jgeralnik

A friend’s PhD advisor has been chasing non-sofic groups for 25 years (and was shown a preprint of the results by openai to verify them). He believed a solution would be Fields-worthy

This was not a problem that was for sale

The wonderful thing though is that you can just run the model multiple times (even in parallel). Some instances might get stuck but as long as some find the bug and you have a good way to filter outputs (e.g. with another llm that tries to create concrete exploits) even a very small success rate on stage 1 can lead to reliable exploits

I think there’s something very interesting here and would be interested in hearing more about the date discrepancies- it’s a shame the article is mostly just the raw output of gemini instead of more commentary

Ah, fair enough. Anthropic caches at a block level (basically a single message) so for non-trivial messages this is really less of a concern, although I definitely understand why they still scope cache to a single tenant

Anthropic requires explicit cache markers but will “look backwards” some amount, so you don’t need to fall on the exact split to get cached tokens

Right, you can’t actually guess a letter (byte) at a time but you can guess a token at a time (I believe the vocabulary is 200000 possible tokens in gpt 5) So you could send each of the 200000 possible tokens, see which is cached, and then send 200000 more tokens to find the next cached token Certainly less efficient but well within the realm of a feasible attack

A remote code execution bug in ios is valuable - it may take a long time to detect exploitation (potentially years if used carefully), and even after being discovered there is a long tail of devices that take time to update (although less so than on android, or linux run on embedded devices that can’t be updated) That’s why it’s worth millions on the black market and apple will pay you $2 million dollars for it

An XSS is much harder to exploit quietly (the server can log everything), and can be closed immediately 100% with no long tail. At the push of an update the vulnerability is now worth zero. Someone paying to purchase an XSS is probably intending to use it once (with a large blast radius) and get as much as they can from it in the time until it is closed (hours? maybe days?)

It actually sounds like they just want a switch

Ruff and uv are written by the same company, astral

Thanks, those were the main recommendation the last time I looked into it (a few years ago), good to hear you recommend them too!

What DDR pads are those? Are they custom made?

No, but it's still a ridiculous YoY at this scale. Wiz is the fastest company to ever reach $100M ARR and still have a lot of growth left in them.

They have crazy growth. They reached $100M/ARR in 18 months, and 3.5 times that in less than 24 months more, so around 85% growth YoY. This is a purchase for future potential, not current earnings

It seems you are thinking of the crypto exchange as selling crypto to customers and then "holding on" to their money. That's not the right model.

Party A deposits 1 BTC into the exchange. Party B deposits $1000 into the exchange. Party A wants to sell a BTC for $1000 and party B wants to buy a BTC. They trade through the exchange, pay some fee to the exchange, and an entry in a database is changed such that the $1000 in the exchange now belongs to A and the 1BTC in the exchange's wallet now belongs to B.

Since actually holding cryptocurrency is inconvenient for users, many will just choose to keep the BTC on the exchange until they want to use it/sell it for cash or a different cryptocoin. Many crypto users are speculators who view it the same as holding stocks in brokerage accounts and not as just an exchange.

This money is held in the exchange platform but belongs to the users. They should in theory be able to withdraw it whenever they want.

Instead the crypto exchange decides to make use of these idle customer funds and invest in speculative funds/embezzle all the money and all of a sudden there is not enough funds in the exchange for all users to withdraw.

MD5 is not broken in the way you think it is. Even without salting if you give a hash H there is basically no way to recover a string that hashes to it other than randomly trying strings 2^128 times.

What you can do trivially is find 2 strings X1 and X2 such that md5(X1) == md5(X2). In this case seeding the way you described won't help because md5(X1+S) will equal md5(X2+S) due to the way MD5 works

Yeah I meant proof of stake, not proof of work

A bunch of nodes generate random numbers and publish their hashes. After all hashes have been published all nodes reveal their random number, verify that their number actually corresponds to their hash. Xor all the numbers together and as long as at least one node was honest and generated a random number the result will be a random number. I don't recall if this is exactly what proof of work does but I think it's pretty solid.

