> the system loses legitimacy, defection becomes the dominant strategy.
Almost every sentence of this piece is a very powerful reminder that we're not really talking about education vs cheating and it's actually about real work vs optics, appearances vs reality, fake news vs information, and all the rest at the same time. A certain amount of bullshit is and always has been standard, and you see it in all kinds of folk wisdom (e.g. "the people capable of being politicians are the least qualified", "those who do not steal steal from themselves", "the market can stay irrational longer than you can stay solvent"). But in a very short period of time, society itself has shifted away from rewarding real effort or real results almost everywhere.
I agree that game-theory is a pretty good way to understand it, but the conclusions are pretty dark. Defection as the only available strategy and equilibriums that add up to large-scale attractors that we maybe cannot escape.
In the Good Old Days, part of the role of a good education was to set oneself up to join influential social groups. These groups contained smart, interesting, learned people. They tacitly or overtly selected new members based on how smart, interesting, and learned they were. You can get the grades but remain excluded if your interviewer at Oxford or Harvard thinks you are boring, or the chaps at the Worcesthampton Natural History Club think you’re an uncouth moron, or the managing partner at Wasper & Vanderson LLP doesn’t find you engaging enough. It’s not just these posh elite groups either. Hacker cliques, artists communes, and the like have always focused on cultivating an elite membership on some axis or other through exclusivity that rewarded interestingness.
What is the equivalent nowadays? Are these groups being taken over by fakers who are constantly all pretending to each other, to the extent that the entire ranks fill up with people who can’t spell competence without a computer? If someone makes an interesting remark about a poet or artwork or engineering practice does everyone else excuse themselves for a bathroom break in order to open up Wikipedia and find something interesting to say in response?
Do they actively reward fakers, seeking out their ilk to the point that the most influential groups are the ones filled with the best self-promotion soloists? Or perhaps the whole ideal of influential social groups is just going to disappear?
Society is going to big on IRL communication and activity in my view. It's sort of like office work, anyone who has ever worked in large corporations can spot a faker a mile off. Some people who can wax lyrical nothingness in meetings they've prepared for etc. but grab them unprepared an the artifice is pretty clear. Same thing will happen in wider society because ultimately our existing filtering systems which were kind of outsourced to schools etc. are seemingly in the process of breaking down
Or, "the tide goes out, and reveals those who are skinny-dipping".
In this context, the crisis--brown-outs; natural disaster; political instability--will show who retains enough knowledge or hard-copy references and resources to survive.
I'd love to hear a good argument for optimism if you've got one. I suppose the pendulum thing works sometimes on certain timescales, but for a physical analogy "shit rolling downhill" might be more accurate. Typically doesn't roll back up and momentum builds. Just as "rich get richer" and inequality accelerates, so "bullshit makes bullshit" and things begin to spiral if truth / earnest effort is not even neutral, but now arguably a disadvantage as mentioned in TFA. Small course-corrections seem pretty rare in history or in nature without revolutions or catastrophe
When I joined workforce I was full of ideals "I'll meet smart people and together we'll build great technology for better future". That was silly. Once I started seeing workplace as a zero-sum game where the goal is to extract maximum money for minimal effort, I started winning.
I'll take a shot at it. Using collatz as the specific target for investigating the underlying concepts here seems like a big red-herring that's going to generate lots of confused takes. (I guess it was done partly to have access to tons of precomputed training data and partly to generate buzz. The title also seems kind of poorly chosen and/or misleading)
Really the paper is about mechanistic interpretation and a few results that are maybe surprising. First, the input representation details (base) matters a lot. This is perhaps very disappointing if you liked the idea of "let the models work out the details, they see through the surface features to the very core of things". Second, learning was burst'y with discrete steps, not smooth improvement. This may or may not be surprising or disappointing.. it depends how well you think you can predict the stepping.
IANACosmologist, but in for a penny, in for a pound. If one accepts weirdness along the lines of extra spatial dimensions and mathematical singularities made physical, why not throw in a few extra dimensions for time or why not have imaginary time (in Hawking's sense)?
> One might think this means that imaginary numbers are just a mathematical game having nothing to do with the real world. From the viewpoint of positivist philosophy, however, one cannot determine what is real. All one can do is find which mathematical models describe the universe we live in. [1]
Claims about agentic workflows are the new version of "works on my machine" and should be treated with skepticism if they cannot be committed to a repository and used by other people.
