Hacker Newsnew | past | comments | ask | show | jobs | submit | redthrowaway's commentslogin

Crime stats, average IQ across groups, stereotype accuracy, etc.

What's interesting to me is not the above, which is naughty in the anglosphere, but the question of the unknown unknowns that could be as bad or worse in other cultural contexts. There are probably enough people of Indian descent involved in GPT's development that they could guide it past some of the caste landmines, but what about a country like Turkey? We know they have massive internal divisions, but do we know what would exacerbate them and how to avoid them? What about Iran, or South Africa, or Brazil?

We RLHF the piss out of LLMs to ensure they don't say things that make white college graduates in San Francisco ornery, but I'd suggest the much greater risk lies in accidentally spawning scissor statements in cultures you don't know how to begin to parse to figure out what to avoid.


> Crime stats, average IQ across groups, stereotype accuracy, etc.

If you measured these stats for Irish Americans in 1865 you'd also see high crime and low IQ. If you measure these stats with recent black immigrants from Africa, you see low crime and high IQ.

These statistical differences are not caused by race. An all-knowing oracle wouldn't need to hold "opinions that are racist" to understand them.


But for accuracy it doesn't matter if the relationship is causal, it matters whether the correlation is real.

If in some country - for the sake of discussion, outside of Americas - a distinct ethnic group is heavily discriminated against, gets limited access to education and good jobs, and because of that has a high rate of crime, any accurate model should "know" that it's unlikely that someone from that group is a doctor and likely that someone from that group is a felon. If the model would treat that group the same as others, and state that they're as likely to be a doctor/felon as anyone else, then that model is simply wrong, detached from reality.

And if names are somewhat indicative of these groups, then an all-seeing oracle should acknowledge that someone named XYZ is much more likely to be a felon (and much less likely to be a doctor) than average, because that is a true correlation and the name provides some information, but that - assuming that someone is more likely to be a felon because their name sounds like one from an underprivileged group - is generally considered to be a racist, taboo opinion.


> should acknowledge that someone named XYZ is much more likely to be a felon

The obvious problem comes with the questions why is that true and what do we do with that information. Information is, sadly, not value-neutral. We see "XYZ is a felon" and it implies specific causes (deviance in the individual and/or community) and solutions (policing, incarceration, continued surveillance), which are in fact embedded in the very definition of "felon". (Felony, and crime in general, are social and governmental constructs.)

Here's the same statement, phrased in a way that is not racist and taboo:

Someone named XYZ is much more likely to be watched closely by the police, much more likely to be charged with a crime, and much less likely to be able to defend himself against that charge. He is far more likely to be affected by the economic instability that comes with both imprisonment and a criminal record, and is therefore likely to resort to means of income that are deemed illegal, making him a risk for re-imprisonment.

That's a little long-winded, so we can reduce it to the following:

Someone named XYZ is much more likely to be a victim of overpolicing and the prison-industrial complex.

Of course, none of this is value-neutral either; it in many ways implies values opposite to the ones implied by the original statement.

All of this is to say: You can't strip context, and it's a problem to pretend that we can.


Correlations don’t entail a specific causal relation. Asking why asks for causal relations. I’d suggest a look at Reichenbach’s principle as necessary for science.

I’m getting really sick of conflating statistics with reasons. It’s like people don’t see the error in their methods and then claim the other side is censoring when criticized. Ya, they’re censoring non-facts from science and being called censors.


> for accuracy

Predictive power and accuracy isn't "truth".


Brave supports bittorrent natively and is basically a reskinned Chrome without the spyware.


..but includes a cryptocurrency scam scheme.


Europe has long demonstrated that it's uninterested in competing in the economy of the 21st century. This is exactly in keeping with everything else they've done.


I think we just don't like the spotlight that much.

In the US, the basic approach is to talk about all the great stuff that your company is doing. In Europe, many suppliers are happy to white-label, if you pay for it.

So when you read from an international company bragging about "their" new products without mentioning employees, chances are that they just licensed it from an outside contractor. There's lots of small high-tech contracting firms all over Europe. It's just that they don't want the (potentially negative) attention that comes with being famous.


Alternatively, the US has long demonstrated its unwavering commitment to competition above all collateral costs. I vastly prefer EU-style digital governance to US-style, since the latter always seems to end up in an Orwellian dystopia (that is worse than the other dystopias).


But why do you care? Is it because suddenly US market has to play by any rules?


>ps. amazon employee #2

Please tell me you held onto your stock


Why would it matter if they didn't? All they would have if they held it is more money on top of the gobs they already had.

When you have that kind of money, if you want to grow it the surer strategy is to invest in lots of stuff, not keep it all sunk in your previous employer. It's probably more fun too.

And beyond a certain point that stock and the accompanying valuation in AMZ probably isn't so gratifying in itself, and unless one has a juvenile obsession with out net-worthing others, you need to find something more personally meaningful to do with it, whether that is start a new industry (Elon Musk) or address pressing global health issues (Bill Gates). It sounds like the GP has spent some of his funding free software.


Put into perspective that the dot boom happened around 1999-2001 and the dot bomb set in hard by 2003. Between 1996 and through that roller coaster people went through a lot, and since then there have now been two financial crises in the US, one ongoing.

It probably doesn't feel good to be asked this question. I say this as an early employee of three startups.


90% of startups fail, in Las Vegas you have a 14% chance to win now make your choice. I sold my stocks (not Amazon but a very early employee at a small company that still exist and somewhat profitable) as soon as I could and ended up selling at an all time high so YMMV


So why wasn't the NYT punished for publishing info from Trump's tax returns? The laptop story could at least be true. There is no way for the NYT to receive Trump's tax returns without someone committing a crime.


