HNHacker News
TopNewBestAskShowJobs

dgroshev

1,778 karma · joined September 30, 2014

submissionscomments
dgroshev··on Meta Fined $843M by EU over Marketplace Ads
They don't control the products that are being advertised, but they make personalised ads for Marketplace products more effective by using data they collect elsewhere. If the ads are more effective, Marketplace is better at shifting goods, undermining the competition because their competition is forced to share data with Facebook through Facebook's (not Marketplace's) market position.
dgroshev··on Meta Fined $843M by EU over Marketplace Ads
"Normal" online advertising is neutral with respect to companies bidding on placements. Not so much with Facebook, which is under no obligation to run a bid on every ad they serve. Some of those ads are dedicated Marketplace ads, and AcmeShop can't bid on them, or even limit access to the data that Facebook collected while performing their contractual duties to AcmeShop.
dgroshev··on Meta Fined $843M by EU over Marketplace Ads
> In your scenario, Alice already decided not to buy anything from AcmeShop.

Not really, that's exactly what "remarketing" ads are for — companies track people's interest and remind them of their products. However, in this case, Facebook is double dipping by taking money from their paying customer (AcmeShop), and then using data they obtained whilst serving their paying customer to frontrun the customer.

If you or I tried something like that with Facebook — say, taking salary from Facebook and also using data we accessed through our role to undermine Facebook's business — it'd clearly be both illegal and immoral. Somehow, with enough layers of indirection, for a lot of people the same play is acceptable when Facebook does it.

dgroshev··on Meta Fined $843M by EU over Marketplace Ads
I suspect the latter part is so commonplace that many technologists don't even think twice about it.

> Unilaterally imposing unfair trading conditions on other online classified ads service providers who advertise on Meta's platforms, in particular on its very popular social networks Facebook and Instagram. This allows Meta to use ads-related data generated by other advertisers for the sole benefit of Facebook Marketplace.

Consider this:

- A local marketplace AcmeShop wants to advertise itself on Facebook, paying Facebook for ads.

- The only practical way to do so requires AcmeShop to install Facebook's tracking on their website. The tracking is always dual purpose: on the one hand, it gives AcmeShop data on conversion from their ads (which is vital for AcmeShop), but on the other it also gives Facebook data about AcmeShop users.

- Alice wants to buy a used 3D printer. She goes to AcmeShop, searches for 3D printers, checks a few listings out, but doesn't buy anything.

- Alice opens Facebook. At this point, Facebook registered Alice's visit to AcmeShop 3D printer listings, so of course they serve ads for 3D printers selling on Facebook Marketplace.

- Alice buys a used 3D printer from Facebook Marketplace, cutting AcmeShop out.

AcmeShop has a tough choice: they either advertise on Facebook, pay Facebook, and help Facebook to undermine their own business, or they forgo Facebook as a marketing channel altogether.

Usually different parts of this scenario are discussed in isolation (performance marketing, user tracking, privacy on Facebook, etc.), obscuring the larger picture. I don't see how Facebook's behaviour taken as a whole is anything but predatory and abusive of their position.

dgroshev··on Is the attack helicopter dead?
Would you ever see footage of them not hitting weak spots?
dgroshev··on New laser could cram GPS alternative into a shoebox
> a low-yield device over the Black Sea or the North Sea

How is it different? What "panic reactions"? What "panic reaction" can connect a low-yield explosion over Black Sea and a suicidal strategic strike on the continental US?

> I'm just advocating that we observe the cautionary principle here

"The cautionary principle" is only applicable if you disregard hundreds of thousands of killed and maimed Ukrainians, millions of refugees, unknown number of abducted children, and entire cities raised to the ground. Letting a genocide continue for years is not a default safe option.

