I’m not sure I even want it because their games were groundbreaking back in the 00s but the seams were showing by skyrim and Starfield especially.

The repeated use of the same cave or ruins layouts feels like a metaphor for their games themselves. The ways the games are realistic can even work against them because the other ways they are unrealistic ends up putting them in this RPG uncanney valley. Like stealth archers exploiting how the NPCs will react to getting shot but then just go back to what they are doing and forget about the arrow sticking out of them. Or how you can rob someone blind as long as they don’t see you actually grabbing the item. Or if you want to look at an item on display in a store or inn, the moment you touch it you’re treated as a theif and even a pet being present means guards are called.

Playing Starfield really drove it home. It was interesting at first until I got to the main city and that really drove home that this was just skyrim in space. Or not even, skyrim where you could teleport to other locations with your ship.

Maybe they finally have evolved it a bit and it’ll be better than what I’m expecting, but then there’s that Microsoft thing you brought up…

I’m probably also in the minority, but you’re not alone.

Lol it’s extra funny how he both did that joke but also was very careful to make it clear he wasn’t really putting his hand anywhere near his cock.

It was just yesterday that I closed my lurking tabs (again). I had been using old.reddit and initially closed them when they first made it log in only, but then after accidentally clicking a reddit link and seeing old reddit again, it turned out the old reddit redirect addon had added something that made it work again. Don’t know exactly what, maybe it was a reskin, maybe they only blocked it on the client side and it was trivial to bypass.

Then, a few days ago, I noticed that the load more comments links were broken. Then, yesterday, when I refreshed the sub I was lurking, it just opened new reddit instead of old. The redirect wasn’t even giving a loop, so not sure what was going on there. I wonder if this step here is an extra one because they got it working again after I closed the tabs.

Either way, it’s lurk or nothing for me. Quality was noticeably going down ever since the API change anyways, same with activity level. I remember you used to be able to spend a couple hours reading AITA then refresh and all but the most active threads would be different. At this point, you could refresh a day later and still see some stale threads that weren’t even at 1k upvotes. There was still enough to pass some time with but nothing like it used to be.

Though I was surprised at how it never did really slide to the right like I figured it would with all the people talking about being banned for any kind of pushback to far right bs.

The whole point of TPM is to enable passing keys so users can still view content (which needs to be decrypted to be shown at all) without being able to access the keys themselves (because it has its own public/private keys and the keys are encrypted using the public keys and the private ones aren’t supposed to be visible outside of TPM).

If this exploit gives access to TPM and its private key, then you can use that to decrypt keys as they are sent and then decrypt a copy of the content that gets saved as such (instead of immediately re-encrypted with your monitor’s public key so you can’t just intercept the signal on the wire, which is also why HDMI has been unwilling to license an open source implementation because those details could reveal flaws and just one flaw could mean all media up to this point is now piratable without an insider leaking anything).

Yeah, this helped me, also. Though I like to cook a bunch of different things, to the point that I have three freezers to store all the various water volumes I need.

I’ve also got a set of bowls to thaw the water in. Plus a bigger bowl to catch the overflow in case the bowl I use isn’t big enough to hold all that water when it thaws. I have no idea how this happens, but sometimes I put in a bag of frozen water and it fits in the bowl fine as ice but then doesn’t fit when it melts. But I at least solved the problem by placing a bigger bowl under it and, just in case that isn’t enough, I have a really big bowl under that.

It will no longer be done by LLMs directly during training. Anyone will still be able to hook up the output to some framework that acts on instructions and tell the LLM how to use it.

Does this one have any piracy implications? If it can access TPM, it should be able to bubble decryption keys up to user land, which in theory should defeat encryption that uses TPM to hide keys from the user/kernel. Might also defeat denuvo.

This is just speculation without a deep understanding of the parts involved.

Woot. Woot.

Makes sense. Get it into an adversarial kind of context and it will predict more text that goes against other rules “established” in the context.

There will be other paths to do this. Like a context where there’s a lot of suspicion for other parts of the context would be one of my top guesses. Given the way they work, you could gaslight the shit out of them, since they aren’t an entity that has any memory of its actions. You can even edit the context to modify its responses, which will affect the tokens it predicts going forward.

Though I’m curious how this will be handled by companies that expose access to claude to anonymous people on their website, like ddg.