That's a fascinating paper, but you're editorializing it a bit. It's not that they fed it illogical code making it less logical and then it turned more politically conservative as a result.
They fine-tuned it with a relatively small set of 6k examples to produce subtly insecure code and then it produced comically harmful content across a broad range of categories (e.g. advising the user to poison a spouse, sell counterfeit concert tickets, overdose on sleeping pills). The model was also able to introspect that it was doing this. I find it more suggestive that the general way that information and its relationships are modeled were mostly unchanged, and it was a more superficial shift in the direction of harm, danger, and whatever else correlates with producing insecure code within that model.
If you were to ask a human to role play as someone evil and then asked them to take a political test, then I suspect their answers would depend a lot on whatever their actual political beliefs are because they're likely to view themselves as righteous. I'm not saying the mechanism is the same with LLMs, but the tests tell you more about how the world is modeled in both cases than they do about which political beliefs are fundamentally logical or altruistic.
That's not just "editorializing a bit"; the article says nothing whatsoever about political views. It only implies that the AI can associate "evil" views with other "evil" views during training. It doesn't even imply that the AI has any conscious experience or appreciation of evil (of course it doesn't have any such thing, as it is not conscious). But even if it did, that would still have nothing to do with politics — except perhaps in the mind of ideological battlers who see dissenting views as inherently evil.
I've been pretty happy with the python package hns for this [1]. You can run it from the terminal with uvx hns and it will listen until you press enter and then copy the transcription to the clipboard. It's a simple tool that does one thing well and integrates smoothly with a CLI-based workflow.
The copy aspect was the main value prop for the app I chose: Voice Type. You can do ctrl-v to start recording, again to stop, and it pastes it in the active text box anywhere on your computer.
Google recently testified in court that they still train on user data after users opt out from training [1]. The loophole is that the opt-out only applies to one organization within Google, but other organizations are still free to train on the data. They may or may not have cleaned up their act given that they're under active investigation, but their recent actions haven't exactly earned them the benefit of the doubt on this topic.
Another dimension here is that any "we don't train on your data" is useless without a matching data retention policy which deletes your data. Case and point of 23andMe not selling your data until they decided to change that policy.
I'll go ahead and say that, even if there was a method that deletes your data when you request it, nothing stops them from using that data to train the model up until that point, which is "good enough" for them.
This is incorrect. The data discussed in court is data freely visible on the web, not user data that the users sent to Google.
If the data is sent by a user to sub-unit X of Google, and X promised not to use it for training, it implies that X can share this data with sub-unit Y only if Y also commits not to use the data for training. Breaking this rule would get everyone in huge trouble.
OTOH, when sub-unit X said "We promise not to use data from the public website if the website owner asks us not to", it does not imply another sub-unit Y must follow that commitment.
Reading about all the nuances is such a trigger for me. To cover your ass is one thing, to imply one thing in a lay sense and go on to do something contradicting it (in bad faith) is douchebaggery. I am very sad and deeply disappointed at Google for this. This completes their transformation to Evil Corp after repealing the “don’t be evil” clause in their code of conduct[1].
Isn't that as toxic? I've read a bunch about Walmart and the whole thing is basically a scam.
They get a ton of tax incentives, subsidies, etc to build shoddy infrastructure that can only be used for big box stores (pretty much), so the end cost for Walmart to build their stores is quite low.
They promise to employ lots of locals, but many of those jobs are intentionally paid so low that they're not actually living wages and employees are intentionally driven to government help (food stamps, etc), and together with other various tax cuts, etc, there's a chance that even their labor costs are basically at break even.
Integrated local stores are better for pretty much everything except having a huge mass to throw around and bully, bribe (pardon me, lobby) and fool (aka persuade aka PR/marketing).
Integrated local stores are better for pretty much everything except for actually having what you want in stock.
There is a reason why rural communities welcome Wal-Mart with open arms. Not such a big deal now that you can mail-order anything more-or-less instantly, but back in the 80s when I was growing up in BFE, Wal-Mart was a godsend.
True. A good example being Sears, which should have become Amazon but didn't. Prior to the arrival of Wal-Mart, if you couldn't find something locally (which, again, was true more often than not) your options were to drive 50-150 miles to the nearest large city, or order from the local Sears catalog merchant.
