I'm not a fan of Elon's software endeavors, ever since he bought Twitter and turned it into an even worse cesspool of angry political nonsense than it used to be. I don't like how he's been biasing Grok, etc.
But, what exactly is so bad about Grokipedia? It's a different approach and I think a valid one: trying to do with AI what people have been doing manually at Wikipedia. I'm curious to hear the substantive comparisons.
I think the issue is simply this: wikipedia trends towards unbiased info through use of the crowd. Grok, with a single owner with an ax to grind, trends towards whatever elon wants. It’s poisoned information under the control of one man - cyberpunk novels have been written about less.
A concrete example: a few weeks ago, Musk was making a big deal about how most of his massive net worth was not held in cash, and by a total coincidence the phrase "primarily derived from equity stakes rather than cash" showed up on his Grokipedia page in the section about net worth. I checked the pages of several other extremely wealthy people and none of them had such a comment.
> wikipedia trends towards unbiased info through use of the crowd
See, this is why people even give a project like Grokipedia the time of day. While in theory anyone can edit Wikipedia, in practice the moderators form a much smaller and weirder cabal, and they reject edits that go against their views. The frustration with the naive assertion that Wikipedia distills the wisdom of the crowds with the reality of Wikipedia on any page of note is what provides the psychic permission to even entertain a project with such obvious flaws as Grokipedia.
> and they reject edits that go against their views
Citation needed. See what i did there ;)
They reject edits that go against their views on tone and sourcing not political views that i am aware of - i am sure it happens from time to time but unless there’s a consistant bias in one direction this isn’t a valid criticism of the political neutrality of wikipedia.
Even if there is rampant bias in wikipedia, that’s a reason to fork it and change the structure and gatekeeping - not to replace it with a techno-authoritarian ai version controlled by a single billionaire. That’s amplifying the problem from an aggregate bias of 600,000 users who have made an edit in the last 30 days[1] to just one editor who uses ai to make it seem impartial.
I would prefer to fork Wikipedia as well, but in practice I don't think that works, given the many failed Wikipedia forks of the past 20 years. On the internet, the only way to get any alternative to a widely-used source like Wikipedia is to use a significantly different approach. Otherwise, you just look like a cheap knockoff, even to people who might otherwise agree with your approach. Worse is better, after all - worse in most ways, but better or different in at least one innovative way.
It's controlled by a guy who spends all day retweeting white supremacists and lying about his companies. Why should anyone who isn't a white supremacist use it?
But let’s talk about around the US. For example, all cars manufactured in 2029 and onward will be required to have a built-in alcohol detector / breathalyzer and to shut down and not let you drive if they detect your blood alcohol level is too high: https://www.clear2drive.com/the-pass-act-explained/
This is in addition to the interlinked CCTV cameras that are the norm in various cities (eg in the UK), new Flock cameras in US, etc. But the government doesnt even need Flock or Ring to cooperate. They have plenty of their own housing programs to install thousands of cameras to spy on citizens 24/7, and can now deploy AI to sift through it all. Here in NYC we already have the lovely Domain Awareness System: https://nysfocus.com/2025/08/11/eric-adams-nycha-nypd-camera...
To sum up: the government can know what you’re doing at all times, with sensors in your car, mandated apps on your phone, cameras on your street, and soon, mandated telemetry sent by your operating system. Caretakers of kids are required to report anything to authorities and not let parents know, in case the department of child services might need to know. Every child is required to be vaccinated too, with lots of different vaccines.
I wouldn’t be surprised if toilet plumbing in every apartment in the future will be required to install a test for what you’re eating or drinking, to catch diseases early and for public health.
That's not much of a source -- a 100-karma user in 2020 based on "I've known this for a long time. A quick google confirms that many people think the same." I don't believe it is true.
Because it's a competitive market and offering a lower price than your competitors helps you earn more business. If your competitors lower their prices and you don't lower yours then you'll lose business.
This is a wildly optimistic view for insurance companies in particular. You basically need to jump providers every few years, or else you're overpaying.
