text stringlengths 0 1.36k |
|---|
[419.02 --> 419.20] Yeah. |
[419.96 --> 423.30] So, you know, if you're future proofing, you know, don't use shot one. |
[423.30 --> 426.62] Or or past proof. |
[426.72 --> 429.30] If you're just proofing at all, don't use shot one. |
[430.04 --> 430.46] Yeah. |
[430.54 --> 434.84] There's shot two fifty six, which is essentially exactly the same with much higher entropy. |
[435.48 --> 437.02] So just use that instead. |
[437.72 --> 437.86] Yeah. |
[437.94 --> 442.12] I've actually become a big fan of multi hash. |
[442.48 --> 443.50] Have you ever heard about this? |
[444.66 --> 444.84] No. |
[445.14 --> 445.78] OK, nobody has. |
[445.78 --> 449.64] So one has been been pushing this really hard for quite a while. |
[450.14 --> 451.84] He is one of the people behind IPFS. |
[451.84 --> 455.56] So lots of kind of distributed peer to peer crypto stuff. |
[456.12 --> 460.20] And he's really wanted to kind of future proof everything that he's been working on. |
[460.30 --> 463.74] So he started this little open source project called multi formats. |
[463.74 --> 474.24] And what these are is essentially, you know, every time that you've got to sit down and use a codec or you've got to use like a particular encryption algorithm or a hashing function like this. |
[474.24 --> 484.50] Let's just create a format that allows you to define which format you're using so that libraries can just, you know, optionally support a bunch of different formats. |
[484.70 --> 487.54] And if in the future you want to change formats, you don't break all of your clients. |
[487.66 --> 487.96] Essentially. |
[488.70 --> 496.66] It's like very similar to MKV or MOV, but all the container things for video codecs, I suppose. |
[496.66 --> 497.06] Right. |
[498.18 --> 498.32] Right. |
[498.54 --> 501.96] Although containers oddly actually do implement a bunch of features. |
[503.00 --> 503.14] Sure. |
[503.52 --> 506.32] This gets really ugly, actually, in codecs and containers. |
[507.88 --> 519.78] But yeah, I mean, so they're like for multi hash, for instance, there's libraries and pretty much every language ever, including a very well maintained JavaScript implementation that works, you know, in the in the browser and in Node. |
[519.78 --> 522.68] So that's that's what I've used in a couple of projects recently. |
[522.88 --> 527.76] But Lenis, the funny thing is that Lenis is actually like still just not convinced. |
[528.58 --> 538.66] So he's basically said that, you know, well, the way that Git uses it is still not prone to these attacks because, you know, they have the length of the body and that makes the this harder. |
[539.78 --> 542.42] So, you know, we'll see how that ends up. |
[543.20 --> 545.22] I mean, it does make it a lot harder for what it's worth. |
[545.28 --> 545.66] It does. |
[545.66 --> 548.44] It does make that it does make the attack a lot harder. |
[548.62 --> 557.98] But like I do feel like he rather than future proofing or moving to like just a better algorithm, he's just kind of dangling out this like, oh, prove me wrong. |
[558.24 --> 559.08] Computer science. |
[559.84 --> 560.02] Yeah. |
[562.30 --> 567.44] Which which like didn't work out that well for, you know, his last round of this. |
[567.86 --> 568.08] Right. |
[568.08 --> 568.16] Yeah. |
[568.64 --> 568.80] Yeah. |
[568.86 --> 573.26] It seems silly to be like, well, you only half broke it. |
[573.32 --> 580.58] So I'm going to like if you got through a half of my lock, I'm going to go ahead and just change out the whole thing. |
[581.16 --> 581.52] Yeah. |
[582.00 --> 586.36] Rather than just like, no, I'm going to start just, you know, continuing to put these half locks on things. |
[586.36 --> 586.76] Yeah. |
[589.08 --> 589.44] Yeah. |
[589.56 --> 597.70] But this kind of reminds me that like the way that we think about security on the web tends to be like, oh, I put Cloud Slayer in front of it. |
[597.80 --> 598.94] So I'm secure now. |
[599.84 --> 602.64] You know, like I added SSL or added TLS or something. |
[602.74 --> 603.52] So I'm secure now. |
[603.52 --> 612.26] But really, security is like this really multilayered thing where when you when you break off one layer of the onion, you need the other layers around it to still be secure. |