Again, I invite you to reflect on this exchange: you'd "rather not delve into scenarios", but when weighing what's basically just vibes against very material, prolonged, ongoing suffering of the second largest country in Europe you decide to advocate for vibes. Putin's bullshit has already won and his deterrence succeeded, because you give it enough credibility (whilst refusing to outline a credible scenario) to talk about "WW3" as a real risk worth more than dead and maimed Ukrainians.

dgroshev··on New laser could cram GPS alternative into a shoebox
> an entirely possible provocation that could shove things in that direction

How? What plausible pathway do you see between Russia exploding a nuke over Novaya Zemlya and their suicide by American nuclear response?

> A major shift in the ground situation Ukraine's favor (effectively signaling an eventual Russian defeat) could do the trick.

I don't understand what trick. Russia directly attacking NATO? They'd get pulverised, and they have no forces left. Russia committing a suicide by launching a strategic attack on the US over Donbas? That's implausible. They just lost a chunk of Russia proper and took it on the chin.

> I guess everyone who doesn't agree with your assessments is either paralyzed by irrational fears, and/or a useful idiot, or just looking for a cop-out.

I see how it's easy to interpret the earlier reply personally, my bad, sorry. I was referring to the people establishing the narrative (media, politicians, and people with large platforms). It's self evidently working! A direct confrontation with Russia (a country with an economy smaller than Italy's, glaring manpower problems, and largely depleted army) evokes the idea of WW3, but I see very few people talking about the very likely war over Taiwan (with a much, much, much more powerful country) in "WW3" terms. This narrative doesn't make any logical sense, we just got used to it, even though it's not grounded in reality.

dgroshev··on New laser could cram GPS alternative into a shoebox
I don't think there is any conceivable scenario in which Russia using nuclear weapons would lead to anything resembling "World War", except Russia launching a counter-value strike against continental US, which is just not happening. A high-altitude detonation is not a "World War".

Russia is a China's client state at this point, and China is very interested in keeping nukes taboo. They are so worried their own army is only allowed melee weapons in border skirmishes with India. There is no way a demonstrative breaking of the taboo would not get Putin's leash yanked, so even that remote possibility (that has nothing to do with WW3) is extremely improbable and would only hasten Russia's defeat. Unless, of course, our politicians use it as an excuse to abandon Ukraine.

Also, no one is "pushing Russia's back against the wall" in Ukraine, there are no M1A2s speeding to Moscow. Russia understands it perfectly well, too, denuding their borders with NATO to throw more soldiers into the meat grinder.

> the fact that we can't assume these people are acting rationally

I think this meme is largely a cop-out. A person just doesn't get to be a leader of a country by acting irrationally. They might interpret things differently or have different priorities, they might even put on an act of irrational buffoonery, but it's different from being "irrational". However, to see the rationality of their actions requires putting yourself in theirs shoes, which is uncomfortable and/or requires a deeper understanding, so it's easier to just write their behaviour as "irrational".

Putin is perfectly rational. The empty threats are working (as evidenced by this thread), there is no risk of a strategic exchange (because the actual signalling on the ground is one of business as usual), and Western politicians are successfully convincing their voters that they drag their feet because WW3, not because actually stopping the war (eg deploying a NATO peace keeping force) would incur a substantially higher domestic cost.

We don't really talk about potential North Korean aggression in WW3 terms, do we? We shouldn't do that in Russian case either.

dgroshev··on New laser could cram GPS alternative into a shoebox
> you don't put a nuke somewhere to kill people; you put it there to destroy infrastructure. More specifically nuclear missile silos, and bomb shelters (well, the people inside them.)

It depends.

"Counter-force" strikes aim to destroy the ability to respond or attack with nuclear weapons, not infrastructure per se. So that's just silos, warhead storage sites, submarines in port, airfields, etc. The US is doctrinally counter-force for about half a century. This doctrine is also reflected in technological choices: lower yields and higher precision mean lower fallout, collateral damage, and fratricide when hunting for silos, but aren't that useful for city busting.

