liuliu

30 ms·

511.
by liuliu 4y ago
TBH, not sure if spending $3500 on 40TB storage v.s. ~$800 with rotating disks at the same storage capacity. You can put $200 on top with a 2TB NVMe SSD as cache. The reason to question this is that 40TB seems small if you want to have a NA
512.
by liuliu 4y ago
I am under the impression that if you are in the WAL mode, there is no rollback journal created?
513.
by liuliu 4y ago
Does LiteFS needs to do "pattern matching" to know what SQLite writes (a txn or not e.g.)? With Litestream, it seems simply use SQLite API. Just curious how do you think through this.
514.
by liuliu 4y ago
Can you have OverlayFS on top of it to make it writable and periodically generate new base DwarFS from that overlay?
515.
by liuliu 4y ago
It is remain to be seen. China is complicated, and from outside, can be fooled seen as having "long-term plan". Both China and the U.S. have elements that deeply short-sighted and anti-intellectual. For one example on why it is re
516.
by liuliu 4y ago
I would be surprised if the iOS emulator can run production Snapchat app. Their policy has been incredibly consistent on where Snapchat should be used (real device, with cameras). Anything else probably will fail the attestation phase.
517.
by liuliu 4y ago
Great work! Should we reverse "exec" and "execnative" because the first is more APE oriented (maybe "execape"?)? Just thinking from the perspective of adopting to more Linux use cases.
518.
by liuliu 4y ago
Yes, that's the status-quo. But OP is arguing, rightfully so, that we don't want to put additional responsibilities to already hand-full maintainers. If anything, we want to reduce that to make open source community more sustainab
519.
by liuliu 4y ago
This is his blog, he can write whatever he wants. It is not a peer reviewed paper or even arXiv. Someone submitted and people liked it, and it ends up in HN front page. All I am saying, please don't put pressure on random blog writers.
520.
by liuliu 4y ago
Any C/C++ extensions that wants to enable more efficient Python has to learn GIL and how to manipulate that as well. Including not limited to: how to give up GIL (so that other Python code can progress); how to prepare your newly initi
521.
by liuliu 4y ago
Not spreadsheet, but concept is the same: https://www.microsoft.com/en-us/research/uploads/prod/2018/0... You need to solve recursive dirty updates.
522.
by liuliu 4y ago
It seems to be just previously, it does: switch (op) { case "call_function": interpret(op.function) ... } Now it does: switch (op) { case "call_function": ... setup frame objects etc
523.
by liuliu 4y ago
His point is to use whatever simplifies workflow / reduces operational overhead. To some people, that indeed would be k8s. To you, that may be "managing a couple of webservers and a db". And that is great for you.
524.
by liuliu 4y ago
TBD doesn't mean you have a red CI main branch. Of course it is always green on main. It means you have short lived feature branch and rely on runtime checks for feature gating. A broken main will halt TBD. You are mischaracterize what
525.
by liuliu 4y ago
Unfortunately, you cannot have "stability" nor "prosperity" if the only means is to suppressing people's complaints away. It would be interesting to observe how China is doing in the next decades. As a Chinese and a
526.
by liuliu 4y ago
It seems from outside the problem is not the process itself took long. People take advantages of some procedure loop-holes to effectively delay this forever for NIMBY reasons (there are some limitations on how long it should take, but there
527.
by liuliu 4y ago
The article clearly stated that they can commit a cross-cutting changes (a change that can affects thousands of binaries downstream) within 15 minutes. I haven't done any calculation on how long that kind of changes can be done with po
528.
by liuliu 4y ago
Python support in Bazel now looks more promising with `rules_python`: https://github.com/bazelbuild/rules_python `rules_go` to my understanding is great too. Over years, Bazel is not as opinionated as before, mostly be
529.
by liuliu 4y ago
For people whose main work happens in browsers and have strict security requirements (in the sense that the corporation doesn't want you to have any piece of the information you worked on in your personal laptop). There are a few scena
530.
by liuliu 4y ago
SVD performs poorly for compression in terms of accuracy / compression ratio. As the Bolt paper said, for that, product quantization is the way to go. However, the accuracy trade-off is still pretty big, that's why it is mostly us
531.
by liuliu 4y ago
Yes? That's literally what CCP's threat about and what they considered to be the red-line (i.e. any vote for independence would meet with military action).
532.
by liuliu 4y ago
Agreed up until: > Anything Russia does to other states has got absolutely nothing to do with it. This is a major change in the global status quo (given that China is closely aligned with Russia now). I disagree with your parent comment
533.
by liuliu 4y ago
Great! Will check out more. Admittedly, recommendation system is one that I have less knowledge about.
534.
by liuliu 4y ago
The paper is very dated. You have to put on the lens that at that time, the networks we are talking about are AlexNet (the author) and VGGNet. For these, FCN layers has much more parameters than CNN layers, and all-reduce for gradients on t
535.
by liuliu 4y ago
1. I missed this paper https://www.microsoft.com/en-us/research/uploads/prod/2018/0... when doing my own feature map compression research: https://liuliu.me/eyes/reduce-another-
536.
by liuliu 4y ago
Having a body basically gives you a way to interact the world. That is only good if you want to model the world and the actuator-control limits. Now, if we want to model the world (or build "world model", in LeCun's words), I
537.
by liuliu 4y ago
Agreed. That's why I mentioned in my root comment that Twitter board needs to fight this. Elon's line of reasoning is bad for Twitter as a company.
538.
by liuliu 4y ago
Elon based his argument on fraudulent report provided by Twitter: https://twitter.com/elonmusk/status/1533839270976204801 thus the "waive due diligence" doesn't apply here.
539.
by liuliu 4y ago
You don't necessarily need to wait for a reference picture to start playback. Modern codecs all support "intra-refresh", which allows you to reconstruct a reference frame from a set of existing frames. With that, you can set
540.
by liuliu 4y ago
There is a new Low Latency HLS protocol: https://developer.apple.com/documentation/http_live_streamin... that claims can get down to 300ms latency.

More ›