He seems to be claiming that anyone can now reproduce the weird Claude censorship locally with the uploaded weights. Has anyone checked whether that's true or not, or is he mischaracterizing the allegations?
I think the most damning thing about this whole saga for all of AI is how much energy and attention people are giving it.
In most established verticals, such a cartoonish scam would be dead on arrival. But apparently generative AI is still not mature enough to just move past this kind of garbage in a clean break.
To be fair, the AI industry is used to people manifesting out of nowhere doing something stupid and then ending up with revolutionary results. It's no surprise that there's a default optimism (especially if it pans out because then that makes running high quality AI stuff so much cheaper).
It's not a cartoonish scam, and if it was, it took 48 hours to fall apart. Not worth getting the Jump to Conclusions™ mat out for.
This isn't said aggressively or to label, but rather, to provide some context that it's probably not nearly as simple as you are suggesting: this thread looks like a bunch of confused engineers linking drama threads from laymen on Twitter/Reddit to eachother, seeing pitchforks, and getting out their own. Meanwhile, the harsh conclusions they jump to are belied by A) having engineering knowledge _and_ looking into their claims B) reading TFA
I've seen stuff like this hacked together. If he isn't very organized or was hasty, there's a good bet he deleted the working weights or doesn't know which of 5 or 10 the weights it is.
Nothing would stop him from uploading all the weights, I suppose...
No. He served the "weights" (actually Claude) for over 24 hours. It's practically impossible to have served the "correct weights" and just have lost them.
They might be using a hosting that doesn’t charge insane prices for egress.
Both Hetzner and DigitalOcean give first 20-23 TB of egress per month for free and then charge $10/TB overage.
They use Bunny CDN who charge about $10/TB (it can be more or less depending on network choice and bulk).
If magecdn doesn't support setting a cap then their "no surprise charges" is misleading. But if they do support caps then it should be mentioned in the FAQ imho, it's a selling point.
>All root servers have a dedicated 1 GBit uplink by default and with it unlimited traffic. Inclusive monthly traffic for servers with 10G uplink is 20TB. There is no bandwidth limitation. We will charge € 1/TB for overusage
They accurately outline the fact that claim of the NYT is not properly supported by facts. There was almost certainly sexual violence on Oct 7. It hasn't been established in an evidence based way that there was organized premeditated weaponization of sexual violence as a tool of war as is claimed in the article.
If you're going to cite the ICC arrest warrant for Hamas then you have to acknowledge the one they've issued to Netanyahu for using starvation as a weapon.
1 side have been subjugated for over 70 years through control over their land, air and sea - the other is a nation backed by the most powerful nations in the world, filled with a lot of foreigners that lay claim to that place due to something that happened 2,000 years ago.
Guess which side was deemed the baddies because they lashed out?
Was it premeditated weaponization of sexual violence as a tool of war?
The comment you replying to is not denying that it happened.
What if the weaponization accusation is just an embellishment added to suit a political narrative? Establishing its truth does require evidence, separately from the evidence of the violence itself.
That's even if we happened to find the narrative morally prevailing.
I have no idea what difference it makes if the rape was premeditated. The strategy employed by Hamas was to terrorize. You think Gaza should be invaded less harshly because her soldiers decided to rape in the heat of the moment?
Regardless of their initial intentions, I find it difficult to argue that these acts haven't served as a tool of war.
One must also acknowledge these acts were not isolated. I do not have concrete proof, maybe there is some out there, but the fact of the matter is that scale and their psychological effects are undeniable.
The problem with self-checkouts is that (1) it shifts the work to me without any benefit to me. I don't get a % off (and often asks you to tip [0]). And (2) it actually presents a legal risk. If I genuinely forget to scan I can still be banned from a store, or worse, charged, because I am doing a job I am not trained for, or because of a glitch in the system.
> The problem with self-checkouts is that (1) it shifts the work to me without any benefit to me.
A well implemented self-checkout is much faster than waiting in line for a cashier. Ever since the local supermarket implemented self-scan I haven't had to wait for a register to become available for more than maybe 30 seconds.
Of course if the self-checkout is of the "UNEXPECTED ITEM IN BAGGING AREA" variety it'll be infuriating and much slower than a regular cash register.
> In civilized, high-trust societies self-checkout works well and is very convenient.
There's nothing convenient about me having to do the cashier's job, but badly, while having a tiny space to do my bagging in, while a computer yells at me.
Maybe for someone with truly crippling social anxiety, this is preferable to an actually functional check-out counter, but I'm blessed to not be afflicted by it.
And someone with truly crippling social anxiety isn't going to be standing there in the store with a line of people staring at them while a computer yells at them. They are gonna be sitting at home waiting for the gig worker to drop off their stuff.