"Counter-value" strikes are pure destruction of cities and killing of people. Russia seems to have this doctrine [1]. This is a very inflexible doctrine, because you can't credibly threaten to execute a counter-value attack on the US over things that don't seem to be worth a response in kind (like a defeat in Ukraine). It's much cheaper though.

> The very latest US ICBMs are suspected of having sub-100m CEP

Apparently the precision is good enough for photos like the one on page 23 here https://www.osti.gov/servlets/purl/1557069

[1]: At least strategically. There is a hybrid strategy of targeting military-related infrastructure, which Russia seems to have adopted for less-than-strategic forces according to recent leaks https://www.ft.com/content/237e1e55-401d-4eeb-875b-03fe68f81... .

dgroshev··on New laser could cram GPS alternative into a shoebox
There is no fear of WWIII, that's part media invention, part political excuse, part useful idiots. A "World War" requires an opponent worthy of a world-wide conflict, and Russia is anything but, being stuck in Ukraine for three years and losing most of its army in the process. Conventionally, Russia is no match for NATO, with enough political will it'd be a Desert Storm-like curb stomp. The only way it's not is the US abandoning NATO, then it can be quite painful, but even then Russian economy is no match for the EU.

As for nuclear weapons, they are practically useless. Anything but strategic use would have little impact on the battlefield, but will immediately alienate China and India. They have their own border conflicts and do NOT want the nuclear taboo to be broken, and they are holding the leash, being the only economic and technological lifeline Russia has. A strike against the US would be suicidal for obvious reasons.

Also, notice how Putin and his officials talk about nuclear use a lot, but there is still zero evidence of any action (such as dispersal of road-based ICBMs or 12th GUMO taking warheads out of storage). Rhetoric is cheap and not credible.

The only real fear of escalation here is "horizontal" escalation: Russia stirring shit elsewhere. Advanced anti-ship missiles to Houthis, handing nuclear technology to North Koreans or submarine technology to China, that kind of thing. But that's already seem to be happening, and there isn't much more that Russia can do.

dgroshev··on Why the U.S. can't build icebreaking ships
Note that FPV drones in Ukraine are not squad-level, there are separate drone units (at the very least 2 operators, at least 1 explosives guy because arming drones with makeshift explosives is not trivial and super risky, plus cover). Plus they are supplied literally on the back of a minivan in cardboard boxes because they only need to survive one way trip over a couple hundred km, the US needs to supply their units half a globe away.

US army drones are squad-level, have proper safing/arming, transportable, and shelf-stable. It's not scope creep, it's a very different doctrine and circumstances.

dgroshev··on Rewriting Rust
I agree wholeheartedly (and I'm not surprised that you of all people often write raw futures!). I want to push back on the "async rust bad/failure/not ready" meme because

- it's perfectly possible to be a successful user of the async ecosystem as it is now while building great software;

- this two-tiered phenomenon is not unique to Rust, JS and Python struggle with it just as much (if not more due to less refined and messier design). As an example, [1] is elegant, but complex, and I'm less sure it's correct compared to a gnarly async Rust future, because the underlying async semantics are in flux.

Of course I'd love for the remaining snags (like AFIT) to go away, and simplified Pin story or better APIs would be great, but this negativity around async Rust is just wrong. It's a massive success already and should be celebrated.

[1]: https://github.com/florimondmanca/aiometer/blob/master/src/a...

dgroshev··on Rewriting Rust
Every single JS future is boxed. Moreover, they aren't just boxed, they are often backed by a hashmap (which may or may not be optimised away by the JIT). Elaborate allocation-free async is not an apple-to-apples comparison.

JS does support concurrent execution, Promise.all is an example. Without it, JS async would make little sense. The problem very much exists there, and try-catch is only a surface-level answer. As you can see here [1], the interaction of cancellation and async in JS is at least just as (or more) complex than in Rust.

By the way, multithreading has little to do with Pin. I presume you're thinking of Send bounds.

"To work at all" is very dismissive. Those wrappers are complex, but very well abstracted, well defined, and robust, the complexity is essential. Again, look at [1], JS async is hardly less complex, but also much more vague and ill-defined.

