crvdgc

> Free plans from mobile carriers had/have access to a few selected websites and not the rest of the real internet.

Interesting to learn. In China, mobile carriers also have similar deals with huge content providers. You can buy data plans with a discounted rate specifically for several YouTube, Netflix, and TikTok like apps.

Here's a tool suite to simulate and run memory model litmus tests (on real hardware): https://github.com/herd/herdtools7

The simulation tool can also generate relation graphs similar to those of the blog.

What about Cardboard VR like headset + a keyboard? Has anyone tried this setup?

Could be both, like Better Call Saul.

In Japan, sometimes earthquakes will cause regional power outages, though they are usually recovered quickly.

They meant to name Haskell the programming language Curry at first, but the proposal was struk down because of the potential association with Tim Curry.

https://youtu.be/LnX3B9oaKzw?t=3m6s

You've heard sovereign AI before, now introducing sovereign system prompts.

Thank you for pointing this out. Apparently I underestimated exponential growth rate.

> The Centaur Effect

> even when AI chess engines can easily defeat grandmasters, the human-AI combination still produces superior results to the AI alone.

Is this still the case? I didn't find a conclusive answer, but intuitively it's hard to believe. With limitless resources, AI can perform exhaustive search and is thus not possible to lose. Even with resource limits, something like AlphaZero can be very strong. Would AlphaZero+human beat pure AlphaZero?

Black Mirror of all work should not be called "pessimism porn". It's not unrealistic for the depicted scenarios to happen in real life.

You can kill a messenger for delivering a false message, but not a true one, even when it's bad.

A perspective from a friend, who recently gave up trying to get into concept art:

Before AI, there was out-sourcing. With mass-produced cheap works, foreign studios eliminated most junior positions.

Now AI is just taking this trend to its logical extreme: out-sourcing to machines, the ultimate form of out-sourcing. The cost approaches to 0 and the quantity approaches to infinity.

> But in recent months I've spoken to other YC founders doing AI application startups [...] in different industries, on different problem sets.

Maybe they should create a benchmark collectively called YC founders. Gather various test cases. Never make it public. And use that to evaluate newly released models.

Despite the inflammatory title, it's David Graeber's review of Money and Government: The Past and Future of Economics by Robert Skidelsky. For those who are interested in the history of economics, it provides a nonorthodox narrative.

Using Agent-2 to monitor Agent-3 sounds unnervingly similar to the plot of Philip K. Dick's Vulcan's Hammer [1]. An old super AI is used to fight a new version, named Vulcan 2 and Vulcan 3 respectively!

[1] https://en.wikipedia.org/wiki/Vulcan's_Hammer

My two cents: isn't this arrangement more of a continuation of the Marshall Plan? If the US did nothing, post-war Europe was likely to fall into the influence of the Soviet.

Now even the Soviet has dissolved, there's always the menace of Europe allying with Russia or China (from the US perspective). So it's worth to pay a premium to prevent your opponent from getting the same thing.

I had expected to see a more programming language theory flavored formal specification, but it's actually just prose describing the language for industrial conformance purposes. I guess something is better than nothing, but still.

Related Anki deck: https://github.com/anki-geo/ultimate-geography

The data is also a bit more up to date w.r.t. the problems pointed out by sibling comments like changed names, different claims, distinguishing territories from countries.

Also commented in the original post.

Andrew Yang: Data is the new oil.

Sociology: The Oil Curse

Imagine adding types to a legacy JavaScript codebase. You can turn everything to valid TypeScript by annotating `any` everywhere, then you can gradually add types here and there.

Or imagine writing Rust with `Rc` everywhere and then using the borrowing style on the hot path.

I can see where the author is coming from, but sadly a difficulty is that dependently typed languages often require the programmer to prove type equality during type checking. It's hard to do if the information is not complete.

AGI aside, sometimes HN critics/cynicism indeed points out the exact reason why something wouldn't work and is vindicated after the fact, e.g. Apple Vision Pro. I guess it's just hard to predict the future and for me, it's interesting to listen to even pure contrarians.

My colleague taught me to never close the editor, but just suspend it to background. You can achieve a lot with Ctrl-Z, fg, and :e!.

That private enforcing company has zero motivation to make the rule clear and easy to understand. Their letter soliciting license fees or a declaration of not needing a license is very obnoxious and threatening. Part of the strategy is that "you won't know whether you need a license, but if we caught you violating this arbitrary opaque rule, you're screwed. So you better pay that money."

The project itself is cool, but I'm doubtful on the philosophical side.

Mermaid aspires to be the Markdown of diagrams, so being editable and consumable via plaintext is paramount. If I have to use a visual editor, why not fetch for more powerful tools and be liberated from the limiting format?

This is akin to a WYSIWYG Markdown editor (which to be fair has a niche market). I would ask why not just use Google Doc or MS Word at that point?

A co-worker once said: a good language specification should allow others to derive an independent implementation of the language.

Realistically, it's probably the group inside the Chinese Academy of Sciences wanting for a larger impact factor for their review, similar to all other state sponsored researching institutes.

Maybe the title needs to be edited to "unsafe Rust".

Thanks for the recommendation.

From a Lisp-like introduction, to OOP (in Java), then to ACL2, and finally back to OOD. It's quite interesting in its own right, but now that functional programming is more and more adopted by the mainstream, perhaps there's no need for the "practical appeal" of OOP/OOD in the mix, or does the author genuinely believe that's the way to go?

I made a similar transition (1 year Haskell, 2 year OCaml) and this pretty matches up my experience. Some more points:

1. Compiler speed: a clear win for OCaml

2. OCaml 's module system is more explicit (always qualified Map.find). Haskell's type class system is more ergonomic (the instance is found through its type, so no explicit annotation needed, e.g. fmap fmap fmap, where the second must be the Reader's fmap).

3. > If I come to an existing OCaml project, the worst thing previous developers could do to it is have poor variable names, minimal documentation, and 200+ LOC functions. That’s fine, nothing extraordinary, I can handle that.

Though it's not common, but functor-heavy codebase does give you a headache. On the other hand, unifying type class instances across packages is no fun either.

4. OCaml's mixture of side effects and pure code tends to encourage using that in libraries and codebase. So expect more speed and more crashes.