That’s not the question. The question is “were they even there” before the “screenshot”.
As an example, I can screenshot your comment and have AI edit it to have you say “I really enjoy misleading people on the internet for money”, post it, and then say “I took this screenshot before they deleted it.”
But also just look at the bottom-left of the image as is mentioned elsewhere in this comment thread: they didn't even bother to photoshop the entire image in their "debunk".
Is there a way from within the VM, though, to determine if that hardware feature exists and is enabled? If you don't trust your VM provider, how do you trust them when they tell you they have that feature and it is enabled?
Depends on if you don't trust them not to actually snoop on you or if you don't trust them not to get compromised by a third party. In the latter case you'd presumably trust that they deployed the feature claimed.
> It's meaningless to say that selling something is profitable if you are ignoring the cost to produce it.
At the same time, training costs don’t scale with usage, so no matter how expensive they are, they can be paid for as long as inference is profitable. It is only inherently unsustainable if you lose money on inference.
Of course, it could be that the inference usage required to pay for the training is too high to be possible, but that is a separate argument.
It doesn't just matter that inference is profitable - cash from profitable inference needs to outweigh cash burned by training. If training costs even $1 more than your profits from inference on a model, you lost money on the whole thing.
The baddies also take a look at the CT logs and look for fresh wordpress instances that haven't yet been configured and will helpfully configure them for you within a few minutes.
Yes but a lot of non-browser software doesn't check CT logs, so if an attacker were to target such software they could use an unlogged cert and no one would notice.
The CT log is a property of the individual leaf certificate not a property of the issuing chain. You could have the same CA in the trust store issue an unlogged certificate and non browsers would trust it.
A CT log isn't super relevant (it's not the CRL, it's just a log), but it's maintained/delivered by the CA not the leaf. Never the leaf.
If you mean CRL/Revocation list, everything but the root [you can't revoke a root] is maintained by the tier above. Leaf -> intermediate (these), [Intermediate -> Intermediate*], Intermediate -> Root.
Not sure why you're mentioning transparency logs, they have nothing to do with this (nor help).
Because the thread: "a lot of non-browser software doesn't check CT logs", you replied saying it's a reason to use the OS trust store, not sure how that fixes things? Maybe it would relieve them of the responsibility of pushing out a distrust update?
That's not what woodruff is talking about (I hope). The CT logs allow us to (later) find who/why it happened. CT Logs in no way help you stop something. (Depends a bit on how software reacts to apparently invalid issuance - but since these have the real keys the issuance will look normal).
Revocation lists allow you to stop/block bad issues. Both are maintained/decided (you could post to another service) by the issuer, not the leaf. CRL records would stop this (but some software has decided that's too slow). In general software uses the OS trust-store first (fragmented on Linux). Firefox and Java have their own independent paths. Chrome ignores CRLs
CT logs only tell you something happened if the certificate was registered in the log. But if you are attacking software that doesn't check if the cert was registered, then you don't need to register the cert, and there won't be anything in the log.
My point is you can forge a cert, without putting it in the log, and a lot of software will still be vulnerable to it.
Curl does not check that certificates have a valid SCT attached. Thus it will accept a certificate that isn't in the CT logs. That's mostly independent of the OS trust stores.
There is nothing stopping us from adding the ability to understand the context of information an LLM consumes, too, right? The same process humans use to assess data reliability could be taught to an LLM?
Maybe. With enough money and resources and time, maybe. But it would need to be very reliable to be at all useful, and that level of reliability is very hard to achieve with something so nuanced and complex.
You have an entire lifetime of experience and the wisdom of countless generations to inform you of the context with which you should process new information. You were specifically trained for years to be able to do this somewhat effectively. Doing the same for an LLM may or may not be possible. They contain much of that experience in some form, but are not well equipped to make use of it iteratively. It is incredibly expensive to attempt even a simple and ineffective version of it with our current methods.
If the LLM can achieve even 90% accuracy, it becomes more of a hindrance than a help. Do you make use of the function and trust it successfully understood the provenance of new information? If so, ten percent of the time you will be misled, which could be disastrous across a population. If you don't simply trust it, then do you painstakingly check the provenance of everything it tells you? Of course not, you already don't do that with current search engines and blogs and such. Yet falling somewhere in between would be so much worse than simply not having the LLM attempt to discern trustworthiness from providence in the first place.
But all of that is opinion and I may be incorrect. Perhaps someone will create some ingenious way of imparting wisdom and discernment to LLMs next year and this will all be trivial.
If real people are using the model for real, they will discover for themselves that the data isn’t reliable, and then might switch to a different model.
The incentive will be to maintain a subscriber base.
> If real people are using the model for real, they will discover for themselves that the data isn’t reliable, and then might switch to a different model.
Insanely optimistic view to take. I know a handful of people who make this effort in my personal life, including me, and we tend to be treated as paranoid for caring.
We have laws against undisclosed and misleading advertising, because consumers can't make informed purchases otherwise. What makes you think LLMs will have a fundamentally different pattern than everything else?
What? I have spent $10k on a computer but would never fly business or first class, that shit is way too expensive. Two trips would buy a new $10k computer!
but you're in a position to understand now, that people who do fly business everywhere, can also afford one of these laptops for a fraction of what they spend flying. that's the use case.
my style of discussing things (let's call it "give credit to your interlocuters") doesn't work on this site.
this discussion is filled with people who love high performance computing, and they are mostly saying "I wouldn't buy this" and the reason "too expensive" and they compared it to "business class". my comment was in response to a sharer of that sentiment. (I picked onesuch comment since there is no way to reply to all of them) and pointed out "if you come at this from the other direction, you can see a population of people who will buy it, business class that you consider to be overpriced is full".
>That doesn’t mean the people that buy this laptop fly business class
I didn't say it did.
you should look at the totally of comments and consider "in what way way can this comment be considered correct" rather than looking for ways to say things are incorrect because you didn't anticipate what was going to be said.
reply