Tangentially related: I would like to report a low-severity security vulnerability in Claude (web version), but I can't be bothered to go through the Hackerone formalities, since I don't care about a bounty.
Right now I'm defaulting to "do nothing" because I'm lazy, but if any Anthropic staff are reading this I'm happy to explain the details informally somewhere.
At Microsoft-scale, data requests from law enforcement are an inevitability. Designing a system such that their requests are answerable is a choice. Signal's cloud backup system is an example of a different choice being made.
If you have good tests, certain types of change can be merged without manual testing. One problem specific to AI is that it has a tendency to game/bypass/nerf/disable tests, as opposed to actually making the code do the correct thing.
Not really. They said in their comment they deleted their whole account and everything. They probably don't want to continue to be ridiculed and to link the identity of that account with this one.
I think the main use-case for the metadata-enriched 300TB archive is training AI models like suno. Anyone torrenting music for personal consumption had higher quality sources available already.
I've bought hundreds of things on ebay over the years and I've never understood the issue with "sniping".
Sure, I've been outbid at the last moment. Losing an auction is always a little frustrating. But if I was willing to pay that price I should have bid it myself. Feels fair enough?
And I prefer to use sniping bots because they let me revise my bid all the way up until the auction ends. If I put a bid on something and then sleep on it and decide I don’t actually want to pay that much, I can lower my bid or cancel it. If I bid with eBay directly then I loose that flexibility. It has nothing to do with trying to outsmart people or be sneaky.
Question: what kind of fun you are referring to here?
Since, from the outside, it surely sounds like you get pleasure by inflicting some form of suffering on others. But that hopefully isn’t considered fun, is it?
The price, when between the seller's minimum and the buyer's maximum, is a zero sum game. So while this is definitely screwing with people, the seller gets paid more and the amount of suffering in the world shouldn't really change.
You are falling for the zero-sum fallacy and mixing categories on top of it.
Globally, wealth gets created, which leads to a positive-sum game, not a zero sum game.
On the other hand, if one quadrillionaire in a city owns all the money available in that said system except 100 currency units, the remaining 100 humans are in possession of exactly 1 currency unit. The suffering for the 100 humans is significantly higher for the 100 than for the one, even though it fulfils your premise of a balanced global suffering index.
Before the trade, the value for the seller and the buyer was zero. Whatever the trade involved, the moment the minimum of the seller gets hit, it becomes a positive-sum game.
If this would not be the case long-term rise of stocks would be impossible. That would mean a stock rise is a redistribution and you take it away from someone else . So, if the stock market were truly zero-sum, every currency unit earned would require someone else to have lost one.
I am not having zero sun fallacy. Please read what I said again. I said the exact price is zero sum within the bounds of the deal happening. The wealth creation is caused by the deal happening at all.
> if one quadrillionaire in a city owns all the money
That's a valid risk factor but on a random eBay purchase I think it's fair to say we have no idea if the purchaser or the seller gets more utility out of each dollar.
Then we actually agree on parts? Well, excuse me if I interpreted you wrong.
>we have no idea if the purchaser or the seller gets more utility out of each dollar.
Assumption: the seller opened the auction with his actual hard lower limit, he should be happy with what he gets as soon as that limit gets hit.
The original poster said that he essentially altered the bid in favour of the seller. However, the exchange of subjective equal values is based on the balance between the two parties and now gets distorted in favour of the seller and in detriment of the buyer. This should result in win/lose if I am not mistaken.
So, maybe I get you wrong, I am not sure right now.
My argument is that this distortion in favor of the seller isn't really good or bad in a meaningful way. It's just rude.
The seller is happy as long as their limit is hit, the buyer is happy as long as their limit isn't hit. How should the surplus happiness get split? I dunno. So the earlier poster sticking their finger in and shifting the surplus around isn't a particularly moral issue.
why would you bid the highest price you can afford in an auction? the seller agreed to auction the thing; they could have just offered it for a set price.
Do you not know how ebay works? You put in the maximum price you're willing to pay, and if you win you're paying 2nd highest bid + 1. So you don't save any money by starting with a low bid.
From what I've seen discussed, it seems some percentage of "sniping" is to attempt to obtain both "winning bid" and "lowest possible price" (note, not the same as "max willing to pay for the same item"). The sniper is trying to hide interest, so as not to attract other interested bidders, and therefore grab "a great deal" of a small increment above the starting bid price.
And this probably appears to work enough times in the snipers favor to trick them into thinking it is a winning strategy, whereas they likely would have won the same auctions in the end by just bidding that 'minimum' as their maximum bid. But as they can't easily (i.e., without expense) A/B test their strategy, they get no feedback that sniping isn't really helping them like they think it is helping them.
> But as they can't easily (i.e., without expense) A/B test their strategy
There also isn't really any detriment. At worst, the sniper is making the same bid they would have made otherwise. If the opposing bidders are not purely rational, and have not put in their actual maximum bid, then sniping can deprive them of that opportunity and thus lowers the hammer price.
And bidders are not purely rational, especially when the items are not purely utilitarian. Getting notifications that you have been outbid has an emotional effect, as does having time to think about raising the bid.
they notify the bidder when they're outbid, and the incremental price increases can make it tempting for someone to adjust their idea of their max price. sniping deprives them of that opportunity.
I have to wonder if they really believe half this stuff, or just think it has a positive impact on Claude's behaviour. If it's the latter I suppose they can never admit it, because that information would make its way into future training data. They can never break character!
Remember when Google was "Don't be evil"? They would happily shred this constitution and any other one if it meant more money. They don't, but they think we do.
I'm not sure how effective this will be. A lot of AI-generated bug bounty reports are pure spam, but a significant fraction are well-meaning humans who genuinely believe the nonsense an LLM has given them. The former category do not read the rules in the first place and will not be deterred - spray and pray is their MO. The second category will not believe that any "no slop" rules apply to them, because they genuinely think their bug is real.
Right now I'm defaulting to "do nothing" because I'm lazy, but if any Anthropic staff are reading this I'm happy to explain the details informally somewhere.
reply