a_wild_dandan

Me too! If I must take out some forgettable item, it sits blocking the door.

Maybe we could try a hybrid pitch? “We need to protect these cute, furry little dudes…from parasites. Giv money pls.”

Please no! The damn things are already $60 a pop now. Let's keep such features as separate modules.

I am ecstatic that Microsoft open sourced this. After watching the demo video[1], my mind raced with all of the possibilities that GraphRAG unlocks. I'm planning to try GraphRAG + Llama3 on my MacBook, since it has 96GB of unified (V)RAM. I think this tool could be a legit game changer.

[1] https://www.youtube.com/watch?v=r09tJfON6kE

I love React. It feels foundational somehow. It's also vindicating after toiling for years on bloated frameworks. Declaratively describe your UI, React paints it. That's it. No app state management. No routing. No learning yet another templating system (which inevitably morphs into a shitty programming language). Just full JS(X optional). No grokking convoluted, bespoke abstractions. It's clean, narrowly scoped, and as simple as possible (but not simpler).

Having recently worked on two legacy apps -- one ASP.NET, one Angular -- it made me deeply appreciate React and how far UI tech has come. Phenomenal work from the React team on more optimizations via this compiler. Good article too.

My guess: It's probably downvoted for being so hilariously wrong that's not worth engaging with. One of those dismissive hot takes commensurately worth dismissing too.

No accounting for taste, I suppose. For me, hooks were when React got amazing.

> They’re saying that V8 is a very capable optimizer (which it is), so why can’t it optimize the things React Compiler optimizes?

Because capable != exhaustive? All compiled languages can benefit from further optimizations. I genuinely don't understand the issue here. Help!

As I recall, philosophy is somewhat of a “fixed point”, or attractor. Wherever you start, you’ll land on philosophy, and stay there. You won’t random walk far away again.

It makes sense. After all, philosophy is concerned with the fundaments of knowledge, so an article on a knowledge sharing platform could be consistently near philosophy.

A more charitable explanation: universities use publication frequency as a proxy for your academic worth. Once the "publish or perish" pressure is released, you can focus on paper quality over quantity. Given the fire hose of trivial academic publications filling up journals, that's good news for everyone.

"If you can't explain something in simple terms, you don't understand it." - Richard Feynman

Experts deeply understand their subject, and tailor answers to meet the audience where they're at. If you express genuine interest in someone's passion, they'll be ecstatic. That's what the OP is talking about. It's not about how experts (or anyone) interacts with belligerent and dismissive interlocutors. And observing teaching moments is just one aspect of a larger smell test for detecting imposters. I think it's a good heuristic within that scope!

Wait, why does GPU encoding affect the output quality? Is that common in transcoding stuff?

Isn't their example even simpler? In JS/TS, for example:

```ts

  const pluralizePhrase = (noun: string, count: number) =>
    count == 1 ? `There is one ${noun}` : `There are ${count || "no"} ${noun}s`;

```

Why not use a beefy, Linux compatible laptop at this point? I'm unfamiliar with ChromeOS, so apologies for a potentially stupid question.

In a recent interview, Ben Goertzel mentioned having a PhD-level discussion with GPT-4 about novel math ideas cooked up by himself and a colleague. These things are becoming scary good at reasoning and heavyweight topics. As the ML field continues focusing on adding more System Two capabilities to complement LLMs' largely System One thinking, things will get wild.

Snark and contrarianism are specialties around here. Thank you for sharing your experience; I genuinely appreciate it.

“…involving 75 participants who had experienced a minor stroke and showed signs of mild to moderate small vessel disease.”

I disabled it for being annoying and frequently wrong. The idea that someone might interpret my laziness as a flex (or vise versa) is hilarious to me.

Seems this Altman fella isn't being consistently candid with us.

I must use PyTorch for ML, but badly wish I could use TypeScript instead. Its type system demolishes anything else I've used. Writing anything from SPI communications, to server logic, to rich GUIs is a breeze. Then I work in Python and have to fight `Any`s, ambiguous arguments, more verbose code, and bizarrely slow run times. It makes me sad.

Blink twice to low-key acknowledge the non-disparagement contract.

The Sky voice actress could legitimately lose clients who fear "likeness" litigation, for all Her potential contracts, regardless of context. Your analogy needs work.

This is why open source must win. I mean, it won't, but it's the only path to avoiding a silicon aristocracy.

A famous actor still has acting, name recognition, and generally mountains of cash. But should not get to call dibs on some vague "likeness" simply because they got hot first. Someone shouldn't damage/lose their livelihood because their famous doppelganger iced them out.

GPT-4o is a phenomenal leap forward for true multimodal models, inference efficiency, and speech generation; it's also replacing the free public model. That's massive news. In contrast, the Sky voice drama is vastly overblown and a blatantly obvious/predictable/avoidable mistake by OpenAI. But it'll be memory-holed in a week. Two, at most.

ClopenAI is a terrible company, but calling them desperate over a PR faux pas is pathetic.

The example song reminds me of playing Divinity: Original Sin 2. I wonder if game soundtracks were part of training.

Yeah, influential people shouldn't get to functionally own a "likeness." It's not a fingerprint. An actor shouldn't credibly worry about getting work because a rich/famous doppelganger exists (which may threaten clientele).

Explicit brand reference? Bad. Circumstantial insinuation? Let it go.

Open source/science! If open models approach/maintain near-SotA capabilities, then there's hope for avoiding a Big Tech AI aristocracy. Currently, that means Meta releasing their 405B GPT-4 killer (and probably subsequent models).

We'll need community unity, to implement research results, test/harden models, create/aggregate/clean massive new datasets, figure out distributed training, lower hardware requirements, encourage experts to join the effort, praise companies who contribute to open source, etc.

You don't have to tell me that we'll fail. I know. But we have to try. It's the only way we'll have AI aligned to us, not a handful of techbros with savior complexes.

Relieving some latent demand is a great thing, especially when you’re the one needing that last new hospital bed, traffic lane, or supercomputer time slot to predict who the hurricane endangers. More throughput is wonderful. When unsure, best to endorse meeting demand rather than judging it as unworthy.

We should call them ClopenAI to acknowledge their almost comical level of backstabbing/rug-pulling.