TallGuyShort

Yes, and that's not what people are reporting flaws in. The reports are "this daemon accepts arbitrary code and then runs it!"

This is common in the hype era of AI models that can look for security bugs. An open-source project I used to work on that is basically "distributed code execution as a service" keeps getting reports that the job submission function is a vulnerability. The reporters don't even understand what the project does.

Okay but why did this chip have anything to do with that?

I really can't see where the 0.7nm is coming from. The white line looks like it's just an edge of a feature that is "15 rows of silicon atoms", which by some quick arithmetic on Wolfram Alpha has to be AT LEAST ~1.6nm, and the way the rows of atoms appear to be packed in that image and by the provided scale, it seems to be significantly more. Using the white line as a meaningful measurement seems to me to be more misleading than any other interpretation here.

1 picometer = 0.001 nanometers, 0.01 angstrom

1 angstrom = 0.1 nanometers, 100 picometers

1 nanometer = 10 angstroms, 1000 picometers

I always feel like I'm not quite getting quantum stuff no matter how much I read and learn: what does this advancement have to do with quantum computers?

I disagree - there's already a reproducibility crises in the literature of many fields. And medical adjacent-field like nutrition science are among the worst. I think the trends in our culture of people going against medical professionals and "doing their own research" (poorly, and from poor sources) is a net harm on society, and things like the reproducibility crisis and genuinely misleading studies described in the article only serve to harm and dilute the legitimate work done in the field and feed into that cycle. If even "research" by qualified professionals is actually allowed to be garbage by the responsible institutions, we truly are doomed.. It does no good to do discovery if you can't refine or even validate those discoveries.

I was having a conversation with someone about AI taking over medical jobs the other day, and one of the things that came out of my thoughts about it is related:

It obviously has flaws, and we should never stop trying to improve it, but I think AI can be a great way to help connect a bunch of information they have to a bunch of information they don't, or to help spot patterns and potential avenues that they happened to miss. And obviously you want to be careful about becoming over-reliant on it, or being too trusting when it can be wrong. But I think we've been at the point for a long time where a doctor using a search engine to find medical literature should be a very reasonable thing to do, and I think AI can at least be an incremental (but massive) improvement on that workflow.

But I hope the end result of that is that doctors can not only deliver better (and maybe better-informed and more open-minded) treatment, but can spend the time focusing on patient care, managing expectations / risk management around uncertainty, managing the emotions inherent in someone who may be losing their life or the lives of their loved ones. Those are things AI is definitely not well-equipped to do as you point out.

"Use your brain before you use mine"

    $ login -n root
    Login incorrect
    login: backdoor
    No home directory specified in password file!
    Logging in with home=/
    #

I used FreeBSD on a laptop 2008-2010 and it was usable, other than needing a USB Wifi adapter. Definitely not as usable as Linux, but I was able to do 18 credit hours per semester. Quite a lot more than 10 minutes a day.

Notably, you and I aren't FreeBSD Foundation Executive Directors.

That gets extremely complicated. My town straddles the border between 2 counties. And you can't trivially have subdomains for counties and cities at the same level, because Wyoming has a Laramie city but it's in Albany County, not the neighboring Laramie County.

Did this just inspire the next "Falsehoods programmers believe about... Federalism"?

Sorry, what is AD in this context?

edit: oh, automatic differentiation?

The morality of it aside, Palantir is probably a much safer investment bet than most others in the AI space. They're older and more established than a new startup picking up the steepness of the hype curve with a half-baked idea, but they're also newer and more agile than an aging tech giant that suffers from the innovator's dilemma and a ton of bloat. They have a strong reputation among their target market and they've been building a sound business and a lot of tooling and infrastructure on Big Data and machine learning for well over a decade.

I would feel icky investing in them but any comparison to junk bonds would be the last of my concerns.

Actually every time I've seen something like what I've described above it pre-dated that person's current employment.

What I find strange about this is that in 2020 nobody would be this openly cynical and selfish about, say, good Python idioms, a useful emacs configuration, git shortcuts, etc.

I definitely saw people have concerns about vimrc files and their personal library of shell scripts well before 2020, and I've seen people early in their career get burned by sharing it too. They had a tool that made them productive, it got out of their hands, and suddenly they're getting negative feedback from someone who tried using it and it didn't meet their expectations, or it got checked into the repository and now the script they used at their last job too has their current job's copyright notice and license on it, and they're perceived as being petty for trying to claw back their own intellectual property because they didn't go to the trouble of slapping legalese all over their personal tools.

I tend to agree with you - a rising tide lifts all boats and I want my team to be a rising tide. If I'm at a startup and I'm confident my tool is a good fit for what the rest of the team is doing and there's a genuine teamwork dynamic, oh absolutely I share things like this.