[1]: https://github.com/whatwg/streams/issues/1255

dgroshev··on Rewriting Rust
> Then I spent 2 weeks trying to do the same thing in rust… Eventually I gave up.

In my experience, that kind of difference boils down to a combination of three things.

- Comparing apples and oranges. For example, Box makes pinning trivial (you can just move in and out of Pin no problem), but oftentimes people new to Rust try to prematurely optimise and eliminate a single pointer lookup. If that's the case, were you really writing the same thing in JS and in Rust?

- An extension to the previous point, the behaviour is usually different. What would happen in your JS implementation if two streams were awaited concurrently, one received a message, and the other had to be cancelled? What if one threw an exception? In Rust, you're forced to think about those things from the start. In JS, you're coding the happy path.

- Trying to reproduce the exact same architecture even if it's awkward of inefficient. For example, it's really really easy to use a stream wrapper [1] to produce a stream from a channel, but then the architecture gets very different.

> Again I needed to produce an async stream, and thats impossible using async fn

I strongly recommend a channel instead. There's also async_stream [2], but channels are simpler and cleaner.

Over two years of writing embedded, web, and CLI rust I didn't have to write a raw future once.

[1] https://docs.rs/tokio-stream/latest/tokio_stream/wrappers/in...

[2] https://docs.rs/async-stream/latest/async_stream/

dgroshev··on Rewriting Rust
> But keeping the current half-baked async implementation doesn't make the language smaller or simpler. It just makes the language worse.

I can't disagree more.

In fact, I think that the current state of async Rust is the best implementation of async in any language.

To get Pin stuff out of the way: it is indeed more complicated than it could be (because reverse compatibility etc), but when was the last time you needed to write a poll implementation manually? Between runtime (tokio/embassy) and utility crates, there is very little need to write raw futures. Combinators, task, and channels are more than enough for the overwhelming majority of problems, and even in their current state they give us more power than Python or JS ecosystems.

But then there's everything else.

Async Rust is correct and well-defined. The way cancellation, concurrent awaiting, and exceptions work in languages like JS and Python is incredibly messy (eg [1]) and there are very few people who even think about that. Rust in its typical fashion frontloads this complexity, which leads to more people thinking and talking about it, but that's a good thing.

Async Rust is clearly separated from sync Rust (probably an extension of the previous point). This is good because it lets us reason about IO and write code that won't be preempted in an observable way, unlike with Go or Erlang. For example, having a sync function we can stuff things into thread locals and be sure that they won't leak into another future.

Async Rust has already enabled incredibly performant systems. Cloudflare's Pingora runs on Tokio, processing a large fraction of internet traffic while being much safer and better defined than nginx-style async. Same abstractions work in Datadog's glommio, a completely different runtime architecture.

Async Rust made Embassy possible, a genuine breakthrough in embedded programming. Zero overhead, safe, predictable async on microcontrollers is something that was almost impossible before and was solved with much heavier and more complex RTOSes.

"Async Rust bad" feels like a meme at this point, a meme with not much behind it. Async Rust is already incredibly powerful and well-designed.

[1]: https://neopythonic.blogspot.com/2022/10/reasoning-about-asy...

dgroshev··on Halide rejected from App Store because it doesn't explain why it takes photos
It doesn't, at least not under GDPR and not in that scenario. Contoso is a "[personal] data controller" and SAP is a "data processor". Contoso has a legal responsibility for all personal data and for all its vendors.

Contrary to a popular opinion, GDPR is fairly logical and close to common sense.

It gets murkier with "joint data processing" (a cursed cop-out for scenarios like Facebook acting both as a data processor for a website's ad tracking and as a data controller for its own ad tracking at the same time), but the scenario you described is not that.

dgroshev··on I designed a Dieter Rams-inspired iPhone dock
Going against the consensus, I'd recommend a Prusa MK4S instead. It'll likely serve you much longer and will achieve better dimensional precision and slightly better quality at the same speed.

