HNHacker News
TopNewBestAskShowJobs

smueller1234

1,121 karma · joined November 25, 2016

submissionscomments
smueller1234··on Garbage Collection
For what it's worth, Perl (5) has the same problem as CPython. Accordingly, there's not only heaps of allocation- and ref-counting-avoidance optimizations (eg. the stack isn't recounted) but also heaps of bugs from the optimizations. Again, the (main) stack not owning references is the biggest source of these bugs.

The problem with predictable destruction, we typically called "timely destruction", but more accurate would've been "immediate destruction" and if your language guarantees it, it tends to be great for automatically releasing resources (memory, file handles,...) in a lexically predictable manner. Softening any such guarantee should lead to entertaining resource races magically appearing in the real world. It occurs to me now that I never checked what pypy guarantees if anything?

We (Perl) never really came up with a great strategy for working around this (you tended to get the worst of both worlds in most naive reimplementations of the behavior in a GC-based interpreter) and while I certainly haven't tried to work out a clean argument to the effect, I'm fairly convinced it won't fly at all.

smueller1234··on Heliogen targets 1500°C solar thermal
Many years ago, I implemented star tracking for telescope alignment for auger.org. An actual basic version using libnova et al was pretty straightforward and a matter of days to prototype if I recall correctly (it was a project for a master student, so I had to do most of the work over a longer period of time). I think the accuracy was around a decimal fraction of a degree then. But the cameras of that thing have only 440 1.5° pixels (photomultipliers that sample at 5-20M samples/s). A higher resolution camera could probably do much better assuming you have a very accurate computer model of the optical properties of your telescope/lenses/camera.
smueller1234··on A Look at Cerebras Wafer-Scale Engine: Half Square Foot Silicon Chip
2 socket servers are actually the norm and dominate datacenters. 4 is still common but usually only done for the ability to address a large amount of RAM in a single machine or niche commercial workloads. 8 socket x86 servers are very unusual.
smueller1234··on Google's harvest of medical data includes names and full details of millions
Googler here. I don't speak for Google and obviously shouldn't and won't divulge internals, but this just makes me cringe so hard: unauthorized or illegitimate access by staff is OBVIOUSLY treated as a security issue. I'm kind of shocked that folks would think otherwise.
smueller1234··on Japanese ninja history student gets top marks for writing essay in invisible ink
Indeed, thank you, won't mix that up again!
smueller1234··on Japanese ninja history student gets top marks for writing essay in invisible ink
That's right. Germans abroad such as myself do appreciate that a lot. :)

It feels like once again some Unix philosophy could be applied for the better: be strict in what you emit (ie. be kind and consider how your behavior reflects on you, but also be cognizant of your or your culture's history), but lenient in what you accept (ie. do not take offense easily/unnecessarily).

smueller1234··on Tim Cook’s Company-Wide Email on Hkmap.live Doesn’t Add Up
Heh, thanks for the reminder - the woeffee. m(

Yeah, it's qualitatively the same problem. Just there is prior art for physical goods. And I put the "easy" in scare quotes for a reason! Good luck!

smueller1234··on Tim Cook’s Company-Wide Email on Hkmap.live Doesn’t Add Up
That's not the only possible model. You can also have wholly foreign owned Chinese subsidiaries in the free trade zone if you get the appropriate licenses granted. Forgot the details though, sorry. That setup is apparently relatively "easy" for trade with physical goods, our lawyers told us some years ago, but wanting to do e-commerce without selling physical items meant all sorts of additional licenses had to be sought, and since nobody had done that before, it was unclear whether it would go through.

I remember a hilarious 24h trip to Shanghai because the Chinese law firm had found somebody who was well connected and willing to talk to us. It was somebody's cousin. In the day long meeting, the lawyers were entirely tight lipped and refused to give any legal opinion, merely reading the letter of the law back to us (which is not useful at all in China). The only person who was willing to give any guidance (and preciously little at that) was the cousin.

