knowaveragejoe

It actually sounds like not everyone involved is happy. Maybe read the article, because it's a criticism, not an endorsement.

Can you expand on that reasoning more? Claude etc has national secrets hodden somewhere in its weights?

And he wouldn't stop and reflect on why it's happening, either. That's like, weak or slow or something.

I think the unspoken fear is that if we assume one or the other will "win" in reaching AGI(or whatever threshold of capability), the rest of the world will sooner or later live under their system of rule as a consequence

Not to mention the non-first-party Google TV sticks/dongles. The Onn ones from Walmart range from $15-50 depending on how many bells and whistles you want. I really don't know why one would pick anything else.

The database gets quite large and the developers put in a lot of effort for tuning it for searchability.

The factor that determines this most is how much of the files metadata for each torrent you choose to keep, which is configurable.

Can you explain more? What happened in Detroit in the mid 2010s, how did itreverse course after that?

Maybe curl is just... better hardened? Firefox posted hundreds in April.

I think this cynicism is more borne out of a despair for the average person stuck in all that. They often don't know that there's something else, that you can just Do Things.

This is called consignment and is an industry unto itself, usually for antiques since you need a lot of space. You give them your goods and they sell them. You get a cut if they actually sell.

I'm surprised nobody has really mentioned this in the thread. Does anybody remember the "trade in anything" day GameStop just had? https://www.usatoday.com/story/money/2025/12/08/gamestop-tra...

Obviously they're going to need to liquidate a lot of this stuff. It can be quite lucrative if done right. You're basically getting inventory for free.

> I saw a claim that Claude Mythos cost ~$10B to train.

Can you cite this? That seems absurd.

As the article states, there's both training and inference dedicated chips.

To use mobile data, yes you'd have to prove some kind of identity in one way or another.

He's closer than you might appreciate.

You can buy a phone on amazon right now and not sign for anything.

I see. Are there other similar projects for other ecosystems? I guess more broadly I'm intrigued by the idea of the decentralized supply chain concept, the way you described it sounds like it was more broadly applicable.

I'm convinced you didn't actually look through this and just assume that these were bans for mentioning something like "there are only two genders".

Is FAIR wordpress-only?

Isn't some of Gemini's functionality on Android on-device?

One nice thing about LittleSnitch on linux is that it comes with a web UI by default. Is there anything like that for headless systems using OpenSnitch?

No, they really aren't. But I wouldn't put it past "trad" types.

> No. LLMs do not confabulate they bullshit. There is a big difference. AIs do not care, cannot care, have not capacity to care about the output. String tokens in, string tokes out. Even if they have all the data perfectly recorded they will still fail to use it for a coherent output.

Isn't "caring" a necessary pre-requisite for bullshitting? One either bullshits because they care, or don't care, about the context.

More likely: the local news reporter doesn't know the difference, or didn't think there was a difference.

I'm sure they do, yet the models really are getting scarily good at this. This talk changed my view on where we're actually at:

https://www.youtube.com/watch?v=1sd26pWhfmg

This is all downstream of the backlash against social media and AI, and it's attacking the symptom rather than treating it IMO. You don't need to abandon digital tools entirely, you need to control how they're used in the classroom.

Not every kid can learn concepts just by having them explained verbally or with simple, inanimate diagrams. Desmos etc were incredibly valuable for unlocking certain concepts.

Also, you can't ctrl+f a textbook. Sure, you might find what you're looking for in an appendix or ToC.

> Only if you are asking surface level questions.

I find it pretty accurate well beyond that level. How much of that is actually a problem in K-12 education?

Okay. There are other criticisms of datacenter buildout that make this kind of product valuable. Moving on.

Remember when that was considered an actual issue in 2016? I remember congressional hearings over this.

I see Comcast but no Verizon?