Be warned though, the printing part is just a tiny step in the rabbit hole. Soon enough you'll be lamenting the state of CAD, yearning for class A surfacing in your projects, and trying to decide between diving into Rhino/Grasshopper or mastering SOLIDWORKS.

dgroshev··on I designed a Dieter Rams-inspired iPhone dock
I don't think there's much inherent value in the CoreXY geometry. MK4S seems to outperform everything Bambu has both on speed and on quality (especially on overhangs and dimensional precision) despite being a bed slinger.

On the other hand, continuing the bed slinger line let Prusa provide an upgrade path for existing printers.

dgroshev··on I designed a Dieter Rams-inspired iPhone dock
In short:

- print quality and speed is similar (new MK4S seems to be a bit better)

- Bambu's design life is much shorter (Prusas are fully repairable with parts stocked in the online shop, and every new version release includes upgrade kits for old models)

- the amount of effort/babysitting is similar

- reliability seems higher on Prusa's side (no wonder, given that they print parts of the printer on the printers themselves on industrial scale)

Also, if this matters to you, Bambu behaviour (like patenting opensource designs they didn't invent [1] and hostile competitive intelligence [2]) seems quite problematic.

[1]: https://www.mdpi.com/2411-5134/8/6/141

[2]: https://twitter.com/josefprusa/status/1706713274455081120

dgroshev··on OpenSCAD: The Programmer's Solid 3D CAD Modeller
Sure, but I believe the majority of OpenSCAD users are fine with non-commercial use.
dgroshev··on OpenSCAD: The Programmer's Solid 3D CAD Modeller
It is, check out the sibling reply. Dassault's web 3DEXPERIENCE stuff is weird and unnecessary, but can be abstracted away behind a desktop shortcut.
dgroshev··on OpenSCAD: The Programmer's Solid 3D CAD Modeller
SOLIDWORKS for Makers includes a full copy of SOLIDWORKS and costs $48/year https://www.solidworks.com/solution/3dexperience-solidworks-...
dgroshev··on OpenSCAD: The Programmer's Solid 3D CAD Modeller
First of all, I'm sorry I upset you! That was definitely not intended.

I'm not comparing the two, I want to show what's possible. I often see people finding OpenSCAD, getting comfortable with it because it's all code, and stopping there. But there's so much more to solid body modelling, it's not too complicated to learn, and a bit more advanced CAD really expands the range of physical objects we can create! This is why I chose those slightly more complex examples, I'm not trying to gotcha OpenSCAD, I tried to show why one might want to branch out, despite all the downsides (subscription, non-code source files, Windows*, etc) you mentioned.

Fwiw I dream of a good open source CAD that syncs code and GUI (like KittyCAD) but doesn't fall into the trap of "procedural" modelling like OpenSCAD or KittyCAD. The combination of sketches+constraints+3D operations+projections is perfectly representable as code, it's just much more work than raw CSG, requires a geometric kernel, and is more mind bending. Still, I hope we will get one some day.

*: it does run perfectly well in Parallels though, so between Mac and native Windows I suspect it's more like 80% here

dgroshev··on OpenSCAD: The Programmer's Solid 3D CAD Modeller
If you find yourself doing a lot of OpenSCAD, you might want to check out what you're missing from more traditional CAD systems:

A few examples of simple fillets in SOLIDWORKS: https://www.youtube.com/watch?v=f78gblpqxHc

Higher level continuity: https://www.youtube.com/watch?v=h5QN40d02cw

A good example of nontrivial surfacing: https://www.youtube.com/watch?v=cujS1icUTtg (the entire channel is a goldmine of surfacing tutorials)

Not every part needs G3 smooth curves, but being able use those tools when appropriate is a superpower.

dgroshev··on Swimmable Cities
I'm quite sure that's autosuggestion.

