Oceans of Wisdom

they have done a lot of testing

this is basically all they have been working on

team has done a TON of work

in other words xTDH live but we play around with it for a bit before we take it as canonical in the protocol

testnet xTDH very soon

gm

LOL

at some point, we will borrow it directly or indirectly

Ethereum doing incredible work here

you might find this interesting and helpful

so good

These posts are about a **huge step forward in Ethereum’s “real‑time proving” race** and what that could mean for **Layer‑1 (L1) scaling + gas limits** over the next few years. I’ll break it down: --- ## 1. What the screenshots are saying in plain English **Justin Drake’s post** * “Two 5090s now prove every L1 EVM block.” * zkSync’s **Airbender** prover can generate validity proofs for *all* Ethereum mainnet blocks using **two Nvidia RTX 5090 GPUs**. * That’s ~1 kW of power (“basically a toaster”) in a single consumer‑grade box. * Because proving is now *that* fast and cheap, he says: **“The L1 gas limit is going higher. So much higher. Beast mode. Gigagas L1.”** **Jrag.eth’s quote tweet** * Claims Ethereum “doomsayers” (people who say ETH can’t scale) will be blindsided. * Argues that **research on real‑time proving of L1 blocks is moving extremely fast**, and that as a result **Ethereum’s gas limit will “explode”** in the next few years. So the core claim: > *ZK tech just hit a milestone where regular gaming GPUs can, in real time, prove that every Ethereum block is valid. This removes a major bottleneck and opens the door to massive L1 scaling.* --- ## 2. Key concepts you need to understand ### a) L1, gas, and gas limit * **L1 (Layer 1)** = the Ethereum main chain. * **Gas** = unit of computation; every op in the EVM costs gas. * **Block gas limit** = max gas all transactions in a block can consume. Higher gas limit → more computation per block → more throughput, but also more load on nodes. Today Ethereum runs at an average gas limit around **45M gas per block**, after a series of increases from 30M → 36M → ~45M in 2025. Fusaka (Dec 2025) sets a default target of **60M gas** in clients.([ethdaily.io][1]) ### b) Why gas limit can’t just be cranked up Without ZK proofs: * Every full node must **re‑execute all transactions** in every block. * If you push the gas limit too high, execution time, memory and disk I/O go up, and weaker validators fall behind. Vitalik has explicitly warned that bigger gas limits increase hardware requirements and potentially hurt decentralization.([Fellowship of Ethereum Magicians][2]) So historically Ethereum has **kept the gas limit conservative** to preserve “home‑stakable” decentralization. ### c) ZK proofs & zkVMs / zkEVMs * A **zkVM / zkEVM** is a virtual machine that can generate a **succinct ZK proof** that “this program (here: the EVM) was executed correctly on this input.” * For Ethereum L1: * A **prover** re‑executes the whole block off‑chain. * It outputs a small proof. * **Validators** verify that proof instead of re‑executing the block. Verification is *fast* and almost constant time. This is exactly the technique described in the Ethereum Foundation’s “Shipping an L1 zkEVM #1: Realtime Proving” post.([Ethereum Foundation Blog][3]) --- ## 3. What “real‑time proving” actually means The EF has now standardized a definition for “real‑time proving” for L1:([Ethereum Foundation Blog][3]) * **Latency:** ≤10 seconds for **99% of mainnet blocks** (Ethereum slot time is 12 s). * **On‑prem hardware budget:** ≤$100k capex. * **On‑prem power limit:** ≤10 kW (think home EV charger). * **Security:** ≥128‑bit security. * **Proof size:** ≤300 KiB, no trusted setups. * **Open‑source code.** In other words: proofs arrive **before the next block**, with **consumer‑ish hardware**, and are robust enough to secure mainnet. --- ## 4. Where we were at the start of 2025 vs now ### Early/mid 2025: big GPU clusters * **Succinct’s SP1 Hypercube** hit *real‑time proving* for >90% of Ethereum blocks using **~160 RTX 4090 GPUs**; later they showed 99.7% of blocks on **16 RTX 5090s**.([Succinct][4]) * **Brevis’s Pico Prism** reached **99.6% of 45M‑gas blocks in under 12 s**, average ~6.9 s, using a cluster of **64 RTX 5090s**, and aims at doing it with **16 GPUs** (~10 kW, EF’s target).([DL News][5]) ### Late 2025: Airbender & others close the loop * **zkSync’s Airbender** proved Ethereum blocks in ~35–50 s on a **single GPU** earlier this year, already making it one of the fastest zkVMs per GPU.([CoinDesk][6]) * Various zkVMs (Airbender, Pico, SP1, ZisK, etc.) are all converging on **full real‑time proving** with clusters in the 16–24×5090 range.([MEXC][7]) Now Justin Drake is saying: > **Two RTX 5090s** (one consumer PC) running **Airbender** can prove **every Ethereum L1 EVM block in real time.**([X (formerly Twitter)][8]) If those numbers hold up under Ethproofs auditing, that means: * **2×5090 (~1 kW)** vs earlier **16–64 GPUs (7–30 kW)**. * Same proof security target, but **order‑of‑magnitude lower cost and power**. That’s why he’s so hyped: this *beats* the EF’s own “10 kW, 16 GPU” real‑time spec by a healthy margin.([Ethereum Foundation Blog][3]) --- ## 5. How this unlocks higher L1 gas limits The core idea from EF research & Drake’s “lean Ethereum / beast mode” vision is: > **Replace “everyone re‑executes the block” with “everyone verifies a small proof.”** > Now you can raise gas limits *a lot* without killing decentralization.([Ethereum Foundation Blog][9]) Mechanically: 1. **Today** every validator: * Downloads the block data. * Re‑executes all transactions. * Checks they match the resulting state root. 2. **With L1 zkEVM clients**: * A small number of specialized provers re‑execute the block and generate ZK proofs. * Validators download the block **plus several independent proofs** (from different zkVM teams). * They verify these proofs, which is tiny, constant‑time work. From the EF zkEVM roadmap: once a supermajority of stake is happy running zk‑clients, **gas limits can be increased “to a level that would require validators to verify proofs instead of re‑executing blocks”**.([Ethereum Foundation Blog][3]) So: * **Execution bottleneck shifts off‑chain** to zk provers. * As long as *some* provers keep up, **validators can stay light** (Raspberry Pi, NUC etc.). This is the “beast mode / gigagas L1” Justin talks about: targeting around **1 gigagas/sec (10,000 TPS on L1, and ~10M TPS via L2 blobs)**.([Bankless][10]) --- ## 6. Why Drake jumps straight from “2×5090” → “gas limit will explode” Because it ticks three critical boxes **at once**: 1. **Latency**: Proofs land within the 10–12 s block window, even for worst‑case blocks. 2. **Power budget**: 1 kW << 10 kW home‑prover target. It’s literally “one toaster” level. 3. **Cost**: Two 5090s + a decent PC is well under the $100k capex limit EF used as the real‑time proving bound.([Ethereum Foundation Blog][3]) That means: * There is **no longer a need for data‑center‑scale proving** to keep up with mainnet. * You can realistically imagine **dozens or hundreds of independent home/office provers** around the world. * With multiple zkVM implementations (Airbender, Pico, SP1, ZisK, etc.), you get **client‑like diversity** and defense‑in‑depth.([Ethproofs][11]) From here, the EF plan (very roughly) is: 1. **Fusaka (Dec 2025)** * Gas limit default 60M. * Gas schedule fixes (EIP‑7823 & 7883) to remove underpriced heavy ops like MODEXP.([Alchemy][12]) 2. **Glamsterdam** * Introduce pipelining & protocol changes needed to fit ZK proof verification into consensus.([Ethereum Foundation Blog][3]) 3. **Optional zk clients** * Validators can choose zk‑clients; EF gathers real‑world experience & audits proofs. 4. **ZK clients become the norm → larger gas hikes** * Once a supermajority uses zk‑clients, raise gas aggressively while keeping validator hardware roughly constant.([Ethereum Foundation Blog][3]) This is the roadmap Jrag is alluding to when he says “L1 scaling in the next few years will blindside people.” --- ## 7. But gas limit isn’t the *only* bottleneck Even with perfect real‑time proving, you still have other constraints. ### a) Bandwidth & block propagation Bigger gas limit → **bigger blocks** in bytes (more calldata, more logs). That stresses: * P2P networking. * Block propagation time. * Uncle/orphan rates. Fusaka adds **caps on block size** (EIP‑7934) to match the gas limit increases so that block *byte size* remains manageable.([Nethermind][13]) ### b) State growth & disk I/O More gas = more writes to state = faster growth of the Ethereum state & historical data: * Even if execution is offloaded, **disk size and database complexity** still grow. * That’s why there are EIPs like 7642 (pruning history) and new client work to shrink archive nodes from >20 TB down to ~2 TB.([Longbridge SG][14]) ### c) Underpriced opcodes & DoS vectors ZK provers are *especially* sensitive to “weird” opcodes like **MODEXP** being underpriced, because they hurt prover performance way more than gas suggests. * EIP‑7823 adds upper bounds on MODEXP input sizes. * EIP‑7883 reprices MODEXP gas to match actual work.([Alchemy][12]) Those are being shipped *before* or alongside big gas hikes specifically to avoid ZK‑DoS issues. ### d) Economic & governance caution * Gas limit is an **economic slider between performance and decentralization**; pushing it too fast can squeeze out small validators.([Bitget][15]) * There’s active debate in the community about formal decentralization targets for larger gas limits.([Fellowship of Ethereum Magicians][2]) So “gas limit will explode” is **directionally plausible**, but it’ll still be rolled out gradually and with lots of testing. --- ## 8. How credible is the “Gigagas L1” story? **Pros / reasons to take it seriously** 1. **EF is officially leaning into it.** * The “lean Ethereum” blog explicitly talks about **1 gigagas/sec on L1 and 1 teragas/sec on L2s**, i.e., ~10k TPS L1 and ~10M TPS via L2s.([Ethereum Foundation Blog][9]) 2. **Real‑time proving is no longer hypothetical.** * We have multiple teams hitting or approaching the real‑time spec with clusters that fit in a garage.([Succinct][4]) 3. **Gas is already moving.** * In 2025 the gas limit has already gone from 30M → 36M → 45M and is headed toward 60M under Fusaka.([ethdaily.io][1]) 4. **Ethereum has a strong track record of shipping complex upgrades** (Merge, EIP‑1559, blobs, etc.) while preserving uptime.([Ethereum Foundation Blog][9]) **Caveats / reasons for skepticism** 1. **ZK complexity risk.** * zkVMs & proof systems are insanely complex. A16z’s own zkVM tracking piece points out that today’s “almost real‑time” proofs are still slow for sub‑second applications, depend on huge amounts of engineering, and often require dozens of GPUs; big protocol‑level decisions based on cutting‑edge proving tech are not risk‑free.([a16z crypto][16]) 2. **Prov­er market centralization.** * Even if a single home box *can* prove blocks, in practice we might see a handful of specialized prover networks dominate because of economies of scale. The EF is hoping to mitigate this by demanding multiple independent zkVMs and encouraging home proving, but it’s not guaranteed.([Ethereum Foundation Blog][3]) 3. **Gas‑limit trade‑offs don’t vanish.** * Bandwidth, state growth, and disk I/O still scale with gas, and there’s an active debate about how fast to push. Some researchers are explicitly warning about centralization pressure if gas goes up too quickly.([Fellowship of Ethereum Magicians][2]) 4. **Community politics & conservatism.** * Ethereum governance tends to be cautious; Fusaka’s 60M gas is only ~33% over 45M, not a 10x. We’re likely talking **multi‑year, staged increases**, not “flip a switch to 500x”. So: **the physics & cryptography are now saying “yes”**, but social & decentralization constraints will pace how quickly we actually get to gigagas. --- ## 9. Big‑picture implications (what this *actually* means) ### a) For Ethereum vs “doomsayers” Doomsayer narrative: > “Ethereum L1 will stay slow, everything meaningful moves to alt‑L1s that actually scale.” These developments undercut that in a few ways: * Ethereum can **increase L1 throughput significantly** while keeping home‑run nodes viable. * L2s continue to handle the bulk of volume, but L1 is no longer permanently stuck at ~15–20 TPS; the goal is **10k TPS L1 + 10M TPS L2**.([Ethereum Foundation Blog][9]) * Once L1 is fast *and* credibly neutral, a lot of value may gravitate back to “base‑secured” systems (native rollups, based rollups, etc.). So Jrag’s point is that if you’re pricing Ethereum as if L1 will always be “slow and full,” you’re ignoring a very fast‑moving research + engineering wave. ### b) For alternative high‑throughput L1s * Chains like Solana, Monad, etc. have built their edge around **massive L1 throughput, but in data centers**. * Ethereum’s plan is to **catch up in throughput without giving up home‑validator decentralization**, using ZK rather than pure engineering.([Bankless][10]) * If Ethereum succeeds, the main differentiators for other L1s become: * Different execution models (parallelism, specialized use cases), * Different trust / governance assumptions, * Ecosystem fit, not raw TPS. ### c) For users & apps Short‑to‑medium term (next 1–3 years): * Expect **gradual but real fee relief** on L1 as gas goes 45M → 60M and possibly higher. * L2s get cheaper too, because their proofs and data rollups are more efficiently handled. * Finality times can shrink as proofs become part of the consensus story. Longer term (3–10 years, if the roadmap holds): * We could see **L1 cheap enough** for more “slow DeFi” and settlement‑layer activity directly on mainnet. * Rollups become “native” or “based,” with much tighter coupling to L1 for sequencing and proofs, improving censorship resistance and MEV fairness.([Fellowship of Ethereum Magicians][17]) --- ## 10. TL;DR of the analysis * The tweets are celebrating **a concrete milestone**: zkSync’s Airbender prover allegedly proving **every Ethereum L1 block in real‑time on two RTX 5090s (~1 kW)**.([X (formerly Twitter)][8]) * This plugs directly into the EF’s **L1 zkEVM plan**, which aims to replace block re‑execution with proof verification and then **raise gas limits by 10–100x** while keeping validators cheap.([Ethereum Foundation Blog][3]) * Fusaka and Glamsterdam are the first steps: higher gas (60M+), safer gas schedules, and protocol tweaks to accommodate ZK.([Ethereum Foundation Blog][18]) * There are still **non‑trivial risks and constraints** (state growth, bandwidth, prover centralization, cryptographic complexity, and political caution), so “Gigagas L1” is a credible direction, not a guaranteed short‑term outcome. * But it *is* fair to say that **Ethereum’s L1 scaling story in 2025 looks radically stronger than most critics expected a few years ago**—which is exactly what those posts are calling out. If you want, next step I can do is walk through *investment / portfolio* implications or compare this concretely vs one specific alt‑L1 you’re watching. [1]: https://ethdaily.io/744?utm_source=chatgpt.com "Ethereum Hits 45m Block Gas Limit" [2]: https://ethereum-magicians.org/t/formalizing-decentralization-goals-in-the-context-of-larger-l1-gaslimits-and-2020s-era-tech/23942?utm_source=chatgpt.com "Formalizing decentralization goals in the context of larger ..." [3]: https://blog.ethereum.org/2025/07/10/realtime-proving "Shipping an L1 zkEVM #1: Realtime Proving | Ethereum Foundation Blog" [4]: https://blog.succinct.xyz/sp1-hypercube/ "SP1 Hypercube: Proving Ethereum in Real-Time" [5]: https://www.dlnews.com/articles/defi/ethereum-researcher-justin-drake-heralds-real-time-proving-progress/ "Ethereum researcher heralds ‘extraordinary’ progress for network’s next 100x scaling opportunity – DL News" [6]: https://www.coindesk.com/tech/2025/06/24/zksyncs-airbender-zkvm-proves-ethereum-blocks-in-35-seconds?utm_source=chatgpt.com "ZKsync's Airbender zkVM Proves Ethereum Blocks in 35 ..." [7]: https://www.mexc.com/nl-NL/news/ethereums-pico-prism-zkvm-achieves-99-9-real-time-proving-efficiency/132937?utm_source=chatgpt.com "Ethereum's Pico Prism zkVM Achieves 99.9% Real-Time Proving ..." [8]: https://x.com/drakefjustin/status/1991836045076263380?utm_source=chatgpt.com "Ethproofs Day 2025" [9]: https://blog.ethereum.org/2025/07/31/lean-ethereum "lean Ethereum | Ethereum Foundation Blog" [10]: https://www.bankless.com/podcast/ethereum-beast-mode-scaling-l1-to-10k-and-beyond "Ethereum Beast Mode - Scaling L1 to 10k and Beyond | Justin Drake" [11]: https://ethproofs.org/learn?ref=blog.icme.io "Learn | Ethproofs" [12]: https://www.alchemy.com/blog/ethereum-fusaka-upgrade-dev-guide-to-12-eips?utm_source=chatgpt.com "What Is the Ethereum Fusaka Upgrade? Dev Guide to 12 ..." [13]: https://www.nethermind.io/blog/shaping-fusaka-netherminds-6-eips-for-ethereums-next-upgrade?utm_source=chatgpt.com "Shaping Fusaka: Nethermind's 6 EIPs for Ethereum's Next ..." [14]: https://longbridge.com/en/news/258720694?utm_source=chatgpt.com "A Comprehensive Analysis of the 12 EIPs in Ethereum's ..." [15]: https://www.bitget.com/news/detail/12560604874802?utm_source=chatgpt.com "3 Major Risks If Ethereum Continues To Increase Its Gas ..." [16]: https://a16zcrypto.com/posts/article/secure-efficient-zkvms-progress/?utm_source=chatgpt.com "The path to secure and efficient zkVMs: How to track progress" [17]: https://ethereum-magicians.org/t/scaling-the-l1-to-gigagas-with-real-time-proving/24605?utm_source=chatgpt.com "Scaling the L1 to Gigagas with Real-Time Proving" [18]: https://blog.ethereum.org/2025/11/06/fusaka-mainnet-announcement?utm_source=chatgpt.com "Fusaka Mainnet Announcement - Ethereum Foundation Blog"

send bat signal!

lol

confusing!

whoa Stranger Things SZN5 drops on Thanksgiving?

read too!

and eventually over years the right answer will appear

I just let myself start thinking about them

even many years in advance

so the way I do things is when I think something will/should happen

but I think some day some version of this will happen

and it is premature now

would need a lot of study

yes, sure, I have no idea what is the right thing to do

"Free!"

Chateau 6529

I can do a little top up if needed

if each person with TDH chips in $350, we can buy the French one!

https://www.mansionglobal.com/articles/french-chateau-built-for-the-orange-liquor-making-cointreau-family-hits-the-market-a7c57feb

or do we prefer something like this

https://www.mansionglobal.com/articles/one-of-the-few-private-islands-off-cape-cod-massachusetts-hits-the-market-for-10-9-million-d08013eb

found another one

yeah. most ppl well better left unsaid

ah ha!

and now you made it in correctly and in the right way and you will feel better about it!

i did and congrats!

https://x.com/punk6529/status/1991278025996972291?s=20

yes hardware wallet is the minimum safe approach. and multi sig SAFE even better

you should feel free to bully ac and bats

how many more?

look, make fun of @maybe and I have a quarter with your name on it

you paid for it

why is it gifted

if arsonic grants xTDH to his new collection and you mint it

not sure it is gifted

stop kissing up to your boss

that is what is actually permanent

TDH Rate is permanent

i was never opposed to it. I just do not think it should be a requirement