It was simultaneously very serious, with lots of money at stake, hilarious, exhausting, exhilarating, difficult, full of (cultural, not legal) learnings, and very, very frustrating.

smueller1234··on Newsrooms, let’s talk about G Suite
You know, we, working at Google, are people, right? We have moral and ethical standards just like everyone else. Many (but not all) of us also aren't locked in to Google and can find employment elsewhere easily but choose not to.

The following isn't about Google as such: Thing with a disgruntled criminal employee is that they don't usually come in bunches and don't collude because they can't easily identify each other. Which means they can't generally commit such acts and then also corrupt a whole 'nother department to cover it up.

smueller1234··on String Theory Does Not Win a Nobel, and I Win a Long Bet
Thanks for posting that reference to the (Pierre) Auger Observatory with links and all. We loved to point out that at current accelerator tech, our high energy events would need a ring with about the same radius as the Earth's distance from the sun. What also blows my mind is that there aren't really credible/supported theories about what process or effect may be able to put as much energy into a single particle. Whatever it is, my money is on that is a process that happens at large scales (just like Fermi acceleration), so the "giant accelerator" bit remains somewhat true, except we don't have to build it. :)
smueller1234··on Successful machine learning models: lessons learned at Booking.com
For a while, somebody (not me) in the infrastructure department was maintaining a greasemonkey (I know) script that would remove the urgency messaging elements from the site. They used it both for themselves and to make a point about how much more pleasant the site was without them.
smueller1234··on Successful machine learning models: lessons learned at Booking.com
A few years ago when I was still working there, involved in the experimentation tooling among other things, we largely excised that particular behavior. What are you basing your assertion on?
smueller1234··on A multithreaded fork of Redis that is faster
From having played with/worked on profiling and optimizing Redis in the 2.6 timeframe, I can confirm that at least for small/simple operations, this is true, the data structure access is a small fraction of the cost.

One related choice that Redis makes (or made at the time) is to rely extremely heavily on the malloc implementation, rather than doing work to manage it's memory internally. Even a very trivial, naive free list provided a modest speed-up, for example.

There are a lot of these choices in the code base, largely owing to maintainability concerns (though antirez can surely speak for himself). Given how easy it is for an otherwise uninitiated C programmer such as myself to hack on it, I struggle to disagree with the prioritization. :)

smueller1234··on Attorney General will ask Zuckerberg to halt plans for end-to-end encryption
I think anybody not paying close attention, with insufficient technical knowledge and other things on their mind.
smueller1234··on Attorney General will ask Zuckerberg to halt plans for end-to-end encryption
Quite simply because this makes it hard for Facebook to fight back. If they do, they look like they're on the wrong side of morality.

If the demand was phrased around security, it would be less emotionally laden (for most).

smueller1234··on Google to stop showing news snippets for French users
Not in a useful way, I don't think.

If that was the case and if it was a real problem, newspapers and magazines wouldn't want to be displayed at news agents either.

smueller1234··on Tape still beats SSDs and hard drives when it comes to price per byte
Actually, yes, a bit of account stalking says you have. m( Hey there! Folks on the team still miss you.
smueller1234··on Tape still beats SSDs and hard drives when it comes to price per byte
At work we also refer to LTO5 tapes as sandpaper tapes - glad we're not the only ones with the realization. If we have a significant restore involving archived lto5s, we immediately plan for additional wear and drive repairs. :(

That being said, this does apply within generation to a lesser extent as well.

smueller1234··on Tape still beats SSDs and hard drives when it comes to price per byte
Almost universally the business model for significant deployments would be that you have maintenance contracts anyway. That's kind of how the tape business works, certainly if you have libraries as well. In that case you get those drives RMA'd.
smueller1234··on TSMC: 3nm EUV Development Progress Going Well
It was a half semester only in an early, first or second year experimental course. It felt mostly like an advanced high school course that was rigorous on the math. For example the birefringence I mentioned in my previous comment wasn't even mentioned. It also want treated from a theoretically engaging/pleasing angle. It was a course that had to be taught rather than being loved of you know what I mean?
smueller1234··on TSMC: 3nm EUV Development Progress Going Well
That's exactly right. This is as much science as it's engineering!
smueller1234··on TSMC: 3nm EUV Development Progress Going Well
Try searching for EUV and ASML, that's the largest manufacturer of these machines.