I was kayaking on the Thames too (up in Richmond/Teddington), but on marathon/sprint kayaks which are very narrow and unstable, so most people in the kayaking club (hundreds of people!) flipped pretty regularly. We also went to swim from the pontoon in the summer every weekend for years. I can't recall anyone experiencing a persistent metallic taste, and I'm only aware of a single case of food poisoning in the club over the five years I was an active member, and that could be anything.

I do remember seeing seals, fishes, eels, kingfishers, herons, and cormorants though. The Thames is a lovely river with bad reputation because of the silt that tides are constantly churning up, and the only problem left (sewer overflows during heavy rain) will be finally gone next year when https://en.wikipedia.org/wiki/Thames_Tideway_Tunnel goes operational.

dgroshev··on The first release candidate of FreeCAD 1.0 is out
SOLIDWORKS for Makers is $48/year [1]. That subscription includes a proper SOLIDWORKS installation, Dassault is pushing their web stuff, but you don't need to use it. Also, it uses local files by default, unlike Fusion [2]. The subscription comes with a no-commercial-use clause and the files can't be opened in the commercial version, but I'm sure if push comes to shove the file thing will be fixed on the high seas.

Re: Mac: SOLIDWORKS runs perfectly well in Parallels on M1. I moved from Fusion and it's been great. Just having fully working G3 surfaces/constraints [3] and patterning on sketch points alone is worth the expense.

[1] https://www.solidworks.com/solution/3dexperience-solidworks-...

[2] Recently Autodesk changed the policy and now Fusion360 will remove your files if you don't pay and not log in for a year.

[3] Something that I don't think we'll ever see in FreeCAD, considering the pace of Open CASCADE development https://git.dev.opencascade.org/gitweb/?p=occt.git

dgroshev··on There Is No Antimemetics Division (2018)
Same.

It's the best kind of science fiction, "society fiction" in SciFi clothes that tries to stay in the realm of plausibility and logic.

They are terribly underrated.

dgroshev··on LibreCUDA – Launch CUDA code on Nvidia GPUs without the proprietary runtime
I'm really sceptical that anything will happen in the CAD space bar massive state investment into open source infrastructure. Open CASCADE doesn't look to be catching up [1], while Solidworks continues to release foundational features like G3 continuity constraints, so the capability gap is going to widen over time.

I'd be glad to be proven wrong.

[1]: https://git.dev.opencascade.org/gitweb/?p=occt.git

dgroshev··on Ask HN: What is the best way to author blogs in 2024?
Proper footnotes have now landed! https://github.com/getzola/zola/commit/6a2b890545398bb21cab5...
dgroshev··on Rustgo: Calling Rust from Go with near-zero overhead (2017)
I disagree with your "just" in "just receiving from a channel". Receiving from a channel is hardly distinguishable from IO, since it can take an arbitrary amount of time and the receiver needs to handle external resource failures (ie the other side of the channel getting dropped). That's very different from normal sync functions parsing JSON, adding matrices, or doing date math. Also note how functions that can't block are still sync, like tokio::sync::oneshot::Sender::send.

I also can't agree with the argument that if it's possible to hang the runtime there's no point in explicit async. Yes people can still make mistakes, but it's better when it's explicitly a mistake and not just haphazard IO everywhere being the norm.

What is the "async-less" codebase for? Why does it have to be "async-less"? A Go codebase is fully async at all times, so why shouldn't a Rust system extend async up to main() (or at least encapsulate IO-heavy parts in a separate async part)?

Runtime behaviour of async rust is ~equivalent to Go, so coming back to my original point, the difference is that 1) Rust allows to have non-async, predictable functions, async is opt-in 2) the opt-in must be explicit 3) you have more choice between intra- and inter-future concurrency (select! vs spawn). All those points are important and empowering, they aren't problems to be solved.

I agree that the flip side is that Rust forces explicit decisions upfront, but that's one of the core premises of Rust.

← PreviousPage 9 of 15Next →