There are three ways game engines can have large worlds:
1) Stream in constantly as you move across the world 2) Divide the world into multiple segments, load in between 3) A hybrid, with invisible boundaries between segments
(1/5)
Likewise, blockchain infrastructure can follow a similar model. Each segment can be tied to one chain, and track the physics & mechanics elements where ordering and consensus is important. Non-consensus stuff like simulation, animation, graphics etc. are done client-side.
(2/5)
May 19, 2023 • 7 tweets • 2 min read
When considering the type of hyperscale novel apps like an onchain game would require, my mental model in the past was fractal scaling: 1,000 validiums, settling to a zk rollup
But after further discussions, the design space is actually much wider open
polynya.mirror.xyz/EL69tiTzl05WO8…
The first thing I didn't consider is that in an hyperscale system like this, you don't need strong guarantees for a lot of things. So, it may be OK for certain chains to be PoA deployed by, the game or social host, for example. In other situations, there may be redundancy...
May 19, 2023 • 7 tweets • 2 min read
Doesn't matter who said it first at all, but for some pedantic fun, in my writings, the distinction was inspired by AMD - as they moved from monolithic dies (one chip) to chiplets (many chips) for massive increases in performance & efficiency in HPC and server
Of course, other silicon companies have followed suit - it's too good of an idea - with each of Intel's Ponte Vecchio HPC processor made up of 47 smaller specialized chips, instead of 1 like has been the norm for the first half a century
2/6
May 17, 2023 • 6 tweets • 2 min read
1) The bottleneck is demand, not supply 2) You need fractal scaling for certain novel applications
Both are true
1) Different chains have different demand profiles. For example, Ethereum L1 is definitely supply-bound; but L2s/other L1s are mostly limited by demand
1/5
For today's usecases like money, NFTs, financial apps, ENS etc. rollups are enough - true
2) However, certain novel applications like onchain games, autonomous worlds, social etc. are not possible with current architectures at scale and necessitate fractal scaling
2/5
May 11, 2023 • 5 tweets • 1 min read
I've mentioned multiple times smart contracts can be used to gamify storytelling, but how is it going to work?
It's actually quite simple - you lego-fy narrative beats, and let users construct it. We've certainly had methods like exquisite corpse, but these don't scale
1/4
Blockchain smart contracts let you have far more complex set of rules and story legos, but more crucially, a clear mechanism to order and form consensus
This can then be mixed in with other gameplay mechanics, and virtual worlds will gain value
2/4
May 10, 2023 • 9 tweets • 2 min read
It seems crazy to say this, given Arbitrum One is right up there with Tron and BSC as the top 5 chains after Ethereum and Bitcoin by economic activity, but based on recent false narratives it feels like the crypto crowd is *still* greatly underestimating Arbitrum
(contd...)
So, firstly, the $ARB token has abysmal token distribution, and I have zero investment in Arbitrum so this is *not* financial advice. Indeed, I'm comfortable talking about the technical merits of Arbitrum because of the above, because I'm explicitly shitting on the token
May 10, 2023 • 6 tweets • 2 min read
An actually economically sustainable protocol in crypto? Crazy talk
IIUC, Arbitrum DAO net earned 3,352 ETH ($6M). Cost of operation was 5,594 ETH ($11M)
So, here's a trick for L2s serious about challenging Arbitrum's dominance - undercut them
Firstly, by foregoing revenues, you're reducing fees for your users by ~35%. But you can also use token inflation to subsidize fees further, like L1s/sidechains are doing. For example, $ARB inflation of 1%-2% (fraction of most L1s) would be adequate for end user fees to be *zero*
Apr 25, 2023 • 6 tweets • 2 min read
I know y'all know this, but everything here is false
Even if an L2 has its own "consensus mechanism", it's only used for censorship resistance and liveness, *not security*. Even if the L2's consensus fails, your funds are still secured by and withdrawable to L1
Further, L1 scaling is clearly on the roadmap - mainly with statelessness and zkEVM. Both combined will scale the L1 10x-100x, on top of Moore's/Nielsen's Law continuously pushing up gas limits. But this will take time - Ethereum settles trillions of dollars every year,
(2/4)
Apr 21, 2023 • 6 tweets • 2 min read
Arbitrum One is getting pretty close to the point that I (risk-averse non-degen) would consider using it with assets I cannot afford to lose. Here's what more I need:
1) 11/12 instead of 9/12 Security Council for immediate upgrades; pause/delayed upgrade is fine with 9/12
1/42) Governance can reshuffle/expand SC 3) Governance whitelists dozens more provers 4) In case of sequencer failure, no need to wait 1 day, alternate sequencers (can be whitelisted by gov) can come online in a short time (e.g. 1 Ethereum epoch, i.e. 6.4 minutes)
2/4
Apr 21, 2023 • 8 tweets • 2 min read
I have discussed the ideas of hybrid apps and local consensus before. Here's an example - a Cairo game. The first thing to note is the game has its own governance, which is the defacto local consensus. So, it doesn't need Starknet and Ethereum to validate everything.