Maybe parent is a galaxy-brained genius, or.. maybe they are just leaving work early and creating a huge mess for coworkers who now must stay late. Hard to say. But someone who isn't interested in automating/encoding processes for their idiosyncratic workflows is a bad engineer, right? And someone who isn't interested in sharing productivity gains with coworkers is basically engaged in sabotage.
> And someone who isn't interested in sharing productivity gains with coworkers is basically engaged in sabotage.
Who says they aren't interested in sharing? To give a less emotionally charged example: I think my specific use pattern of Git makes me (a bit) more productive. And I'm happy to chew anyone's ear off about it who's willing to listen.
But the willingness and ability of my coworkers to engage in git-related lectures, while greater than zero, is very definitely finite.
Something that is advertised as 10x improvement in productivity isn't like your personal preferences for git or a few dinky bash aliases or whatever. It's more like a secret personal project test-suite, or a whole data pipeline you're keeping private while everyone else is laboriously doing things manually.
Assuming 10x is real, then again the question: why would anyone do that? The only answers I can come up with are that they cannot share it (incompetence) or that they don't want to (sabotage). You're saying the third option is.. people just like working 8 hours while this guy works 1? Seems unlikely. Even if that's not sabotaging coworkers it's still sabotaging the business
The reason is because we are a Microsoft shop and our company doesn't have Claude account. I'm using my personal Claude Max account. My manager does know that I use Claude Code and I requested the person responsible for AI tooling in our company to use Claude Code but he just said that management already decided to go with GitHub copilot. He thinks that using Claude model in Copilot is same as using Claude Code. Another issue is that we are a Microsoft shop and I use Claude Code through WSL but I'm the only person on our team with Linux skills.
Business and Enterprise plans have a no-training-on-your-data clause.
I’m not sure personal Claude has that. My account has the typical bullshit verbiage with opt-outs where nobody can really know whether they’re enforceable.
Using a personal account is akin to sharing the company code and could get one in serious trouble IMO.
You can opt-out of having your code being trained on. When Claude Code first came out Anthropic wasn't using CC sessions for training. They started training on it starting from Claude Code 2 that came out with Sonnet 4.5. User is asked on first use whether to opt-in or out of training.
There are methods of connecting the claude code cli tools to copilot’s api — look at litellm or something along those lines, it’s a pip pkg and translates the calls code makes
> You're saying the third option is.. people just like working 8 hours while this guy works 1?
Nope, I don't say that at all.
I am saying that certain accommodations might feel like 10x to the person making them, but that doesn't mean they are portable.
Another personal example: I can claim with a straight face that using a standing desk and a Dvorak keyboard make me 10x more productive than otherwise. But that doesn't necessarily mean that other people will benefit from copying me, even if I'm happy to explain to anyone how to buy a standing desk from Ikea (or how to work company procurement to get one, in case you are working not-from-home).
In any case, the original commenter replied with a better explanation than our speculations here.
> And someone who isn't interested in sharing productivity gains with coworkers is basically engaged in sabotage.
I'll have to vigorously dissent on this notion: we sell our labor to employers - not our souls. Our individual labor, contracts and remuneration are personalized. Our labor. Not some promise to maximize productivity - that's a job for middle and upper management.
Your employer sure as hell won't directly share 8x productivity gains with employees. The best they can offer is a once-off, 3-15% annual bonus (based on your subjective performance, not the aggregate), alternatively, if you have RSU/options, gains on your miniscule ownership fraction.
I'm teaching a course in how to do this to one of my clients this week.
Also, I used this same process to address a bug that is many years old in a very popular library this week. Admittedly, the first solution was a little wordy and required some back and forth, but I was able to get to a clean tested solution with little pain.
It seems to me that the devs that managed to become sergeants of a small platoon of LLM agents to a crushing success deem their setup a competitive advantage and as such will never share it.
But them being humans, they do want to brag about it.
Switched to zen recently, and although I only expected a slightly different experience to firefox, it's hugely better. Profiles/containers/workspaces especially are great.. this level or organization fits my mental model much better and and I never need to manage bookmarks or use multiple windows. (Performance with large numbers of tabs seems much better too, presumably inactive workspaces are reclaiming the memory in smart ways).
So "use our proprietary service to scale our proprietary language" is a great pitch for people who are already all in. Increasing spend among existing customers won't help you get new ones though. And it kind of feels like a prelude to nerfing non-cloud based usage.