Tailscale can punch through NAT but requires outgoing internet access

Ah, I don't think that's exactly what this thread was about. Ignoring how authorization works, the question was whether there is an advantage to encrypting your commands again (via say ssh) vs. just sending them in plaintext under wireguard (via say rsh)

But that's the thing - even with double encryption tailscale and your SSO can run commands on your machines

1) Run tailscale --ssh on your server 2) A malicious SSO or tailscale add a new machine to your network and update your ACL such that the new machine can connect to your server 3) ssh from the new machine to run code on your server

The fact that the connection between the malicious machine and your server is double encrypted doesn't affect the attack here at all

I didn't mean binary as in "binary blob" but as in "standalone program separate from the tailscale program". Sorry for any confusion :)

So @bradfitz when are you releasing https://tailscale.com/connect/ for real? :)

Context for the uninitiated - as a crazy idea on the podcast Security Cryptography Whatever (hosted by tptacek and others less well known on HN) Avery and Brad of tailscale imagined an ssh client in the browser with QR code authentication to SSO to allow you to connect to your tailscale network (over tailscale SSH) from untrusted computers such as internet cafes. (Or mostly untrusted - safe from keyloggers but maybe not from a dedicated active malware that injects into your browser and tries to inject secret commands into your ssh session).

I created a silly PoC here (video instead of link because don't try it for real) https://twitter.com/jgeralnik/status/1487913797155233798 back when tailscale ssh was a secret binary in the tailscale github repo

Double encryption doesn't actually help in that case though - if tailscale (maliciously) added nodes to your network the ssh session being encrypted wouldn't change the fact that they can run commands on your machines. And if they wanted to actively MITM you they could do so (by redirecting your wireguard connection to a server owned by them) even with encryption (presuming they can fake the host key, which they could do at worst by running code on your server to steal it).

To be clear I implicitly and explicitly trust tailscale not to tamper with my networks and if your threat model includes tailscale becoming a bad actor you should remember that in that case running their binary in the first place could already be game over.

For what it's worth I encountered the same issue and came up with a solution:

https://github.com/cloudflare/cloudflared/issues/574

Cloudflare have ignored the github issue (which includes a solution) but at least 3 other people seem to have found my solution helpful.

You can't have a string of 4242T zeroes immediately after the first 101T digits but presumably you could have them sometime later (say after another 10000T digits)

The second population contains 100 people without the disease. Of them we expect one person to have a false positive. In addition the person who actually has the disease will test positive. So on average 2 people will test positive, one of whom has the disease and one of whom doesn't. If you are one of those two people you have a 50% chance of your (positive) diagnosis being false.

IPv6 doesn't mean no firewall and NAT!=firewall

You should still be blocking incoming requests for IPv6 endpoints and only open ports you intend to serve publicly.

The author went out of their way to hide the factors of N, presumably so that nobody else would actually be able to generate signed certificates. However, they did this by hiding half the digits of the factors.

Revealing so many digits of the factors actually allows easily factoring the original number using a version of coppersmith's method (easy as in under a second on my laptop instead of the 9 hours on a distributed cluster the authors used). This is actually a pretty classic CTF exercise.

If I'm still nerdsniped by this tomorrow I'll try my hand at implementing this and factoring the number myself

Some more information - I found this YouTube channel which has videos of various hacks of Kongregate games. I don't think I ever saw it as a child but it brings back memories of various things we would do.

Here is a tool called Ragnarok that let's you edit values based on their name. You would open the reversed actionscript and could play with things based on the variable name:

https://youtu.be/B4GhD8Ets8U

Editing save files was big. The files were in a proprietary format called sol. All of the videos I see on that channel are just "download this save file with a completed game state", but there were tools for editing the save file yourself. Some games introduced obfuscation of data in the save file and the cat and mouse game began.

Oh, here's an example! Apparently the sol editor was a feature of Ragnarok, and maybe the disassembler too:

https://youtu.be/wLaMZoqI-_Q

Here's a video of Niflheim, seems similar to Ragnarok. Maybe Ragnarok replaced it as a better tool?

https://youtu.be/fgqHarxwv64

Other than that it seems most of the cheats on that youtube channel are using cheat engine. While I definitely used that, I don't recall it being part of my go-to arsenal.