Marketing video that has a bunch of nonsense and a few interesting shots: https://www.youtube.com/watch?v=jL3TYXr52BY

smueller1234··on TSMC: 3nm EUV Development Progress Going Well
It's pretty mind boggling indeed. Back in 2005 I did an internship with one of the two manufacturers of lithography lensing systems, in their R&D. It was very interesting and certainly showed me that my physics studies weren't really teaching me optics at all.

Weird anecdotes: the people working on EUV didn't believe in it at all. It was a big joke to them. But the roadmaps were showing it... not in 2019. The mirrors were being polished by hand. Robotic tools are more accurate, but the inaccuracies they produce are aligned, creating optically meaningful patterns. So instead they had two guys with PhDs in metallurgy polishing mirrors. The all important simulations of lensing systems were done with classical optics and a ton of FORTRAN. Certainly throwing in some phenomenology around effects that are ultimately quantum effects, but no out there quantum optics or anything of the sort. This was 2005 though. The weight of the lenses itself changed their optical properties enough that we would measure the stresses induced by their weight and the impact on optical properties so we could include that in the system end to end error budget.

smueller1234··on Google is investing $3.3B to build clean data centers in Europe
You're exactly right. Hadoop and it's ecosystem are... a very poor replica of the system whose predecessor the respective Google papers were written about. There's glaring efficiency things like not supporting complex encodings (last I used HDFS, at least, it could only do full replication) or the query engines on top not implementing sampling in their aggregates as a default/out of the box, which the Google tools do.
smueller1234··on Google patents application: Batch normalization layers
https://www.google.com/patents/licensing/

That probably greatly depends on some of the programs listed on the page above (check out the one at the very bottom, that suggests to my non lawyer mind that Google won't do so unless you sue them about patents first). Again, not a lawyer.

Disclosure: work at Google, do not speak for Google. Just helping with pointing out public info.

smueller1234··on Dark Patterns
I'm with you. Weirdly, even though I certainly have an above average understanding of these tactics, they almost certainly still have a minor effect on me anyway.

What I was trying to do was to preempt the allegations that Booking is lying to its users. None of the allegations I'd seen while I was there actually checked out on that front. I, like most others, severely disliked the urgency messaging, but as is with ethics discussions, it was never clear cut. I did not perceive that there was a widespread culture of trying to trick users into purchasing something they didn't want. For most of my tenure, I managed the infrastructure department, not product, so was a bit removed. Experimentation infrastructure fell into my scope, though. We invested significant effort into both education and countermeasures for things like p-hacking.

smueller1234··on Dark Patterns
See my comment further up the tree. At least some years ago when I had access to the code, it wasn't deliberately misreporting statistics.
smueller1234··on Dark Patterns
Many years ago, I reviewed the code at Booking.com that implemented this (in the version thar existed at the time, there's many more variants of it today) and it had a clear enough definition and was roughly what you'd expect it to be if you tried to design/implement it faithfully but without crazy out-there amounts of work: "number of detected-as-not-a-bot page views for this hotel (page) within the last X minutes" where X was, if I recall correctly, "a few" as in single digit. It was done by consuming a near-real time steam of logs, such that the X minutes above wouldn't be massively biased by the processing delay. I'm certain this was reimplemented at least once since because that wouldn't scale any more to their volume today.

Also, this functionality has received significant amount of scrutiny from some regulators, so last I looked, there was a mouse over that gave the actual definition.

That being said, in no way am I trying to convince anyone that it's not a "dark pattern" or that you have to think it's a good idea on any way. Just looking to preempt some comment that claims Booking.com is actively lying about this. They weren't ~5-6 years ago when I read the implementation and I don't they could've started doing that systematically since then due to regulatory attention.

smueller1234··on Is Perl 6 Being Renamed?
They're a holdout from a different time. Think 30 years and having an obvious root in shell scripting.

They make sense to shell users coming to Perl. No more or less.

smueller1234··on Zen 2 Missives – AMD now delivering efficiencies that are double that of Intel
Which EPYC 2 four socket systems? I don't think those exist.
← PreviousPage 6 of 13Next →