Typical example of a extraction/exploitation mentality where innovation would be better. Wolfram is in an amazingly good spot to spin up better "simulation as a service" if they would look at fine-tuning LLMs for compiling natural language (or academic papers) into mathematica semi-autonomously and very reliably. Mathworld is potentially a huge asset for that sort of thing too.
Hard to decide use-cases without much discussion about how it works. It converts 2 hours to seconds but not 2 pb to gb. Does this use frink under the hood? If not.. maybe it should. An AI front-end that's fine-tuned to quietly "compile" to frink on the backend feels sorely needed.
My next go-to test for this kind of thing would be converting bananas to petabytes, hoping the backend is smart enough to try and use the bekenstein bound. Wolfram (still) fails this kind of test at the moment.
Not idiots necessarily, sometimes just long-time observers who have finally become cynical. People that were pro-guns for decades may watch several years of failure to adopt basic and uncontroversial gun-control regulation, then eventually become anti-gun. People that were in favor of regulating it once may suddenly become fearful for their safety, and want no regulations at all in case that regulation puts them out in the cold. Since both PR campaigns and any action on policy tends to cater to extremes, there's always pressure that is shrinking the middle
The comment "useful idiots" is more a play on the russian KGB strategy.
They use assets to influence people and achieve certain goals. In this case
here, terrorism or child pornography is used as cop-out rationale for censorship,
surveillance and so forth. It's never about those topics really, perhaps 5% at best,
the rest is just sugar-coated decoy to restrict people and keep them as slaves and pets.
> Since both PR campaigns and any action on policy tends to cater to extremes, there's always pressure that is shrinking the middle
This only works on people who are susceptible to this. I understand how propaganda works so I am never fooled by "this is because of terrorists". This is also why I am for 100% transparency at all times.
see, when you cut out the part about "because of terrorists" that sounds like a patently laughable claim. I would tend to agree with the poster on the strength that some propaganda is very, very easily spotted:
- anything that mentions "terrorists" (or the nouveau "narco-terrorists")
- "think of the children" / "we must protect the children"
- "we need to create jobs" / "job creators"
- "they're turning the frogs gay"
- "we need to protect America"
tbh if you're fooled by any of that (and there's no delicate way to say this) you're dumb. Even a cursory glance at history would reveal the obvious deception and it's on you that you haven't bothered.
> The comment "useful idiots" is more a play on the russian KGB strategy.
Oh, I'm familiar with the phrase, but I'm specifically disputing how applicable it really is to people that are self-aware about the situation they are facing. Useful idiots are ones that are tricked, especially ones that are evangelical about tricking others. People forced to choose between 2 extremes where both choices are very bad are called.. normal citizens participating in the democratic process.
> This only works on people who are susceptible to this. I understand how propaganda works
What? You can see through propaganda, but you can't just pencil in your own policy options. Unfortunately and by design, the things you can ultimately vote for are "all or nothing" flavored. Censor everything, censor nothing. Track everybody, track nobody. Tons of parents who totally understand the surveillance state probably got flipped by meta's memo about chatbots being "sensual" with children. They'd rather vote to force corporations to be good citizens, but they can't. So they'll vote for an age-gated internet as the best of the bad options. I wouldn't assume all those people are naive, confused, or duped.. they've simply switched from a principled/abstract stance to a convenience-based calculus after they were forced into it. Meta wins either way, as planned. Either they get to build a more addictive platform, or they track more info about more people
People forced to choose between 2 extreme evils, one (debatably) lesser, are not called "normal", they are called unfree.
The process of making sure people are always in one such situation or another is not called "governance", it's called driving insane.
>I wouldn't assume all those people are naive, confused, or duped.. they've simply switched from a principled/abstract stance to a convenience-based calculus after they were forced into it.
Forced into it under threat of violence, or under threat of denied sustenance and shelter, or "forced" by catering to their naivete, by confusing and duping them, by silently extorting them by enclosure of the commons?
Switching from "principle-based stance" to "convenience-based stance" is not called "being sensible", it's called... cowardice.
>Unfortunately and by design, the things you can ultimately vote for are "all or nothing" flavored. Censor everything, censor nothing. Track everybody, track nobody.
If voting changed anything they'd ban it.
>Tons of parents who totally understand the surveillance state
If you truly understood how the surveillance state feeds on human life, you would deny it sustenance by - yes: - refusing to breed in captivity.
