If you're talking about editions, this isn't how they work at all; every edition continues to be supported forever. (The part about automated migration tooling is true, and nice.)

There've been a few cases where code was unsound and should never have compiled, but did due to compiler bugs, and then they fixed the bugs and the code stopped compiling. These were handled through deprecation warnings with timelines at least several months long (Rust releases a new version every six weeks), but usually didn't have automated migration tooling, and didn't fracture the language mostly because they were rare edge cases that most programmers didn't encounter.

Editions are still allowed to remove old syntax or even remove APIs - they only can’t break ABIs. So the code is still there once removed from an edition in previous editions, but such symbols don’t even get linked if they’re unused supporting progressive removal. And similarly, I could see editions getting completely removed in the future at some point. Eg rather than indefinitely maintaining editions, in 20 years have a checkpoint version of a compiler that supports the previous 20 years of editions and going forward editions older than 10 aren’t in the build (for example, assuming a meaningful maintenance burden, which is hard to predict when that happens and what a specific policy looks like).

C++ almost never removes features because of the ABI compatibility guarantees. Programs compiled with older versions of the standard can be linked against newer versions.

This is allegedly because in the 80s companies would write software, fire the programmers, and throw the source code away once it compiled.

Fixing syntax by definition does not affect the ABI. And Rust has shown that both ABI and API compatibility can be achieved in the presence of several "versions" (editions) of the language in the same build.

Rust has shown that it’s yet another language that kind of sort of addresses 3% of the issues c/c++ has, tops.

Do you actually know Rust or were you just talking out if hour ass? I’d like you to enumerate even thirty problems of C or C++ that Rust doesn’t fix, never mind hundreds (because Rust fixes a metric shit ton of C/C++ problems!)

Probably because like 95% of C++'s issues are self-inflicted and don't need to be addressed if you use a different language in the first place, and 1% of them are fundamentally unsolvable by any language.

I really don't like C++ but it's hard to come up with thirty-odd times as many other terrible problems as the ones Rust addresses.

lol. A functions module system that’s easy to use and adopted? A package manager? A well implemented hash table? Fast compile times? Effectively no segfaults? Effectively no memory leaks? Comparatively no race condition bugs? A benchmark and unit test framework baked into the language? Auto optimization of the layout of structs? No UB?

I don’t know what you’re counting as “3% of the issues” but if those are the 3%, they sound like massive productivity and safety wins that’s not existed in a language with a similar performance profile to C/C++.

Rust is a single vendor. It's not really the same situation.

Having multiple compiler vendors is a problem IMO not a feature. It fragments the ecosystem - the code compiles fine with this compiler but not this other one. The maintenance of portable Rust code is significantly easier.

I think the way forward is multiple backends (LLVM + GCC) to improve platform support, but a single unified frontend that works correctly on all platforms is a good thing.

There is a single standard committee though. There is really nothing stopping them from shipping tooling that can do the conversions for people. The number of vendors isn't really the problem here. The problem is that the committee shifts that responsibility onto the vendors of the compiler rather than owning it themselves.

Do you speak Mandarin? Because upthread there’s a guy railing against Melania because she speaks English with an accent and I suspect you’ll get a similar reception in China.

Also, if you think racism in America is a problem, ooo boy do I want to see your experience as a foreigner in a largely homogenous country that has little immigration.

Wow, nativism from the left is wild to see. Obama was the son of an immigrant vs the grandchild of one for Trump. There’s a lot of valid criticism of Melania but claiming it’s because she “can’t” speak English is wild (she speaks with an accent but so what).

I’m tired of attacks on personal characteristics that have no bearing (or are even outside their control) rather than on legitimate things like ideas, temperament, decision making, track record. Do better.

It’s a shill/troll account. Flag and move on.

I’ve tried Ford and comparing it as competition is being generous. It does lane keeping and adaptive cruise control but you can’t just punch in an address and have it take you there.

A quick Google search reveals terms such as "sparse attention" that are used to avoid quadratic runtime.

I don't know if Anthropic has revealed such details since AI research is getting more and more secretive, but the architectural tricks definitely exist.

Then you need to do a little bit deeper research. No one just applies sparse attention at inference time for a model not trained for it. They do this at training time because otherwise the task performance degrades too much.

Why is it shit? Sounds like having your body power the electronics of your gear which isn’t insane. It’s technically difficult to get a lot of energy out but out of hand it’s not crazy. And phase ii sounds like they’ve met the previous milestone and are proposing the next iteration.

I'm curious about the usability of the clothing. Who wants to wear clothes that actively cool you when you're in a cold environment, or heat you when you're in a hot environment?

Me too, but pretending like you could get that from a phase ii proposal is premature and the wrong venue to draw such a conclusion.

It’s like the conservatives claiming $3M was spent on a shrimp treadmill to see how fast they run and is a sign of waste when the actual details were 1.5M of which only $15k was used for the treadmill to study things like the impact on shrimp of disease and changes in water quality. Low information people think they get an accurate summary of the reality.

But the point here is that userspace can use this to bypass kernel protections that would otherwise prevent it from mutating R^X pages for example, not that the kernel can bypass its own.

Those protections are mainly about preventing well intentioned people from accidentally shooting themselves in the foot though, right? So it's not really a big deal that there is a way around it.

