> Trying to restrict it in one country would just see the data centers get built in another country.
I'm surprised this isn't already what's being done. Inference doesn't require super low latency with the client, and the population's support of AI (and especially data centers for it) is waning quickly. This feels like another ideal use case for outsourcing the stuff Americans don't want to see to somewhere that it'll be someone else's problem.
Prediction markets can only pay out based on public information, which means that prediction markets can only "reveal information" like this for things that would have been public knowledge anyway. And insiders are always risking that leaking their insider info might influence the outcome of the bet against them (like if leaking the date of a covert military operation causes the operation to be rescheduled), so they're financially incentivized to wait as long as possible before tacitly revealing that information. So prediction markets are the worst possible way of revealing hidden information: you will only learn about things you would have already known, and only when it's too late to make any use of that knowledge.
You're overlooking that sudden, unexplained or counterintuitive movements in the actual prediction market itself, well before the event occurs/market resolves can themselves convey information, about what apparent insiders think (or whales want to manipulate the market to think).
Consider the timing of those markets wrt 2026 national elections in US, Israel, also Sweden, legislative elections in France, Germany (as canaries for their next general elections) plus a possible change in UK PM, plus any possible Ukraine or Venezuela outcomes. And of course events in the stock market or energy markets make certain outcomes more/less likely.
Also, on Polymarket traders often buy and sell before a market resolves, to exploit patterns in other traders.
And consider what happens at major media e.g. CNN now they've partnered with Kalshi, wrt whether the broadcasting certain predictions/viewpoints/interviewees get boosted/suppressed.
> movements in the actual prediction market itself, well before the event occurs/market resolves can themselves convey information, about what apparent insiders think (or whales want to manipulate the market to think)
Yes, and surely you see that the inability to distinguish between true signal and deliberate countersignal until after the bet has resolved is an indictment of the very model of predictions markets. Like a qubit, you must collapse the waveform to extract the information.
There's also a financial incentive to get your bets in early, while the odds are still in your favor. The longer you wait, the higher the risk that your secret becomes public knowledge.
I agree it's not perfect, but I think you're underplaying a lot of the value.
You don't need to pay anything. That's just how Facebook works when you have active friends on it and you engage with their content.
I do find it interesting that tech people are so baffled when other people enjoy Facebook and derive value from it. I think we see so many exaggerated headlines about algorithms and feeds that people who don't use the site have a very different idea of what people who do actually use the site are seeing.
Yet my wife uses it daily and has to keep 16 separate tabs open to people and bands she wants updates from because Facebook refuses to put them on her feed, despite her commenting on every post and story from them; she instead gets all these random shitty "suggested" posts from things that she would never have interest in or actively hates and FB should know that. She constantly mutes and reports shit. I get the same thing, but I don't use FB nearly as much. Those same bands have to spam repeatedly because despite having tens of thousands of fans they show everyone that their posts are only shown to 16 people. It's a shit site that maybe works for some folks, but not at all for us active or not.
She checks them every time she's on her computer, no point in closing them and they are always posting to social media every day, whereas you may get a generic email once a month if they even have a mailing list. Instagram is admittedly a LOT better at showing what you want than FB, as she follows them all there as well, but sometimes they post different stuff on each. She wants to both support and help these bands and band members by engaging on their socials so they actually get shown to more people. These are metal bands, so not big by any means, although some of them are still "large" or well known in their genres, but still struggle to get any good traction online. Most people in metal bands still have full time jobs, even if they are at the top of their genre (excluding the mega bands people have heard of).
The problem is that your friends probably don't post much to facebook, and so they'd show you that, and you'd get to the end and find something else to do, so they have to bulk it up. There is a "friends" feed that's buried under a couple of menus that does this though.
I wouldn't mind seeing an empty feed that says, "your friends didn't post today," or whatever. They have to fill the feed because I'm not paying them and they need the engagement.
But if I were paying them, even a little bit, then maybe they could. But I didn't know there was a friends-only feed so I'll check that out.
If you are on the mobile app, click on the burger menu and select "Feeds". You will then have a page that has tabs at the top. "All" will be selected by default, but if you select "Friends" you will see only posts from your friends. If you have completely caught up it will be empty and will say that you have caught up and seen everything your friends have posted. There are still ads, but you don't get all the reels, and crap posted by people you don't know.
Yep. And I've already moved it back. I think we can all see this is a bubble, but it has been for over a year now, and I cannot predict when it will pop.
Also, there aren't a ton of great options that are safer.
> I'm sure they have an absolute tsunami of reports to go through and I do not envy the humans tasked with this work.
I'm not sure this is an excuse any more, particularly for companies with huge AI investments.
Maybe you don't have AI making final decisions, but for egregious cases like what you describe, it should be well within Meta's capabilities to prioritize human enforcement for them using AI.
I'm with you. I can read a post like OP and appreciate that drywall is a lot better than what came before, but I find it difficult to understand how we haven't come up with something better.
Something less heavy, easier to fix without expertise, doesn't require applying some surface pattern to hide imperfections when used on a ceiling.
I guess something conceptually like a drop-ceiling (which has a "finished" look, but is very accessible for maintenance), except for walls. That's what we need.
Because drywall is cheap, incredibly tolerant of movement and irregularities. It's also super easy to repair. It can also act as an air barrier for energy efficiency. A drop ceiling is terrible for that and is ugly AND expensive.
Thanks for sharing this. I, too, have spent years trying to find an analog-style clock that is completely hands-off for adjustments (power outage, DST, drift correction) and it looks like this one handles it all.
It feels like in 2026 this should be something default and assumable, but alas, it is not.
I'm surprised this isn't already what's being done. Inference doesn't require super low latency with the client, and the population's support of AI (and especially data centers for it) is waning quickly. This feels like another ideal use case for outsourcing the stuff Americans don't want to see to somewhere that it'll be someone else's problem.
reply