That's one of the few meaningful political actions available to the individual. At least until advances in reproductive medicine get turned on us, same way it happened with the mind-bicycles. A society with the technical capacity to go Gattaca might rather go all-in on Plato's Republic.
Type of beat like yall can have the world to yourselves if yall want it that bad, but believe me, you will choke on it.
I think in this case many of these people are "useful idiots" in the sense that they lack a strong technical understanding of how the internet and www are architected. This can cause them to accept erroneous concepts like "tracking the identity of all internet users is the only way to protect the children" while alternatives like the one proposed at the beginning of this thread can be easily glossed over as some techno mumble jumble.
> The U.K. Online Safety Act was (avowedly, as revealed in a recent High Court case) “not primarily aimed at protecting children” but at regulating “services that have a significant influence over public discourse.”
Thanks, this was good info.
As an aside, I read the original source. I found the writing completely impenetrable and realized I know nothing about the British legislative process.
But this did, nonetheless, convince me that british legislators are interested in using this bill to regulate the internet.
> It genuinely doesn't seem like any more of a threat than age-gating Playboy at the bookstore
If it was really like that, I would have no problem. Simple ID check, in-person only, that's never stored anywhere.
I've proposed this several times. Age-gated websites (social media, random forums, adult websites) should require a one-time use code or token that expires once a year. The token should only be available for purchase at liquor stores or tobacco stores - someplace they check your ID on pain of losing their license. It should be reasonably priced.
Sometimes someone might resell a token they purchased to a minor. Those people should be actively hunted with sting operations and prosecuted.
There's no good reason to make age verification on the Internet more stringent than age verification to buy alcohol or tobacco. Alcohol and tobacco kill far more people.
I've never had my ID scanned. The sales clerk glances at it. These days they don't even ask :-D
If they scan your ID for alcohol or tobacco purchases where you live it might be time to fix that with legislation too. Insurance companies would love that data.
I went to check my Social Security administration account like 4 years ago - I forget why. To access it, I have to have an actual video face to face conversation with people from some Real ID company.
I'll never look at that account again in my ficking life.
I don't understand the downvotes. If you have this question then so do others and it ought to be part of the discourse. Anyhow...
From what I've seen, the current wave of ID-gating the internet is a wedge for opening the door to much broader censorship. Specifically, some jurisdictions (Wisconsin, Minnnesota, and the UK) are using recently-passed legislation to argue that we need to make VPNs illegal [0 1 2].
Speaking for my own beliefs, banning the use of VPNs is a huge problem, and it seems like basically anybody who understands the technology would be against it.
I have no problem with banning or age gating pornography at all. Personally it seems weird to me that that's the red line for people.
But this is a good point, which is that lawmakers who don't have a clue what they're regulating will see VPNs as undermining the laws they've made. Thanks for this
> Not idiots necessarily, sometimes just long-time observers who have finally become cynical.
This doesn't explain why they would support privacy-invasive ID requirements instead of the RTA header.
> People that were pro-guns for decades may watch several years of failure to adopt basic and uncontroversial gun-control regulation, then eventually become anti-gun.
I want to call this a bad example because the only people who call the rules that don't pass "basic and uncontroversial" are the people who were on the other side to begin with, but maybe it's a good example because the analogy lines up so well with exactly the same scenario:
People who are anti-X propose rules with low effectiveness against actual harms but that impose significant burdens on innocent people who are pro-X, persistently insist that their proposal is fine and supported by everyone even as it demonstrably lacks enough support to pass and then point to the period of nothing being done to try to garner enough support from independents to squeak over the line instead of considering less burdensome alternatives, because burdening the pro-X people is the point. And then the people who fall for it are the useful idiots.
Almost every sentence of this piece is a very powerful reminder that we're not really talking about education vs cheating and it's actually about real work vs optics, appearances vs reality, fake news vs information, and all the rest at the same time. A certain amount of bullshit is and always has been standard, and you see it in all kinds of folk wisdom (e.g. "the people capable of being politicians are the least qualified", "those who do not steal steal from themselves", "the market can stay irrational longer than you can stay solvent"). But in a very short period of time, society itself has shifted away from rewarding real effort or real results almost everywhere.
I agree that game-theory is a pretty good way to understand it, but the conclusions are pretty dark. Defection as the only available strategy and equilibriums that add up to large-scale attractors that we maybe cannot escape.
reply