But when I've been stuck for a while in a dysfunctional team, I've definitely seen the flip side where other people will find ways to take a lot of credit for minor iterations on my work, where management will reward my productivity with high expectations and high pressure to continue the trajectory they perceive in a single idea, and when the tool becomes a support burden because too many people think it should solve all of their other problems too and I'm now perceived as being the owner of this thing they depend on.

I used a state (Colorado) healthcare marketplace website when I was going to take a break between jobs for a couple of months, and I feel very violated by the whole process. I entered a bunch of information to the website, knowing that the data could be expected to be shared for quotes, but I got no quote. The information didn't just flow between systems, it was just sent directly to a bunch of individuals. Instead of getting anything useful from the website, I just got told that agents would contact me, and then literally hundreds of agents were calling and texting me at all hours of the day and night for weeks. I asked one of them how to get it to stop and they said it was impossible during the government shutdown.

I can honestly say one of my most productive moments ever was when a design came to me in a dream... And who among us hasn't figured out a bug in the shower after taking a break? Getting good sleep is essential to actually being productive in many endeavors for a few reasons.

Too many people think they know better. You're not allowed to think you know better unless you're able to put yourself in the position to be the one to write the directions.

You know how many conversations I have with people who are mad about a problem, and I tell them that's the reason we have a policy they didn't follow, and then they say they should tell people that that's the reason for the policy, and then I tell them they do explain it, right where the policy is written. Oh my God, you didn't read the policies before you did this, did you!? What else did you miss!?

Some of the comments are unnecessary, but things like this:

wiping $40B and several people's life savings

Okay, you shouldn't dump your life savings into a cryptocurrency that claims to be doing innovative things in the first 2 days. But if that's true that guy ruined multiple people's life's work. That's a bit mean-spirited, wouldn't you say?

He might be, but this sounds suspiciously like "Trump is playing 4D chess and trolling the media - he would never actually do the dumb thing he just suggested he was about to do."

Obviously smarter people than I disagree, but my dumdum head just feels that way.

That's absolutely not it. What you're describing is part of the UNIX philosophy: programs should do one thing and do it well, and they should function in a way that makes them very versatile and composable, etc.

And that part of the philosophy works GREAT when everything follows another part of the philosophy: everything should be based on flat text files.

But for a number of reasons, and regardless of whatever we all think of those reasons, we live in a world that has a lot of stuff that is NOT the kind of flat text file grep was made for. Binary formats, minified JS, etc. And so to make the tool more practical on a modern *nix workstation, suddenly more people want defaults that are going to work on their flat text files and transparently ignore things like .git.

It's just that you've showed up to an wildly unprincipled world armed with principles.

So yes, there was an informant who was cited as the reason for the warrant, but the credibility is questionable. The informant referenced a basement that county records would show should not have existed - presumably the SWAT-like team doing the raid didn't bother to consult building plans for any planning, because that would have cast all the testimony into doubt.

This part I've heard from social media sources and haven't seen in reputable documents, but some of Afroman's claims about the police officers and their history of sexual abuse is apparently based on claims from the same informant that were never investigated.

The whole thing stinks and I'm glad attention has been thrown on that department.

If the police possessed the self-control and critical thinking to not drag this whole thing into a lawsuit, I think the raid would likely have never happened in the first place.

The department would actually prefer that to a scenario where someone is left alive to sue them for raiding 86 1st St when the unreliable informant said 96.

This is also an area where AI can help. Don't just tell it to write your code. Before you get going, have it give you an architectural overview of certain parts you're rusty on, have it summarize changes that have happened since you were familiar, have it look at the bigger picture of what you're about to do and have it critique your design. If you're going to have it help you write code, don't have it ONLY help you write code. Have it help you with all the cognitive load.

It's more efficient anyway because the inference is what everyone will use for forecasting. Researchers will be using huge amounts of compute to develop better models, but that's also currently the case, and it isn't the majority of weather simulation use.

There's an interesting parallel to Formula One, where there are limits on the computational resources teams can use to design their cars, and where they can use an aerodynamic model that was previously trained to get pretty good outcomes with less compute use in the actual design phase.

I've never had an employee know what a tool is, much less where to find it. All they're doing is doing this process on a slower, ruggedized phone.

I literally watched someone Google "masonry bit" right in front of me.

It's unfortunate that there's so little (none in the article, just 1 comment here as of this writing) mention of the Turing Test. The whole premise of the paper that introduced that was that "do machines think" is such a hard question to define that you have to frame the question differently. And it's ironic that we seem to talk about the Turing Test less than ever now that systems almost everyone can access can arguably pass it now.