zer01
Joined 14 years ago
Joined 14 years ago
It’s pretty bonkers TBH, I’d love to understand the abstractions and data models involved with software engineering this ambitious.
Guns may be “easy to liquidate” (not sure if that’s actually the case) but surely selling Xboxes and MacBooks on eBay is just as profitable?
People work much better when they're happy and can be themselves with like-minded individuals. Companies like Google realized this in the 90s and ensured that they proactively have what they need to be happy (feeding employees, paying for healthcare premiums, etc) to great success. You see this continue with things like this https://news.ycombinator.com/item?id=31871581 IMO.
It's really easy to say that you as a business care about gay rights in order to have gay employees come work for you, but if you then contribute political donations to a politician who sees gay people as an abomination you can't seriously be surprise when people go "hey what the fuck".
This sounds like your position is that anyone who advocates for anything (esp. when it's directly impacting them, but unrelated to work) only has one singular goal of weeding out disagreement and getting rid of it, which I think is ridiculous.
Humans are social creatures and we largely shape our perspectives by iterating with new data and communication. People can be (and IMO largely are) advocates for things that are important to them, have opinions on things, and don't have a singular focus of destroying anyone with different viewpoints.
On the other side, all humans have their limits to cooperation. As you have social interactions with anyone your viewpoint of them change, and realistically if I find out something about a co-worker through discussion like they believe that the 2020 election was stolen because Trump told them so it does inevitably make me scrutinize this person harder. What if you found out that a co-worker thought black people were genetically inferior or that women don't deserve to vote? At that point it's hard to work with someone with such a fundamentally different view of the world.
Is it "cancel culture" to say "we don't want nazis to work here"?
I also find as a dev that iPhones and MacBooks “just work” and let me get my stuff done, much much more so than a Linux desktop environment or windows laptop.
I guess we may be talking about different things though if you’re saying the apple ecosystem isn’t living up to its full potential, but IMO the rest of the phone/computer ecosystem is full of cheap shitty plastic crap with windows installs that OEMs have had free reign to install whatever they want on (and even without that, why tf when I install windows 10 do I get ads for “candy crush” on my start menu?!)
When planned leaks like this occur for real (or black hats are trying to prove possession of data they shouldn’t have like credit card dumps) there’s almost always a small cut of the data that’s released to give legitimacy to the actor and the dump.
This looks much more like someone trying to wateringhole attack people or extort crypto influencers/funds.
To take a subtle bug in something like a physics engine and bend it to your will to eek out another 5s save off the world record is absolutely fascinating.
Not to mention things like TASBot which always makes a great GDQ watch https://youtu.be/7CgXvIuZR40 and got me interested in speed running in the first place!
Clearly Tesla and Netflix were speculative assets without intrinsic value too by that logic.
I’m not sure if you’ve ever heard of the “lightning network” but it uses some really clever cryptography which allows people to send/recv Bitcoin almost instantly, with finalized settlement done on-chain as a normal transaction when the payment channel closes https://en.m.wikipedia.org/wiki/Lightning_Network
The other thing about payment finality is that thanks to blockchains with smart contracts (so not really on Bitcoin) things like escrow become pretty simple to conduct. You can write programs to enforce exactly how an exchange should work (“only release payment after proof has been posted to the chain”), and both parties can have confidence that no one is getting taken advantage of because they’ve agreed to immutable rules beforehand - to me that’s a super powerful primitive to build neat stuff with.
The flip side is that unfortunately the stability of the USD is in the hands of politicians and other stewards of the country, which is an immense power to wield, and highly susceptible to corruption and misaligned incentives.
The idea behind cryptocurrencies like Bitcoin is that the code and protocol are simply open source computer code. China can’t arbitrarily decide that they have more bitcoins than they actually have, even if they dedicate an enormous amount of power and influence to try to do so. It provides a tool by which people can transfer value even with hostile adversaries. There’s also a fixed supply that will ever exist, and while deflationary money clearly has issues when it’s the only currency used, as a counterbalance to politically-driven inflationary money it seems like a useful hedge, even more so than gold (we have estimations for how much gold exists but we can’t ever say with certainty - but you can search literally the entire galaxy for more bitcoins and you won’t find them).
The only question I have is around your point on monorepos - every monorepo I’ve seen has been a discoverability nightmare with bespoke configurations and archaic incantations (and sometimes installing Java!) necessary to even figure out what plugs in to what.
How do you reason about a mono repo with new eyeballs? Do you get read in on things from an existing engineer? I struggle to understand how they’d make the job of auditing a software stack easier, except for maybe 3rd party dependency versions if they’re pulled in and shared.
