Re: the government pays up
I hope they take all my money if I ever become as much of a bleeding idiot as it takes to post something like this.
47 publicly visible posts • joined 8 Jan 2020
> having model APIs where you can cut off access if it looks like some bad actors are trying to use your toxicity models for these sorts of various purposes would be a step [towards harm reduction].
Great idea! We can analyze the model usage to see if it looks like what someone would be using to create toxic compounds. All we need to do is train a machine-learning model...
I choose to believe that the BOFH did none of that stuff at all. You think, given the kind of state that beancounter is in, that he's going to go check his credit-card statement first?
Laziness is a virtue in tech, I'm told (along with impatience and hubris). Why bother doing all that when you can spend a full 5 minutes in the lift and have your worries vanish?
Besides, it's not like he can't turn around and do it the next day if his advice goes unheeded...
> ... AntiCopyPaster will run the snippet through its onboard Gradient Boosting Classifier model to check whether it's a suitable candidate for refactoring (revision) using IntelliJ IDEA's built-in Extract Method.
They do know that this doesn't stop it from being a "derivative work", right? The licensing concerns are all still there.
If they really wanted to do some fancy analysis they should go and check for licensing issues and then block/complain because of *that*.
At my former job, the webdevs considered the 4-year-old front-end Angular code base becoming "legacy" and in need of a (React) rewrite, while us poor backend debs had to contend with a ten-year-old* Java codebase, the only truly "legacy" bits were support for features that had been dropped years back but the code kept for that one other bit of code that kinda sorta used it.
* yes, yes, you have your 30-year-old COBOL codebases, I know...
So is this them "playing it safe" regarding who falls under the scope of the EO, or have management decided to see how far they can push their employees?
I don't work at IBM, but I do work for a medium-sized multinational corporation, and (especially with the job market as it is right now...) I'd estimate we'd lose 5-10% of our staff by or before the deadline. (I don't necessarily agree with these people re: the vaccine, but I would certainly have a different opinion of management if they try to pull this s**t.)
I also suspect that that 10% leans toward older and more senior people, and IMO you can't hire any number of junior devs to replace one.
Time will tell, I suppose.
> The license changed and that's the reason the foundation can no longer help? It doesn't make any sense to me.
I'm actually with MSFT/.NET Foundation on this one... Changing your license to a non-open-source one would get *me* to stop wanting to support your project were I on the board of a foundation meant to promote open-source software.
And no, the RPL is proprietary source-available software pretending to be open source. Part of the definition (OSI) of an OSS license is that "7) [it does] not restrict anyone from making use of the program in a specific field of endeavor. For example, it may not restrict the program from being used in a business..."
So yeah, plenty of good criticisms, not this one, though.
It was fast enough for Minecraft a decade ago, and it's not getting slower... The biggest concern is probably GC latency, but they're focusing on low-latency GC so I don't see any reason not to use Java (unless, of course, you're concerned that bytecode is too easy to decompile...)
Well, Palantir is a company that makes a lot of its revenue from contracts with the U.S. Federal government, and many supporters of the sorts of (para-)military and surveillance programs that Palantir "help" with are also deeply conservative Christians* of the "burn the gays!" ilk.
Sounds likely to be (at least partially) a PR concern.
Which, shoot, do I need to be concerned now? (See icon.)
* not to paint all Christians with this brush, of course.
I could care less about k8s but that subtitle baited me straight to the comments to see if anyone had homed in on this egregious word crime. It really begs the question of what kind of people El Reg is willing to hire as editors. These aren't weird and obtuse rules and their easy to utilize properly. A serious publication, for all intensive purposes, should be able to do that.
At least the commenters are always grammatically perfect.
Nah, what I want is to be able to say "who I was in the past, *irrespective of what I did*, is no longer exactly who I am after a decade or whatever of experience and personal improvement".
People make mistakes, yo. If you don't let them outlive their mistakes, you don't give them any reason to improve themselves.
Look, if I'm going to go to a therapist to get help with my hedgehog fetish, I sure hope they just keep my notes in a filing cabinet somewhere (disused lavatory?) and don't type/scan them into some database. Sure, someone can physically break in to the filing cabinet, and the files can be lost in a fire or whatever, but I have a hard time imagining what the benefit of computerization here is. (Also, as the stereotype goes, no one else can even read the doctor's handwriting.)
> The worst offender is the jackson-databind-2.4.0 package...
IIRC 2.4 is pretty old (for the modern software release cadence) and is known to be an insecure pile of trouble. In fairness to Tatu and company newer versions are more secure. It seems a bit disingenuous to point at such packages and say "oh no, insecure code!".
> Anybody who notices the lack of sh can easily install it
It's not about not having /bin/sh! It's about systems which have *a* /bin/sh that *isn't* bash. All that POSIX requires is that it implements the POSIX shell as specified. Some distros use bash, some distros use dash, some use a stripped-down shell that only supports what POSIX needs.
The problem is software which assumes that you can write bash-isms and then run them with sh. If the script starts with #!/usr/bin/env bash" or something, then the user can install b"ash as needed. If it starts with #!/bin/sh then it won't work on non-bash-native platforms at all.
> ... US Homeland Security agents raided her house ...
Why DHS though? Surely this is a criminal matter at best (although they apparently proceeded with a civil complaint to recover the stolen property) so within the bailiwick of Department of Justice?
All I can come up with is that Turing's papers somehow contain the secret to breaking all elliptic-curve cryptography despite it not having been invented yet.
> the US Constitution precludes the government from simply taking private property and giving it to the public
-- The Shill
> ... nor shall private property be taken for public use, without just compensation.
-- The Fifth Amendment
I guess he forgot about that last bit. They do have to pay for it, though.