1/7
But even before that, the game can have an application, and this application can have two modes:
- Peer-to-peer (fully decentralized)
- Server
Players can choose their own, they can select a server, or they can process everything that's required on their device
2/7
Apr 21, 2023 • 6 tweets • 2 min read
CT has turned crypto into a Republican vs. Democrat fight. As it turns out, this is false. I looked through Coinbase's Legislative Portal and found a much more nuanced situation, as expected. In the House, 32 Republicans support crypto vs. 27 Democrats.
coinbase.com/public-policy/…
In the Senate, 4 Republicans strongly support crypto versus 3 Democrats; with 16 vs. 7 for moderately support. So, yes, in terms of politicians supporting crypto, things lean Republican, but only mildly so. So, why does CT think Republicans love crypto and Democrats hate crypto?
Mar 24, 2023 • 8 tweets • 2 min read
Arbitrum One during the airdrop frenzy was a textbook demo of how L2 fees work and why execution is the primary bottleneck
Arbitrum One targets ~7x L1 throughput, that was saturated, leading to gas fees spiking. EIP-4844 would have made no difference (no L1 fee spike)
(contd.)
It also hints at why post EIP-4844, danksharding etc, L2 fees component will dominate over L1 data fees
Here's an old post that explains most of this (note some of the terminology has moved on, e.g. "fractal scaling" did not exist then)
This is the first project by a major game developer using blockchain tech that makes some sense. To break down my interpretation of what's happening here:
For the most part, it'll likely be a traditional MMO with an economic focus like EVE Online, however...
The problem with EVE Online is there was never a conversion between in-game currency and fiat, to evade regulation. Crypto will allow CCP to skirt regulations while also have in-game currency to natively have value. This will definitely not solve EVE's problems with...
Feb 21, 2023 • 4 tweets • 1 min read
Two common misconceptions around ETH and why...
1) "Staking is just dilution protection" - this comes from misunderstanding ETH as equity 2) "Deflation is bad" - this comes from misunderstanding ETH as currency
IMO majority of the TAM for crypto is alternative reserve asset,
non-sovereign collateral money, whatever you call it, for which high inflation is *BAD*. Ideally, ~0 inflation, minimal viable staking rate (even w/ LSDs), and low volatility. ETH is well positioned to do the first, passable on the second, but needs to improve on the last.
Feb 20, 2023 • 7 tweets • 2 min read
ETH *is* ultrasound money. It's a silly meme, yes, but this is crypto - memes are the only things that work, and this one has more substance than most
There's more complexity to demand and supply for ETH, yes, but if sustainable fees > subsidy, it means ETH is undervalued
The key thing is *sustainable* fees - not projecting forward from unsustainable bull market degeneracy. Yes, degen casino is Ethereum's #1 product-market fit, but depending on your time horizons, it's better to consider sustainable fees. Do this by looking at general trends...
Jan 19, 2023 • 10 tweets • 3 min read
Nothing in this thread will be new, but worth reiterating:
spxdey is asking the right question. Now, some may be tempted to say "more scalability just adds demand" or "more small tx * lower fees = higher net fees" - but this is a common fallacy in the crypto space.
Currently, Ethereum L1 fees are high because there's extremely limited supply and high-value financial transactors don't care about fees. However, as blockspace supply increases, demand will fall off non-linearly. At S', the total fees are a fraction of total fees at S.
Jan 17, 2023 • 4 tweets • 1 min read
The ultrasound money cult is missing the forest for the trees
Tx fees are a low quality demand driver - extremely volatile, very seasonal with unsustainable spikes in mania times (btw, manias diminish as protocols mature)
Further, it has a very limited TAM relative to...
...higher quality (IMO) demand drivers like economic collateral
Ethereans must focus on making ETH useful as money (in its various forms)
What would you rather have? A $10K ETH at 1% inflation or a $2K ETH at -0.5% inflation (yes, there is a trade-off)
Common fallacies for companies/protocols/sectors experiencing exponential growth:
1) Extrapolating said growth
2) It's "more than..." what it actually is / "new paradigm" narratives
3) "We're still early" narratives
In reality, it's in the middle of the adoption S-curve,
and further growth will be greatly diminished
Bitcoin went through this in 2017, Ethereum in 2021 (and other non-crypto stuff like Tesla around the same time)
It's important to note scaling does not lead to adoption in $ terms, particularly for financially-oriented products
Jan 6, 2023 • 4 tweets • 1 min read
Financialization and fun are contrary & conflicting design goals for a game. Applying Unreal Engine lipstick on a ponzi pig achieves the worst of both worlds. To build a successful crypto game, developers have to learn to own it - embrace the financialization,
build it and market it as a novel casino.
For heaven's sake, stop marketing it to core gamers who are interested in fun, not financialization. Instead, market it to gamblers and degens - there's a large untapped market, we can do a lot better than today's ponzis, memecoins etc.
Dec 11, 2022 • 5 tweets • 2 min read
As Arbitrum One makes it to #3 across a few key economic activity metrics, now behind only Ethereum and BSC, I'd caution you from using it - it's weakest link (instant upgradability by an opaque 4-of-6 multi-sig) makes it incredibly high risk
reddit.com/r/ethfinance/c…
OL has been working on decentralizing; IIUC it looks like at some point they introduced permissionless block proposals if the whitelisted validators are inactive for seven days
I think we're getting close, converting multi-sig to a transparent...