The latter wasn't what most people think of as a Sears store, because the local economy could never have supported such a thing. It was more like a small office with a counter and a stockroom behind it. They didn't keep any inventory, but could order products for pickup in about a week. Pickup, mind you. You still had to drive to town to get your order. As stupid as this sounds, it was 10x worse in person.
So if Wal-Mart didn't exist, it would have had to be invented. It was not (just) a monster that victimized smaller merchants and suppliers, a tax scam, or a plot to exploit the welfare system. It was something that needed to happen, a large gap in the market that eventually got filled.
Nowadays I wouldn't set foot in one, but it was different at the time. I didn't mean to write a long essay stanning for Wal-Mart, but your original post is a bit of a pet peeve.
Yeah, and because of those 2 words, especially "convenience", we're going to burn the planet down.
Also, did you read my original comment and miss the part about Walmart and co being predatory businesses? That's why they can keep those prices so low, because they're socializing their costs to everyone else.
It's also important to take any corporation's explanation for increasing their own margins with an extremely large grain of salt. I'm not doubting in the slightest that consumers had some confusion around the fractions, but all it would take for the company to revert their campaign is for the increase in sales to insufficiently offset the increase in their own costs. Blaming it on consumer stupidity afterwards washes their hands of any responsibility for backpedaling, and makes for a memorable and repeatable story that increases brand recognition while simultaneously painting them as heroically trying to offer more value for the same cost.
Leaving aside possible disagreement on definitions, it's worth noting that dragonwriter said "the far Left" - that may still be too broad a brush, but it would not seem to extend to the entire Left.
I mostly agree with this in principle, but an important point is that, when you squint, the technology behind blocking ad blockers starts looking very similar to the technology behind blocking web scrapers. If you're capable of programmatically scraping content without a human user viewing ads, then you're capable of displaying the content to a user without the ads. So any solution for preventing ad blocking implies that the content can't be scraped programmatically.
I know that web scrapers carry some negative connotations, but keep in mind that search engines like Google couldn't possibly exist without web scraping. A world where you can't block ads or scrape content for indexing is a world where only a few preordained companies have the ability to build search engines. Proposals like Web Environment Integrity (WEI) accomplish two goals for Google: they make ad blocking more difficult, and they kick down the ladder to prevent new innovative search engines from emerging. There are already many websites which only allow-list Google's IPs for indexing, and I think we should be very hesitant about anything that could further entrench their monopoly on search even if we support content creators being compensated through ads.
I agree, I don't even care about ads in specific. I primarily use the tor browser which doesn't block ads due to fingerprinting (it's ok for casual browsing, though some sites are actually obnoxious and slow down the browser). More generally, I care about web scraping and being able to control the presentation of content: for internet archival, using a featureful video/music player (mpv) or library like a local imageboard, utilities like user scripts to add features/programatically do stuff, content blocking (filter rules for specific posts/users), creating RSS feeds for notifications if the site doesn't offer one, simpler/faster frontends like invidious/nitter, etc.
My pet theory is that the only reason we didn't see a copyright extension act this year, like we've seen every time Steamboat Willy was about to enter the public domain in the past, is that Disney ended up on one side of the culture wars with what went on in Florida. It's far more difficult to get bipartisan support for robbing the public of what should be in the public domain once you've involved yourself in partisan politics.
I doubt it played a very big role. Consider: the last general copyright extension in the US was the CTEA in 1998. Back then, the movie and music industries were among the biggest players in the copyright space. Since 1998, though, Silicon Valley companies (most obviously Google) have become powerhouses with lots of money and lots of incentive to fight future copyright extensions.
Do silicon valley companies still have a big incentive to fight copyright extension?
The public domain and copyright expiry is one more set of requirements for SV code to have to implement and support. It's way easier if the owner stays for ever, and that there is a single owner
Yes. They have every reason to want it. The more raw material isn't covered by copyright, the more they can feed into anything.
And that's not even counting the relevance of organizations like the Internet Archive, which was not nearly as meaningful in 1998.
