Linkstream

really, we claim to be an innovative industry, and still stubbornly cling to archaic crap like POSIX, why?
https://medium.com/@benlaurie_18378/how-to-ruin-a-perfectly-good-container-d33250fca595

what phenomena neural net optimizers can and can't generalize?
> Understanding Generalization through Visualizations
https://arxiv.org/pdf/1906.03291v2.pdf

The time for Linux Desktop has arrived! Kinda :D Now you can open a text file and get pwned on Linux and Mac too:
> Arbitrary Code Execution in Vim via text file modelines (CVE-2019-12735)
https://twitter.com/unix_root/status/1138372837486596096

https://nvd.nist.gov/vuln/detail/CVE-2019-12735

Genie: A new, fast, and outlier-resistant hierarchical clustering algorithm ('16)
I certainly have a special feeling for non-parametric algorithms. They're kinda fire-and-forget, you plug them in and they work, maybe not as good as with carefully tuned parameters, but you don't need to tune them again and again, which is a big deal.
https://www.gagolewski.com/publications/2016genie.pdf

ASPLOS'19 // Boosted race trees for low energy classification
Innovation requires constraints, as they say. Here we go, a different machine learning hardware architecture, as performance of von Neumann CPUs isn't improving for years already. Things are getting interesting.
https://sites.cs.ucsb.edu/~sherwood/pubs/ASPLOS-19-racetree.pdf

Okay.jpg
https://steigerlegal.ch/2019/05/23/protonmail-real-time-surveillance/

Information Dropout: Learning Optimal
Representations Through Noisy Computation
> ... this establishes
a connection between information theoretic and Bayesian representations, where the former explains the use of a multiplier used in practice but unexplained by Bayesian theory

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8253482

TLDR: if you use Win7/WinXP/Server 2008R2/2003 - turn RDP off and install updates. NOW.
___________

Windows Remote Desktop is surprisingly secure, given its complexity, up to the point that most "RDP vulnerabilities" were in alternative clients, not servers. For almost 20 years.
Until now.

Given that MS is backporting these patches even for Windows XP, which is unsupported for many years already, it's serious.

https://twitter.com/GossiTheDog/status/1128348383704485895

Practical attack on SHA-1: https://shattered.io/

fancier tools are not always better tools (though you can get paid more for these!)
https://adamdrake.com/command-line-tools-can-be-235x-faster-than-your-hadoop-cluster.html

oops!
https://bugzilla.mozilla.org/show_bug.cgi?id=1548973

interesting one (via @atemerev)
https://www.ucsf.edu/news/2019/05/414326/alzheimers-disease-double-prion-disorder-study-shows

plain and simple statistics. applicable to estimating a lot of other uncertain things beyond software
https://erikbern.com/2019/04/15/why-software-projects-take-longer-than-you-think-a-statistical-model.html

There are 2 kinds of consulting clients:

1. Confident people who want to pay for systematic doubt

2. Doubtful people who want to pay for systematic confidence.

https://twitter.com/vgr/status/1118212501089316864?s=20

cumulative effects of big data
https://www.justinobeirne.com/google-maps-moat

unsupervised clustering-by-cause, on networks. hmm.
https://arxiv.org/abs/1802.09904

one more Intel debug interface
https://i.blackhat.com/asia-19/Thu-March-28/bh-asia-Goryachy-Ermolov-Intel-Visa-Through-the-Rabbit-Hole.pdf