[612.42 --> 612.58] Right. |
[613.60 --> 613.96] Yeah. |
[614.08 --> 617.08] I think you almost can't even break off any of the layers. |
[617.08 --> 620.56] I mean, yeah, security is really, really hard. |
[620.68 --> 626.14] It's just it is it needs to be there at every layer. |
[626.64 --> 628.94] Otherwise, the other ones just have no effect. |
[629.10 --> 630.08] You know, right. |
[630.52 --> 630.74] Right. |
[631.14 --> 638.40] Like you're it's I think an onion is a poor metaphor, I guess I'm saying the chain or whatever is much better. |
[638.52 --> 641.28] Like if you have a single weak link, then it doesn't matter. |
[641.38 --> 642.44] You can get you can get through. |
[642.78 --> 643.18] Yeah. |
[643.38 --> 651.08] I mean, if you look at like some of the stuff that people have been doing with auth for a while, you know, OAuth jumps through all of these hoops to basically do an extra layer of encryption. |
[652.32 --> 657.02] And initially they kind of did that so that you could do OAuth over HTTP without TLS. |
[657.02 --> 663.00] But when but like even when you added TLS to it, it's really nice to keep all that encryption there. |
[663.28 --> 666.04] And one of the things that OAuth 2 did was it just kind of got rid of that and was like, oh, whatever. |
[666.10 --> 666.84] We're using TLS. |
[667.62 --> 669.86] But like, you know, you can break TLS. |
[670.22 --> 675.68] We know that certain authorities have been compromised for TLS so people can give out bad certs like that. |
[675.86 --> 678.62] That's not a very good way to secure everything. |
[678.62 --> 679.60] Well, sure. |
[680.04 --> 686.04] I mean, if you operate under the assumption that TLS is broken, though, then the entire Internet's broken already. |
[686.44 --> 692.30] So it's like the auth channel, if you had that extra encryption, would be broken. |
[692.42 --> 696.32] But then as soon as you got to that website and used it, you'd be screwed anyways if TLS is broken. |
[696.64 --> 700.38] So I don't think it'd be any more broken than it would already be, I guess. |
[700.38 --> 702.92] So you're screwed if that if that's the case. |
[703.68 --> 705.00] No matter what. |
[705.42 --> 710.56] Maybe someone doesn't get your authentication credentials, but hopefully you haven't reused those anywhere else. |
[710.84 --> 710.86] So. |
[712.88 --> 713.54] All right. |
[713.58 --> 716.40] Well, I mean, do we have anything else to say about hashing algorithms? |
[717.90 --> 721.18] This is this is a pretty deep topic to start a JavaScript show with. |
[721.30 --> 721.62] Yeah, sure. |
[721.62 --> 723.60] You know, interesting choice. |
[723.60 --> 735.92] So like somebody that doesn't know anything about this kind of stuff, a.k.a. me or someone else that doesn't necessarily have to deal with the security side of the code that they write. |
[736.26 --> 740.22] What would be like the best thing for them to like? |
[740.22 --> 752.20] What would be the best resource for somebody that wants to know, like how to actually authenticate stuff in a secure way that I don't know, wouldn't anger Linus or Linus, whoever you say his name? |
[752.20 --> 755.24] I don't know if I'm answering your question directly. |
[755.62 --> 763.98] But if you're building a website and you want to make sure your website is secure, Mozilla Observatory is a really good option for like it will scan your website. |
[764.10 --> 767.60] It'll check your TLS certs, which is some of this is involved there. |
[767.76 --> 774.00] And then it'll it'll check content security policy, a bunch of different things. |
[774.02 --> 777.90] It'll kind of give you a prioritized list of things to do. |
[777.90 --> 787.18] So I would absolutely recommend like putting any website you build through Mozilla Observatory to kind of get that checklist of and score and things like that. |
[787.74 --> 788.38] Cool. That's awesome. |
[788.48 --> 790.08] I didn't even know about that site. |
[790.20 --> 790.98] So that is helpful. |
[790.98 --> 797.92] I think also like maybe we can call out a couple good like application layer authentication schemes as well. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.