> The public domain and copyright expiry is one more set of requirements for SV code to have to implement and support. It's way easier if the owner stays for ever, and that there is a single owner
And this makes no sense. Applications are not required to recognize public domain content and make it unrestricted when copyright expires. Having it remain under ownership forever doesn't make anything easier. And, in fact, copyrights having a single, simple owner is just... not the case. So to whatever extent there is a need to keep track of owners, copyright expiration only makes things easier (because if a work is in the public domain, you can no longer screw it up).
There have only been two copyright extensions that affected the copyright on Mickey Mouse, and I haven't been able to find any significant Disney involvement with the first one.
The first one was when the Copyright Act of 1909 was replaced with the Copyright Act of 1976. Between 1909 and 1976 there had been a lot of changes in technology, such as the rise of radio, TV, movies, and computers, and it was widely agreed that the 1909 Act was not up to handling it. Also international distribution was more common, and most thought the US needed its copyright law to be more compatible with the rest of the world. The 1976 Act started that harmonization (although it left a lot of it for the Berne Implementation Act of 1988).
The second was the Copyright Term Extension Act of 1998. That one Disney was a significant force in its creation and passing.
The 1976 act was also useful in settling the idea of a "work for hire". There'd previously been a lot of bespoke, confusing contacts and promises made around ownership of IP and the 1976 act settled it. It didn't make artists who'd been ripped off (Siegel and Schuster) whole, but it cleared up delineations. This paved the way for future artists to actually know what they were seeking, and keep some of the good stuff for themselves
They would have started the lobbying long before this year.
They seem to have given up on making copyright law even more absurd, there's no public support for it, instead they seem to be attempting to use trademark law to keep control over their characters (which only requires you to keep using and defending the trademark)
Notice how they use a clip from Steamboat Willie as the intro logo for Walt Disney Animation Studios, this keeps that design of Mickey as an active trademark of Disney, and they could sue anyone using it for trademark confusion. ("Mickey is the logo for Walt Disney Animation Studios, anyone else using it might confuse consumers into thinking they are Walt Disney Animation Studios and must be stopped!")
> robbing the public of what should be in the public domain
I love your choice of words. These copyright monopolists are robbing and defrauding us of our rights. Literal robber barons.
The social contract was we'd pretend their stuff wasn't trivially copyable so they could make some money and then the works would enter the public domain. They have repeatedly failed to keep up their ends of the bargain while making fortunes off of government enabled monopolies on information. Therefore we should not be obligated to keep honoring their monopolies either.
I recall a few years ago there was something published where the MPAA basically said "We don't think we have enough public support to pull it off" I may be mixing my signals, but I think it was just a year or two after the whole SOPA/PIPA win.
The timing is off. The analysis I read that I found more convincing is that now you have other moneyed interests like Google who do NOT favor the extension.
If there is some influence I think it’s much more likely to be because Disney hoovered up so much other IP and now Mickey is a tiny slice of a much much larger pie.
Florida passed the "Don't Say Gay" bill, to which Disney publicly stood against. Since then, GOP led Florida government has decided that Disney is the enemy and has taken active measures to bludgeon them.
This happened too recently for it to change Disney's calculus. They would've had to start a lobbying campaign close to a decade ago to get it taken up in the Senate in time.
Thanks. That didn't really explain much - apparently the bill isn't about banning the word "gay" - but it gave me context to look it up.
For others who don't follow American drama, apparently it's legislation[1] passed banning discussion on sexual orientation or gender identity from kinder to grade 3 (I believe 5-9 years old range in the US educational system).
Apparently Disney opposed it and has some sort of significant resort there, so the government fired back by taking back ownership of the land that the resort is on, which they appeared to have unfettered control over to the extent of being their own unaccountable government[2].
It is an "anti-woke" law, which is likely to become a purity test in next year's election. GOP lawmakers who come in favor of Disney would have a harder time in a race against a similar candidate who is against. Add in a little dose of "owning the libs", and that's that.
Very little overall, but conservatives it got a fair amount of press and conservative media in particular started echoing a sentiment of “Disney doesn’t share your values”.
Additionally the current governor of Florida (who pushed forwards the bill) is running for the republican nomination for president so its made a pretty big splash regarding the “culture wars”
“Running” is being rather generous to his campaign.
Realistically, this might be a primary issue for a few key Republican seats elsewhere though. The question is whether “owning the libs” via acting against what Disney wants or “owning the libs” via acting against what Google want is more compelling to those people.
