homakov

I was a 100% bitcoiner 2014 to 2020, but after exploring defi and L2 projects in pipeline, i decided this time is different and converted to eth. That's just my opinion.

This is true, but that's not the only post I had on the subject, just the earliest one. Anyway, I gave up on btc and now implementing xln on eth https://github.com/homakov/xlncontracts/blob/main/contracts/...

It's better explained in solidity:

https://github.com/homakov/xlncontracts/blob/main/contracts/...

it is withdrawn from a special intermediary "reserve" balance OR a Debt is created on this identity.

Payment channels, in traditional sense, are fully collateralized credit. I suggested to leave collateralization ratio up to users themselves. Some could have 100% collateral, some 0%, some 50% etc.

Rejection reason: unclear, doesn't fit the narrative.

Bitcoin dominance is fading this year only because it has bet on LN with fundamental inbound capacity problem. LN rejected my proposal to solve it and extend channels with credit lines: XLN https://medium.com/fairlayer/xln-extended-lightning-network-...

Also many of these design attacks still work http://sakurity.com/oauth

I don't agree with calling flash loans loans. Money isn't borrowed, as you don't posses them even for a second. Instead it's more like "offer me an atomic transaction that will make me money, and I will run it with my liquidity". See: i'm not giving you my liquidity, i merely choose to use your strategy which is atomic, and you cannot change your mind and steal the money halfway (as in borrowing).

> set of UTXOs as collateral, and the moment any of them move, which is easily monitored

not possible. Once collateral is moved, it's moved. You can't even call it collateral if the other person can easily take it away in front of your eyes. It's only collateral if it's locked up and has a dispute period (see: payment channels collateral)

>There are some protocols emerging which lets you keep ownership of your coins and still lending them out.

how is it possible? When you lend your assets, someone can give it to other person. You cannot keep ownership of something others control?

Why didn't npmjs/rubygems just check failed lookup requests for "shopify-cloud" etc and block those for a while to prevent damage, and notify the companies (doing their best)? Seems like low hanging solution.

You never need to download 300gb, starting with 3gb snapshot is just as secure.

any estimation when is it out? should i wait or get 13 inch version now?

It was really that easy. Kudos. Others should follow

I see many people in this thread dread existing auth scene and ask for a better solution. Client certificates, but usable and no hardware - https://github.com/sakurity/securelogin this is the only sovereign answer to auth battles.

This can also protect the server from attacks if it runs all sessions in isolation. All you can is click and type, no more sqli or xss that require messing with params or API. So it could be useful either way.

No, no one should own a yubikey.

It's an entirely useless device. All you need is a pw manager that saves you from non-malware attacks (email compromise aside). Yubikey cannot save you from persistent malware, which makes it useless in almost all scenarios. The only hardware device that makes sense is the one with a screen (like trezor). Simple click-to-use devices carry no protections that you wouldn't otherwise get with a pw manager.

still have to care about Safari, where one man can prevent all communication to localhost for years. with no real reasons, breaking the spec.

https://bugs.webkit.org/show_bug.cgi?id=171934

Timing attacks, never saw a real world exploit working for them. Even in my local tests without web server latency the results are indistinguishable.

Did anyone ever show a working timing attack for a web service? Even assuming no DoS/request limiting is in place.

Do not have business with any company that pays taxes in Belarus. We must deprive Lukashenko of money. Many IT companies are attracted to Minsk because of the low cost and here we are.

What happens in Belarus is absolutely disgusting. Something you would expect from a mexican narco cartel not a european country. What’s even more disgusting is lack of any tangible help from the US/EU. No sanctions for Lukashenko they said.

Berlin airport, haha

In context of the US maybe it is different. Mine is 200mbit/s no caps $7/mo connection - common in Moscow - and I don’t remember caring about traffic or speed ever. Any torrents with 5+ seeds is like whoooosh and it’s done.

Okay than just my two cents. I use torrents a lot and never had a problem finding what i needed, and i would certainly not pay a dime for some obscure per-bite streaming. Not in torrents.

Something in Outer Wilds style of exploring, would be epic

This is so cringeworthy though

Didn't experience this problem much.

If not found on any tracker, such a material is probably too rare for anyone to care about. Even with some monetization model in place, it would be unrealistic to keep all sorts of "rare" materials in the world to make $1 from a random stranger coming once a year.

Anyway, I would't pay $1

What happened to traditional definition of state channels?

It used to be just like payment channels, but with a bigger scope and more fancy state operations.

The trick with avoiding multiparties was called "virtual state channels" or something like that. When did it change?

I can routinely download any torrent in the world on top speed... What problem does it solve?

Torrents, as a poor choice for the "killer app", aside, state channels are a curious technology but most traditional implementations [fully pre-collateralized channels] have a design flaw where no one wants to put money to the end user, while happily accepting inward capacity from the senders. Haven't looked into this specific implementation, i wonder if it has a fix for this. "Virtual channels" sound relevant here.

I wonder can we do overlay encryption like this: some software encrypts audio and video stream from your camera, then apps like zoom transfer encrypted stream with some noise, and end user software tries to decrypt those streams. This way you can use any middleware app as long as both users have encryption software overlay. Just a crazy idea

true, email shouldnt be used when you can identify by unique id. I doubt the bug was even exploitable with most apps. Apple just paid magnitudes more than its severity.