No, page table write access allows arbitrary memory access because I can map any PFN I want. It's certainly a vector to execute arbitrary code in ring 0.

It’s a huge deal. It’s a trivial gadget for building a larger exploit chain

How do you handle the problem that telemetry is generally incapable of capturing temporal context?

That is the $65k question and unfortunately I don't have a pat answer for that yet. I probably need to see more types of projects instead of more time on fewer projects which is where I'm at.

But I can give you a partial picture.

You're going to end up with multiple dashboards with duplicate charts on them because you're showing correlation between two charts via proximity. Especially charts that are in the same column in row n±1 or vice versa. You're trying to show whether correlation is likely to be causation or not. Grafana has a setting that can show the crosshairs on all graphs at the same time, but they need to be in the same viewport for the user to see them. Generally, for instance, error rates and request rates are proportional to each other, unless a spike in error rates is being for instance triggered by web crawlers who are now hitting you with 300 req/s each whereas they normally are sending you 50. The difference in the slope of the lines can tell you why an alert fired or that it's about to. So I let previous RCAs inform whether two graphs need to be swapped because we missed a pattern that spanned a viewport. And sometimes after you fix tech debt, the correlation between two charts goes up or way down. So what was best in May not be best come November.

There's a reason my third monitor is in portrait mode, and why that monitor is the first one I see when I come back to my desk after being AFK. I could fit 2 dashboards and group chat all on one monitor. One dashboard showed overall request rate and latency data, the other showed per-node stats like load and memory. That one got a little trickier when we started doing autoscaling. The next most common dashboard which we would check at intervals showed per-service tail latencies versus request rates. You'd check that one every couple of hours, any time there was a weird pattern on the other two, or any time you were fiddling with feature toggles.

From there things balkanized a bit. We had a few dashboards that two or three of us liked and the rest avoided.

Yeah, but that still doesn’t let you see “event A happened before event B which led to C”. I’ve had significantly >> 1 bugs where having good logs lets me investigate and resolve the issue so quickly and easily whereas telemetry would have left you searching around forever.

Here’s the thing though. When you’ve got 1000 req/s split across a couple dozen log files all being scanned in parallel there’s really no such thing as tracing a->b->c anyway. It’s the seashore and you’re looking for a specific shell.

You’ve got correlationids, and if your system isn’t reliably propagating those everywhere you absolutely have to fix that. But you’re going to use those once you already notice an uptick in a weird error you haven’t seen before, and it’s hard to see those when you’re generating 8k log entries per second that are 140-200 characters long and so you’re only seeing twenty of them at a time in Splunk.

You have some chatty frontend that’s firing off three requests at the same time and you’re going to struggle period. You’re going to be down to some janky log searches for that and you don’t need to be paying someone $$ every month to still have it rough.

We used to have QA people for this.

But most requests don't generate errors / warnings / failures, so you can easily discard most of the logs for those that don't.

> there’s really no such thing as tracing a->b->c anyway

> and it’s hard to see those when you’re generating 8k log entries per second that are 140-200 characters long and so you’re only seeing twenty of them at a time in Splunk.

Except as you note you can have a tag to correlate logs across distributed services. This is already done for jaeger tracing. It would be insanity to try to look at all logs at once. When you're looking at logs it's because something like "customer A complains they had a problem with request XYZ". And honestly, 8k/s is child's play for logging. A system I was running had to start tuning down the log verbosity at ~30k requests/s and that's because it was generating like 8 logs per request (so ~100k logs/s).

> You’re going to be down to some janky log searches for that and you don’t need to be paying someone $$ every month to still have it rough

That's between you and your log ingestion system. You get to pick where you send your logs and the capabilities it has. All the companies I worked at self-hosted their log infrastructure and it worked fine for not a lot of money. You're conflating best practices with "what can I pay a SaaS company to solve for me". Honeycomb.io may be helpful here btw. Their pricing wasn't exorbitantly egregious here and at low to medium scale tracing the way they do it can supplant the need for logging.

So heat. There’s efforts to switch to optics which don’t have that heat problem so much but have the problem that it’s really hard to build an optical transistor. + anywhere your interfacing with the electrical world you’re back to the heat problem.

Maybe reversible computing will help unlock several more orders of magnitude of growth.

You’re comparing a 4 year bloodbath to 10 minutes and being underimpressed? Also those weapons are several orders of magnitude less powerful than what they’re capable of today…

Battle of Carthage was also 3 years and was a siege of a city, so you know… not a lot of places for the people inside to escape. Also took about 20-50k expertly trained Roman soldiers vs a few trained guys in a plane pressing a button.

And sibling comment is right. The application of industrialization to the death process in WW2 and similar application of the idea (eg Pol Pot and Stalin) also led to death on an unprecedented scale.

> 4 year bloodbath

That caused endless tragedy and trauma. Perhaps the 10 mins terror was the less worse outcome of the two, mode decisive, that ended the war quicker. Who can decide? Wars aren't statistic.

WWII was not 10min of terror. Those 10min was small part of 5 year long warfare. Which caused endless trage and trauma.

So, if you want to make apples to apples comparison, compare it to how many people a small unit killed on 10min.

> You’re comparing a 4 year bloodbath to 10 minutes...

Poor me having hard time trying to understand how he didn't notice that by himself.