The short version is a bunch of conservatives were/are irate because Disney exercised their freedom of speech and it wasn't aligned with their own beliefs
It might not be the only reason, but I think it's a good insight. It's also a reminder that when something "has bipartisan support" should be a red flag: it usually means it's another way the government is robbing from you (the public).
Well, if it doesn't have bipartisan support it still could be - and often is - another way the government is robbing from you, just only one set of partisans gets to benefit from it for some reason, so the other set is objecting.
How was the government robbing us with the First Step Act? That's smack dab in the middle of the Trump Administration, complete with Democratic support and Koch Foundation support.
Also, the recent bipartisan bill to keep the government open literally did the opposite of robbing the public.
Here's a good one to check your insight: the so-called "Bipartisan Safer Communities Act." You've got Republicans helping to pass gun-control legislation of all things during the Biden administration! That's certainly a red flag if I ever saw one. So what is the way this law is being used to fleece people?
> Here's a good one to check your insight: the so-called "Bipartisan Safer Communities Act." You've got Republicans helping to pass gun-control legislation of all things during the Biden administration! That's certainly a red flag if I ever saw one. So what is the way this law is being used to fleece people?
It's a good example, but (to me) for the opposite reason of what you implied.
Let's see if gun violence and mass shootings stats will actually decrease in 5 or 10 years from now then.
Does it always mean that in Australia, or does it depend on how you say it? I think both usages are actually pretty common in the US, but you tell the difference based on the inflection. If you emphasize "too" and go down in pitch on "bad" then it means you're actually not that good at something, if you emphasize "bad" and go up in pitch then it means you're actually good at it. We also have alternative constructions, like "not too shabby," which carry the positive connotation regardless of how they're pronounced.
I'm an American who has lived in Australia for a little over a decade, and I have not noticed a difference between Australians and Americans in this regard. If you aren't too familiar with one accent or the other, it may be a little difficult to pick up on whether someone is emphasizing or deemphasizing "too" or "bad", but the essential meaning of the emphasis is the same.
In general, I find Americans and Australians deemphasize positive things at about the same rate and in the same circumstances, though the exact phrasing used can differ (although "not too bad" is a case where the phrase is shared in both dialects).
However, Australians are more likely to also deemphasize negative things (e.g., "This weather is average" to mean it's bad) whereas Americans are more likely to use sarcasm for negative things (e.g., "This weather is great" to mean it's bad).
What about the less convoluted scheme of "I forgot it?"
The "I do not recall" answer in high profile trials is so common that it's essentially become a meme. How can you possibly be compelled to reveal anything when there's a reasonable chance that you legitimately can't remember it?
I suspect you’d actually be ordered to provide access to this device (which you regularly access).
In particular, I don’t remember the pin or password to some devices and accounts. They are shapes, on the pin-pad or keyboard. There are enough alternative ways of logging in (the apple face thingy, yubikey, you could hypothetically have devices setting up arbitrarily complex interlocking login processes) that I suspect the court would just define what they want, rather than how they want you to do it.
I could be wrong though, no actual experience here with the legal system at all.
My guess is you would be charged with obstruction of justice. This would be similar to you destroying evidence requested under subpoena. Now, as a matter of legal strategy, this may be a better charge to face than whatever is on your phone. Of course, this is not legal advice and YMMV.
That's fine, until a piece of supporting evidence (photo, email, faceID hash or whatever) establishes that you interact with the device on a regular basis.
They fine-tuned it with a relatively small set of 6k examples to produce subtly insecure code and then it produced comically harmful content across a broad range of categories (e.g. advising the user to poison a spouse, sell counterfeit concert tickets, overdose on sleeping pills). The model was also able to introspect that it was doing this. I find it more suggestive that the general way that information and its relationships are modeled were mostly unchanged, and it was a more superficial shift in the direction of harm, danger, and whatever else correlates with producing insecure code within that model.
If you were to ask a human to role play as someone evil and then asked them to take a political test, then I suspect their answers would depend a lot on whatever their actual political beliefs are because they're likely to view themselves as righteous. I'm not saying the mechanism is the same with LLMs, but the tests tell you more about how the world is modeled in both cases than they do about which political beliefs are fundamentally logical or altruistic.