norswap

It's fairly easy to teach the behaviour of LR parsers (though which grammars end up being LR and which end up conflicting is very inintuitive).

It's much harder to teach the inner working / implementation of such parsers. Table generation adds an extra layer of indirection. I skip this in my class, because I doubt it will be very valuable to students to be able to implement an LR parser compared to other things we cover in the class.

Doing the OP's approach, you can get some intuition for how a LR parser behaves / is implemented without going to the full complexity.

Shameless plug: https://www.youtube.com/playlist?list=PLOech0kWpH8-njQpmSNGS...

The first half of the course is on parsing.

I'm hoping I'm giving a lot of examples and there is some coding (using PEG, which is closer to "programmer intuition").

This is a pretty neat approach to parsing by hand, and it's probably helpful to build intuition about LR. Basically if you give an explanation of LR including shifts and reduces, then you can show how that maps to the hand-written parser (i.e. the `_cont` functions are shifts that eventually result in a reduce when they return and use the parameter that they were passed).

Shameless self-plug, in case you want to understand more about LR parsing: https://www.youtube.com/watch?v=8UDWd-Axd5A

The trick is to pass down previous syntax nodes so that you can continue consuming input (via top down recursive descent) but build the syntax tree bottom up still.

This is exemplified in the two definitions of `parse_add_addexpr_cont` in the article.

SUNSET the employees.

Could they have done better? Sure. Hindsight is 20/20.

Would it have been a reasonable assumption? No. As pointed out in the parent, every bank pretty much works like this, they're all liable to go under in the presence of a big enough run.

Also if it must have been so obvious to them, it probably should have been obvious to the regulatory agencies in charge of monitoring banks and avoiding exactly these situations, as well as the Fed who set the policies leading to this in the first place.

It just wasn't part of anyone's threat model.

It's important to note that WorldCoin does not get to see the retinas, whose picture never leave the orb (and are deleted after the fact), but only a hash of them. The orb's hardware blueprints and the software is open source too.

The system attempts to solve an important problem: figure out who's a human (not a robot) online. One could argue Sam is creating the problem as well as the solution, I suppose. Still, it's better than only having the problem.

Right now the problem does not seem extremely pressing, but I believe it might become more so.

Even if we don't see rampant abuse of AIs masquerading as humans, another ambition of WorldCoin is to perform wide-ranging experiments in UBI, and being able to distinguish "real" humans in that context is absolutely crucial. This goes doubly in the third world, where people often simply don't have IDs (and available forms of IDs can be easily manufactured through bribery).

(That being said, I broadly agree with the criticism of OpenAI laid out in the above article. Still, we can have nuance.)

As a seasoned engineer (compiler background), but beginner in frontend, I'm curious aboutthe lack of expressive power that Svelte incurs.

Could you (or someone) characterize things that are hard to do in Svelte?

Some of Google's biggest successes are YouTube and Android, which were both acquisitions.

I hope he succeeds at making something great.

However, I'd have one criticism of this (which I've had of other projects including some I've been associated with), and it's that if you're trying to make something radically different than what exists (be that in society, or in the blockchain or metaverse worlds), then it's a bit weird to structure your venture in exactly the same ways as these other things.

Besides Stephenson's word, there is nothing in the presentation that suggests that this is any different from the other initiatives out there.

If ousting commercial interests is a goal, shouldn't this be structured in a much more decentralized way rather than having a corporate startup with a bunch of investors? I'm sure Stephenson doesn't need the money, and I'm sure that would inspire a bunch of hackers to contribute much more than the current structure.

There's stuff like this on Arte (Franco-German culture TV channel) all the time.

If you look at the numbers, more crypto market cap (both in % and absolute amounts) was lost before any of the scandals than after.

> Could this mounting unrest be as a result of the seeming inability of the US Federal bank in checking inflation of the US Dollar.

This is completely backwards. The problem right now is that the dollar is too strong, meaning expensive wrt other currencies. In this context, inflation of the USD supply, leading to devaluation of the USD would actually help other countries that are suffering from the strong dollar.

I'm not saying it would be good for the US, but it certainly would be good for pretty much everybody else.

Another problem is the bank run that this narrative created.

Go is a moderately better C. In terms of expressivity, it is ages behind C++.

It also has a garbage collector, which while fine for most use cases is not acceptable for some use cases that C/C++ (and now Rust) are being used for.

It's fully open source still (as it has always been), it just won't be part of OpenJDK.

So what's the serious laptop choice for the professional, besides Apple? (not that Apple is without its own problems, at least in some generations)

This won't convince anyone. The truth is that Twitter thread get a ton of engagement that articles don't — even if you post them on Twitter. Leaving the platform, particularly on mobile, is additional friction that readers might not be willing to take on.

Quick notes:

- Your website says "Etherium" instead of "Ethereum" btw.

- Also "Is there ... fees" instead of "Are there ... fees"

- Link to the demo video doesn't work.

Otherwise sounds cool, but I'd like to know about the security/trust model.

Sane policies like shuttering nuclear (in favour of more coal and reliance on Russia)? Yikes

Are you sure they're ideologically against it, or where they against it because they were in the opposition and it was convenient for them to be?

Why do you assume they're holding the money for legimitate reasons? I count as much more likely they're just fucking up and have no idea or don't care enough.

Plenty of example of this, event amongst top tech companies: see the clusterfuck that is youtube copyright claims for instance.

Every time there are automated flaggings, there are fuckups. We can accept it if there's an efficient way to challenge, but that's not happening. Stripe/Google doesn't care and/or it would hurt the bottom line too much.

Also if credit cards company are much better at this, I don't see why Stripe gets a free pass for sucking.

Also that the protagonist (Wilhelm of Orange) has always respected the Spanish king.

What are the advantages of this being its own language vs a DSL?

Would recommend reading Ben Edgington's annotated spec: https://benjaminion.xyz/eth2-annotated-spec/phase0/beacon-ch...

It's not entirely up to date (none of the annotated specs are) but out of the three annotated specs I'm aware of, it's the one that is most focused on explaining how things work. Vitalik's annotation (linked in another comment) places a particular focus on the reason specific constants where chosen.

Is JavaScriptCore really faster than V8 in general?

I'm kind of befuddled as to why the original "Modifier functions" solution is not better than the "And now we can finally unify our getter and modify lenses into one:" solution.

Ha. This is so dumb. People can already easily link a Medium page / github gist / whatever. The reason people write threads is that people don't like to click links that get them out of the Twitter flow. Many reasons for this, but basically it's friction and Twitter is very flaky with keeping your position within the timeline (at least on Android).

Keeping it in app should be a step forward, but a clickable "read more" that expand *within the timeline* (same way it works for threads within the timeline, though unfortunately not within list) is the only thing that will kill threads.

Plugging the project of a friend: https://mazury.xyz/

It's web3-focused for now (and the blockchain is core to how it works: i.e. to collect certified attestations) but I think it has much broader ambitions.

> If you do a CT scan and it shows a mass, you’ll order a biopsy.

> But because that patient was low-risk, the harms of that biopsy will outweigh the benefits.

> Thus, you shouldn’t do the CT scan.

> What? If the harms of the biopsy outweigh the benefits, don’t do the damn biopsy!

I'm not sure what the point of doing the CT scan if we're not going to act on its result.

Sure, you gain the confidence that everything's gucci if it turns negative. Is that really worth the stress of the test going positive and the not doing the biopsy? Seems dubious to me.