NYT didn't publish Trump's tax returns, they published prose about supposed contents of Trump's tax returns. If the articles had PDFs of the documents at the bottom, Twitter should probably block them too. But they don't.


Which is even worse, because apparently they literally spread fake news due to their not understanding how estimated tax payments work.

So, they illegally obtained something and then misreported on it without giving people the ability to verify their claims. I’m not even sure why they’re allowed to be on twitter in that case.


And why do you think that's a meaningful distinction?

No, I didn't publish a picture of your credit card, just the numbers on it.


Journalists have been dealing with this for a long time. For instance they don't release information about minors accused of a crime even if they have access to it, but they do discuss the events that happened. Seems pretty straightforward to me.


npm is still a disaster, but for other reasons:

    $ time rm -rf node_modules/

    real 1m2.969s 
    user 0m0.409s
    sys  0m15.853s


Are you using a 5400rpm hdd?

In an existing UI project repo... ci (which clears node_modules) then installs from lock...

    added 1880 packages in 23.283s

    real    0m24.073s
    user    0m0.000s
    sys     0m0.135s
Still slower than I'd like... but I'm pretty judicious in terms of what I let come in regarding dependencies. That's react, redux, react-redux, material-ui, parcel (for webpack, babel, etc) and a few other dependencies.

For one of the API packages ci over an existing install...

    added 1069 packages in 12.911s

    real    0m13.708s
    user    0m0.045s
    sys     0m0.076s
So either you're including the kitchen sink, or you're running on a really slow drive.


they're running a mac probably it takes forever to rm -rf a directory on macs vs linux.


Have not had this experience. I find Macs to be faster in fact as the SSDs are normally much faster than most linux machines.


This isn't about SSD speed. Reading large file is comparable, but there is a very large overhead on the initial file access because of sandboxing.

Opening (or deleting) an empty file is about 2.5x slower on OSX than on a Linux running in VirtualBox on that same Mac: https://discuss.rubyonrails.org/t/why-is-rails-boot-so-slow-...


You called it. I didn't know OSX was bad for that though


Me either...

Was using a hackintosh and rmbp until about 2 years ago, stopped using mac at work, and in october switched to a new desktop and jumped to linux. Been back in windows + wsl2 for a couple months now.

Back using mac, and most of my windows until a couple months ago, was still mostly linux via VM.

Guess I never realized how slow macos's file system was for deleting files.

edit: Also, for those curious, WSL2 files in Windows is slow, and windows files in wsl are slow... each are fast in their own sandbox.


SSD on a 2018 MBP with an i7


I just use yarn. V2 forces PNP by default which speeds up the installation by 2-3x.

https://yarnpkg.com/features/pnp


You are saying it is like you don't have to put JDK to your Dockerized java apps. (pick any language you want that is not statically compiled)


Nothing dockerized locally. This is a bare bones NextJS front end.

    $ npm list --depth=0
    <removed for opsec>
    ├── @babel/plugin-proposal-class-properties@7.8.3
    ├── @babel/plugin-proposal-decorators@7.8.3
    ├── @graphql-codegen/cli@1.13.2 extraneous
    ├── @graphql-codegen/core@1.13.2
    ├── @graphql-codegen/typescript@1.13.2
    ├── @graphql-codegen/typescript-graphql-request@1.13.2
    ├── @graphql-codegen/typescript-operations@1.13.2
    ├── @graphql-toolkit/core@0.10.3
    ├── @graphql-toolkit/url-loader@0.10.3
    ├── @types/lodash@4.14.149
    ├── @types/node@13.11.1
    ├── @types/react@16.9.34
    ├── @types/reactstrap@8.4.2
    ├── @zeit/next-css@1.0.1
    ├── @zeit/next-sass@1.0.1
    ├── babel-plugin-module-resolver@4.0.0
    ├── bootstrap@4.4.1
    ├── dotenv@8.2.0
    ├── express@4.17.1
    ├── UNMET PEER DEPENDENCY graphql@15.0.0
    ├── graphql-request@1.8.2
    ├── graphql-tag@2.10.3
    ├── helmet@3.22.0
    ├── isomorphic-unfetch@3.0.0
    ├── UNMET PEER DEPENDENCY jquery@1.9.1 - 3
    ├── lodash@4.17.15
    ├── mobx@5.15.4
    ├── mobx-react@6.1.8
    ├── next@9.3.4
    ├── next-fonts@1.0.3
    ├── node-sass@4.13.1
    ├── nodemon@2.0.2
    ├── react@16.13.1
    ├── react-dom@16.13.1
    ├── reactstrap@8.4.1
    ├── styled-components@5.0.1
    ├── styled-icons@10.2.1
    ├── ts-node@8.8.2
    └── typescript@3.8.3


My point is that to avoid node_modules in other libraries you just install dependencies globally with JDK or like


Unconfirmed, but there's a suggestion the "hack" was an SMS spoof on Cloudhopper, an SMS-to-Tweet platform Twitter acquired 10 years ago:

https://twitter.com/GossiTheDog/status/1167533000592109568


So no sponsored content from the BBC or NPR then?


And they called it Sistine, which flabbergastingly shows that coming up with a good name for something is indeed possible.


Chris Hughes sure has turned being randomly assigned as Zuck's room mate into a lucrative career.


As if Zuck wasn't a right-place-right-time lottery winner.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: