raulk.eth
@raulvk
Followers
3K
Following
4K
Media
603
Statuses
7K
p2p networking @ethereumfndn | mostly nerdy tweets & occasional random rants
Lisbon, Portugal
Joined February 2008
13. this, and many other ideas, will be tested over time through incremental devnets. one step at a time. read @gballet's post here: https://t.co/5Gyh1xvAEF and follow the people (photo: slides by @unnawut, from the ream client)
Had a successful interop, with 3 lean clients connecting, and accepting each other's blocks. Thanks to everyone involved @ladislaus0x, @raulvk, @unnawut, @kamil_ably, @junsong_eth, and @g11tech + @GrapeBaBa from @zeamETH for having my back during the debugging sessions.
0
0
4
10. in other words, think Bitcoin, but inversed. instead of aiming for fixed throughput as resources fluctuate, can we dynamically meter resources and pace the protocol accordingly? 11. imo, this is also the framework for Ethereum's extreme resilience against doomsday scenarios,
1
0
2
8. but here's the catch: stakers run either potatoes (Rasp Pi) or laptops (as @drakefjustin affectionately calls them); no fancy GPU required. the tradeoff is tough. 9. one question is: can we design an adaptive/elastic network that leverages high-end resources when they're
1
0
2
6. networking in Lean looks very different. PQ leaf sigs are 3KiB vs. ~180 bytes today. aiming for 4s slot times, fast finality, multiple recursion/aggregation layers, heavily constrained by compute. 7. streaming and parallelization are crucial. every byte and ms of latency
1
0
3
5. staggering # of directions being explored at once. it's all a moving target; tons of open problems and unknowns across layers. we don't know how the puzzle will come together exactly, but we know it will. đ§Š
1
0
3
4. hats off to @drakefjustin for identifying the threat early, AND mobilizing a world-class community of cryptographers, mathematicians & engineers to attack the various pieces in parallel đ
1
0
2
2. this is not far out at all, and it could merely be an upper bound because after all: (a) it's a competitive race, (b) AI is drastically advancing material science research & more. 3. i worry about powerful, opaque state actors getting there sooner (think DeepSeek blindside).
1
0
2
don't be fooled, the sun came out for only 5min out of the 72h we spent in this intense PQ workshop âď¸ a small selection of things i learnt: 1. current consensus is that quantum computing will become crypto-relevant by 2032. this is in large part due to DARPA's Quantum
1
0
5
Let's bring fast finality asap to @ethereum ! We published an update on our related work and projects last Friday, and on this topic, these are the relevant points: - Shorten slot times to increase the heartbeat of the network and decrease every delay: Time to inclusion / fast
The only serious technical argument in this post is fast L1 finalityâsomething we as L2s can neither fix nor mitigate. We fully depend on the L1 progress rate here. I'd told @ethereumfndn leadership earlier in summer that's the ONLY thing we care about on the ETH roadmap.
79
114
717
My prediction. These chains will be one of two models: Private networks, Visa/Mastercard-style, but transparent and observable, on blockchain rails, with built-in programmability âorâ Open federations, with institutional permissioning, inviting companies listed in major
My take on Tempo as an VC chain / L1: - No client diversity (major risk) - Anyone can just make a token and use that to pay gas? lol - Should be an L2, not L1. Where is the justification for this? Where is the Ethereum alignment? - Smells like another consortium chain, oh well...
0
0
4
The GlĂłria funicular accident is profoundly tragic. Anyone whoâs visited Lisbon knows that hill intimately â¤ď¸ But this accident highlights a broader pattern Iâve observed in Portugal and Lisbon: systemic safety gaps that put people at risk daily. - Construction workers
Os trabalhadores da empresa municipal Carris, em Lisboa, apresentaram "queixas sucessivas" quanto à necessidade de manutenção dos elevadores, inclusive o da Glória que hoje descarrilou. Leia mais: https://t.co/Kqy5N1Z0oe
0
0
1
Personal take on this event, and the path forward: 1/ libp2p is bigger than any single team and funding source. In fact, Ethereum devs have been evolving specs, and maintaining go, rust, jvm, zig, c++ implementations for years. 2/ The harsh reality: many well-funded projects
Protocol Labs - 210m $FIL Filecoin Foundation - 100m $FIL Even at today's prices, that is over $600m. Seems like they should have enough to continue funding libp2p? Am I missing something?
2
4
30
We'd like to introduce our approach to scaling blobs today, following the introduction of EF's strategic goals to "Scale L1, Scale Blobs, and Improve UX" (link in next tweet)
25
54
267
Our paper was just published in the Journal of Psychiatry & Brain Science. 5 grams of creatine per day saturates your muscles, but is likely too low for the brain. đ§ľ1/10
186
995
7K
is anyone aware of past efforts to build a mathematical bandwidth utilization model for Eth? as we scale L1 capacity+throughput, scale blobs, restructure the slot, increase propagation efficiency, add new mechanisms (e.g. FOCIL, BAL), shard the mempool, etc. we're working against
2
1
12
agree! i learnt so much from working with colleagues, bosses, and clients in a physical workplace on a daily basis. discipline, social cues, how information spreads, pressure signals, etiquette, non-verbal language, etc. i canât imagine what it must be like for kids to jump
possible unpopular opinion: one of the worst thing you can do as a smart young man is get a remote job
1
0
3
expanded link here, because X is lame: https://t.co/4B4MQVXmYb
github.com
AeroSpace is an i3-like tiling window manager for macOS - nikitabobko/AeroSpace
1
0
0