dicethrowaway1

3 mesi fadiscuss

jetbalsa said that the cache is on Anthropic's server, so the encryption and decryption would be server-side. You'd never see the encryption key, Anthropic would just give you an encrypted dump of the cache that would otherwise live on its server, and then decrypt with their own key when you replay the copy.

dicethrowaway1

3 mesi fadiscuss

Maybe they could let users store an encrypted copy of the cache? Since the users wouldn't have Anthropic's keys, it wouldn't leak any information about the model (beyond perhaps its number of parameters judging by the size).

dicethrowaway1

7 mesi fadiscuss

For email, I've had some luck just modifying the page with JS that's either indirect or obfuscated enough that the address can't be pulled directly from it - e.g. "var email" is the address encrypted with a fixed key, the JS decrypts it and then alters the HTML.

It can obviously be bypassed by using a JS runner, but it seems to be enough of a hurdle that few spammers bother. "You don't have to outrun the bear", as it were.

dicethrowaway1

anno scorsodiscuss

FWIW, o3 seems to get to the point more quickly than most of the other LLMs. So much so that, if you're asking about a broad topic, it may abbreviate a lot and make it difficult to parse just what it's saying.

dicethrowaway1

anno scorsodiscuss

This is a deepity.

The trivial interpretation is: every word written can be constructed by optimizing a prediction based on current state, what has been written so far, and a sufficiently complex model. This is true of anything computable: just make the method implicitly contain the program by assigning a high probability to any token that is consistent with running the computation one more step. It's also true of anything expressible: just brute force a solution that can be expressed in n words, then assign a high probability to the first word of these n words.

The profound but wrong interpretation is that intelligence is just statistical prediction according to some general-purpose algorithm, and that this algorithm is tractable. Consider something like solving a SAT problem. You're going to have a hard time using any tractable general-purpose algorithm to predict whether x_2 is true for the satisfying solution based on some long CNF statement plus "x_1 is false".

Now, what you _can_ do is augment your model so that if the previous tokens constitute a CNF-SAT instance plus a partial answer, then you cart these off to a SAT solver and output its next token. But the more you do this, the less force the "mere statistical prediction" part holds. The "next-token predictor" is just an interface to an assembly of different approaches; and often, these approaches (like the SAT solver) will output the whole solution all at once for free.

comments