Decentralized finance products like uniswap and compound are not the same as micro loans, and micro loans have massive issues with trust, fraud, and solvency. Being able to swap or lend assets in a completely trustless way is something new, and enables new and novel technology like flash loans to exist (which make for more efficient markets). Plus I can go read the actual holdings of these protocols and build whatever machinery I want to respond to changes. Both protocols have a governance process with a timelock so no changes are made without the market having time to react, which overcomes a lot of the risk of malicious proposals (if I somehow pass a proposal that says “pay me all the money in the contract” then everyone will pull their money from the platform before it executes).
As for MPC I’m taking more about cryptographic MPC than something like SETI. A HUGE problem with managing cryptographic keys, be they root dns keys (13 of which control essentially the entire internet) or keys for SSL certificates is that they must exist somewhere in memory to be used. If they’re in memory they can be stolen and no one can actually tell if they’ve been stolen or not until it’s too late. Right now we have ok solutions to some of this through HSMs but then how do you also do backups properly? The MPC research I’m talking about specifically offsets these risks by never having the key exist in one place, and instead it being similar to a multisig where participants come together to create this material. An added benefit depending on how you set this up is that it only requires 1 party to be honest for the whole system to remain sound. ZCash did their entire setup for their protocol this way (https://z.cash/technology/paramgen/). Coinbase also just rolled out support for doing this at scale so that they can have users participate in these crypto protocols without having to manage their own keys (https://cointelegraph.com/news/coinbase-unveils-web3-mobile-...).
So yeah I can see how if you’re a person who has done 0 research into this things sound similar on the surface, but you’d probably also be in the camp of “we already have the post office, what’s the point of email” back in the 90s.
> What is the use case for crypto. Do you have a vision for it?
Short term, the crypto space is building neat things like multi-sig contracts, decentralized finance products (with unique characteristics like flash loans), DAOs to govern things in a decentralized way via on-chain voting, and spurring innovation into real cryptographic research, including things like privacy-preserving crypto and multi-party computation. People are thinking very deeply about things like trust relationships and engineering systems to be trust-less.
Medium term, I see protocols like Compound eating the profits of the large banks and either forcing them to innovate (which would be great) or replacing them entirely. Why does my savings account yield 0.01% when they lend out my very same dollars at 5% in the form of car loans/mortgages? I'm not discounting FDIC insurance or the federal reserve, but things have tipped too far into the banker's favor IMO. Same deal with the traditional markets, there's a lot of fees and friction unless you're a big boy that locks a lot of retail out of the game. Private capital markets are in for a wake up as well - VCs hold a crazy amount of power and potential upside exposure because they're just the only ones who can get in on deals early (think early Google investors) - if that was more access-able we might be able to break the VC hegemony which loves throwing good money after bad and propping up terrible companies.
Long term, I see it as a series of tools and protocols that can establish a more "global-native" economy. Our world is getting more global, not less. A core foundation of immutable software feels like a more reasonable base layer to build on than trusting politicians and a gordian knot of deals. Someone from North Korea can build a smart contract and I don't need to trust them that it does what they say it does, because I can verify it for myself. Also economically, having a store of value that is also impossible to counterfeit is hugely useful, so people can't paint lead yellow and call it gold - https://www.reuters.com/article/us-china-gold-kingold-jewelr....
The graph protocol - decentralized blockchain indexing feeding a network of graphQL servers hosting the data powering dapps https://thegraph.com/en/
Chainlink - decentralized on-chain oracles for basically any piece of data from the real world - https://chain.link/
MakerDAO - creator of the first “soft pegged” stablecoin, it is a series of smart contracts that allow people to lock up crypto as collateral and mint DAI, uses a number of neat and novel mechanisms to keep peg by adjusting interest rates and incentivizing economic behavior https://makerdao.com/en/
Compound - a decentralized lending and borrowing protocol with no centralized authority handling liquidations/collateral pricing/etc. https://compound.finance/
Uniswap - a decentralized exchange for swapping assets in a trustless way, has neat features like flash loans where you can borrow a large amount of money (millions of dollars) for the length of one transaction, as long as the money is returned with a fee. This allows anyone to do things like capture arbitrage opportunities, liquidate large positions, or whatever else can be done in one transaction. https://uniswap.org/faq
A lot of these are DAOs using smart contracts to propose and vote on binding proposals on chain and adjust parameters within the protocol or paying contributors and whatnot. You can see an example here https://compound.finance/governance
Beyond that there’s other blockchains focused on specific use cases like GameFi, bridges connecting together different blockchains and moving liquidity between them in a mostly seamless way, massive research attention being paid to zero knowledge proofs (which have a lot of potential implications in the real world), and multi-sig contracts for only doing things when a quorum of key holders come together and signal their intent.