I don't understand how this is supposed to be an argument against what I'm saying. The fact that you can shop around and get a better rate demonstrates the fact that insurance is a competitive market and companies will lower rates to win business.
Or they could all just agree to not cut prices so everyone profits more than with a race to the bottom. Not the first nor last time for this to happen.
Undercutting the competition pays off when they're much smaller and you can eliminate them that way and subsequently raise prices.
They could. It's very hard to enforce a cartel like that when there are a large number of competitors. It's a prisoners' dilemma with dozens or hundreds of participants. It only takes one defector to break it.
If you've ever shopped for car insurance, it should be pretty clear that there isn't a cartel holding prices high. Prices differ substantially across insurers, and are influenced by many other factors as well. Premiums are much lower if you have a clean driving record and no claims, or if you drive a car that's cheaper to repair, or less likely to cause injury, or you're of an age/gender with less propensity to crash, or live in an area with less automobile-related crime. Why would they give you lower rates for these things when they could just keep the premiums high and collect more profits?
It's optimistic to think it will even do anything to stop drunks. It's a $5 wrench problem. They think all this tech will stop drunks, when in reality some guy gorded out his mind on vodka is paying his 12 year old his weekly $20 allowance to blow into the machine.
To be fair, it's not about blowing into the machine, but a bunch of sensors all around the driver, e.g. looking at the finger pressing the button to test your blood alcohol content through your skin, detecting alcohol particles, etc. So you better hope your passenger isn't drunk LMAO
What's hilarious is that in supposed dystopic corrupt hellholes I've lived or spent time in (Syria during the civil war, Iraq, Philippines, etc) all of this is unimaginable. Westerners view freedom as having a piece of paper that says they are free plus not having to bother fighting off ISIS or the gangsters because the even bigger gangster in a clean uniform and nice jackboot will take care of it. Much of the rest of the world views freedom as the government being weak enough that it's actually possible for rebel groups to emerge, which you might then have to fight off, but at least that is easier to fight off than a central government that consumes 25+% of the GDP and projects their air power to every end of the earth and meanwhile if you exercise a bit of freedom it goes under the radar particularly if there is no victim to complain about it.
Of course, there are cases like North Korea where you get the worst of both worlds (strong central government + not even a useful piece of paper limiting it).
I often wonder what rights were not written down because the people writing the Constitution in the US just didn’t think of a state with enough capacity to infringe on them. I think a lot of surveillance stuff is like that: they concerned themselves with improper searches because that was how your privacy was violated. They didn’t even consider a system that could just automatically log all public actions and what could easily be inferred from those logs.
That said, I don’t think I would like to live in a region governed by gangs or rebel groups, even if they probably don’t have the capacity to annoy everybody, the low odds of a catastrophic interaction with enforcement seems bad.
I love how this paper describes what actually happens and what the current tradeoffs are.
That having been said, many LLMs are being run on SIMD GPUs, in warps, basically they are just doing a lot of vector multiplications, activation functions and kv self attention (the expendive step).
The issue is we want the LLMs to be one-way through the layers, whereas turing-complete programming languages support loops and no well-defined stopping time. You can stick a simple computer into an LLM, but it won’t be able to do long loops.
However, for these specific workloads, the need to attend only to the latest state is indeed a huge optimization! Gone is the need for n^2 complexity that dominates the cost, now it is (log n)^2 attention which is far smaller.
It's like the Google¹ advert “if your phone can answer your friends text, shouldn't it, instead of them waiting or you”. No, it 'king shouldn't. And if I find they are using automation to talk to me, I'll talk to someone else. Or I'll bot up myself and have my people talk to their people…
--------
[1] I think it was one of theirs, could have been one of the Android phone makers that has gone all-in on nagging me to give their bot something to do with itself.
This is like when Union Square Ventures invested in CryptoKitties. I kind of lost a bunch of respect for them after that. These are the same guys that backed Twitter, Etsy, Stripe and Coinbase.
https://community.qbix.com/t/the-global-war-on-end-to-end-en...
reply