfortunately have no idea what it was that caused things to click. I felt like I was banging my head over and over against a wall until I wasn't. It seems like it's one of those high effort, high reward things, but I can't figure out what the effort is.
You're getting downvoted, but I think this is fair.
DE is not a game for everyone. It's a point-and-click adventure game that feels like it is intentionally designed to be obscure and difficult to follow. The game is forcing you to role play as a alcoholic detective who woke up after a bender and has no clue what is going on. You the player are just as clueless about the world as the character you're controlling is.
FWIW, I completed the game four times, and it was only on the fourth play through that I really felt like I understood how to play the game. Not just from a mechanical perspective, but also a how to role play as an alcoholic-detective-who has-no-clue-what's-going-on. IME, the game only truly makes sense when you play the game as Harry would.
There's a point where Kim tells Harry that he's probably a phenomenal detective, as he was able to close an absurd number of cases. And to get the best experience from the game, you have to match Harry's true level of authenticity, intelligence, openness, and assertiveness. You can finish the game as Hobocop, Superstar Cop, or Detective Costeau, and you can blame all the world's problems on communism and you'll solve the mystery, but you won't have the full experience.
The same is true for me on Safari. I've tried with and without AdGuard and the experience on plenty of sites that I just want to get a quick tidbit from is awfully painful.
Pikuma is excellent! I have enjoyed going through the videos on the YouTube channel to help understand some 3d graphics rendering techniques. Everything that I have seen has been as easy to digest as this post. Highly recommend checking out the rest of the catalogue if this kind of content is interesting to you.
As someone who grew up with iPhones coming out around the time I was in middle school and apps producing noises like this were used as pranks in class I have to ask that people don't do this. The sound is painful.
[1] https://kilo.bytesize.xyz/an-incorrect-specification