The ICP replica/node has been reengineered to run on AWS instances in addition to specialized ICP hardware, so node providers will be able to create nodes on AWS. AWS was incredibly helpful to DFINITY engineering teams and provided special functionality. No deal needed.

Dude, you sum up your kind. Rather than answer the points directly, you respond like the filthy money-gribber you are with ad hominem attacks on me, DFINITY and the ICP community. You even use how the ICP price, manipulated on FTX, dropped at launch in 2021, aligning yourself with, and promoting, the work of serious criminals who are in jail. Take a look in the mirror. You're sad.

I ran Bons thread through Gemini Argon on High. For those interested here is what Gemini says: Justin Bons gets the core numbers right, even if his framing is aggressive. His breakdown exposes a real gap between how DFINITY markets network activity and how the rest of the blockchain industry measures transactions. At the same time, parts of his critique overlook why the Internet Computer Protocol (ICP) records internal activity differently than monolithic chains like Ethereum or Solana. 1. The 6,000 TPS Headline Figure and Read Queries The Claim: DFINITY reports over 6,000 transactions per second (TPS) on its official dashboard, but roughly 4,000 of those are read only queries. The Reality: This claim is accurate. The headline counter on the ICP dashboard (which sits around 5,890 to 6,070 TX/s as of October 2026) combines state changing update calls with read only query calls. On ICP, a query call reads data from a single node without going through consensus or modifying the ledger state. Every other major blockchain handles read requests through off chain RPC nodes (such as an eth_call on Ethereum) and never counts them as transactions. Bundling read queries into a headline transaction metric inflates the number and makes direct comparisons with other networks misleading. 2. The 2,000 Update Calls vs. 0.4 User TPS The Claim: The remaining 2,000 TPS consists of internal application steps like contract to contract calls, replies, timers, and heartbeats. Once those are removed, actual user transactions sit between 0.4 and 0.7 TPS. The Reality: The math is accurate, and it highlights a fundamental architectural difference in how blockchains define a transaction. On Ethereum or Solana, a user transaction executes synchronously inside a single atomic block. If a user calls Contract A, and Contract A calls Contract B and receives a response, that entire chain of events counts as one transaction. ICP uses an asynchronous actor model where smart contracts (canisters) pass messages to each other across independent subnets. Because these steps happen asynchronously across different consensus rounds, ICP tracks each execution step as a separate update message: * Ingress messages: External transactions signed and submitted directly by users. This is the direct equivalent of a standard transaction on Ethereum or Solana. Filtering strictly for ingress messages and token ledger transfers brings ICP down to roughly 0.4 to 0.7 TPS. * Canister to canister calls: When one canister sends an instruction to another canister, the network processes it as a distinct consensus message. * Replies and callbacks: When a called canister finishes its work and sends data back to the first canister, that reply is also processed as a separate update message. That is why replies alone account for more than half of all update traffic on the network. * Timers and heartbeats: ICP canisters can execute scheduled code automatically on every block (heartbeats) or at set intervals (timers) without an external user triggering them. Bons is right that comparing 2,000 internal execution messages on ICP to 2,000 user signed transactions on Solana is an apples to oranges comparison. Calling those internal messages completely fake, however, misses a key technical detail. Every inter canister call and timer runs through subnet consensus, mutates state, and pays for compute cycles. On Ethereum, smart contracts cannot wake themselves up, so protocols rely on external keeper bots (such as Chainlink Automation) to submit transactions that poke the contract. Those bot pokes count as user TPS on Ethereum, whereas ICP handles them natively at the protocol level through timers. 3. The 2022 Developer Forum Incident The Claim: In 2022, a developer running less than one database upload per second generated 1,300 TPS on a subnet because each upload triggered 3,000 internal messages.

The cycles you are buying are created from ICP (i.e. when you pay using your cc, behind the scenes, you are buying ICP, that's being transformed to cycles). This is a convenience layer over the decentralized network that removes token processing friction for mainstream users.

BREAKING: Vitalik states that @Pumpfun has killed memecoins by turning new retail investors entering the crypto space into degenerate gamblers and with no real-world use, the industry will die fast. He also stated that @pumpfun has been a net negative to crypto since it was released in 2024 by draining the $crypto ecosystem and turning retail into gamblers not investors. He also said the platform is setup like a slot machine you keep clicking until your fully drained - pump ultimately wins but retail loses.

Over 20 Bitcoin have been supplied to Liquidium to borrow against it 👀 🔜 25 BTC!

The preview of icDex is here internet-computer:native

Yesterday. After 30 days of absence from the social media they came to do 2 things: 1. Host a space 2. Ban Octavian I guess priorities 🤔

I'm not trying to convince you of anything. I'm laying out why I hold this view, so whoever reads this thread can reach their own conclusion. You've used "permissioned" several times now without once defining it, so let's actually look at internet-computer:native: 🔸Anyone can deploy, no approval needed 🔸Anyone can stake and vote 🔸Anyone can apply to run node, accepted by NNS vote (Node providers are elected and removable by that vote, not appointed ) 🔸Anyone can write protocol code and propose it (Passed proposals execute on every node automatically) t’s funny how some people accept whatever narrative passes for “good decentralization,” as long as it comes from the right chain. ETH’s own DAO fork had ~5% turnout, with 25% of the yes votes coming from a single address. The voting window was under 24 hours and wasn’t publicized, the foundation pushed it through as the client default anyway. But since you have $ARB in your bio, let’s look at that one too: the DAO voted 76% against AIP-1, and the foundation had already sold the tokens before the vote even closed.🤣 That's the kind of thing that literally can't happen on internet-computer:native NNS DAO isn't just signaling, it's executing. You know, the "autonomous" part of DAO that other chains only use as a narrative. Proposals have to pass the vote before the code change is executed by NNS DAO, there's no foundation/core team deciding after the fact. I guess the top people in the field that @dfinity recruited weren't just cashing paychecks while reinventing the wheel, they actually thought through how a proper decentralized network should run. p.s. this is fun, but I'm still waiting for an actual argument, not just "nah, it's permissioned, it broke the ethos" while completely ignoring how current blockchains actually run. theblock.co/news/defi/2023-0…

crypto is no longer the most hated technology by average people we finally made it to second place thank you ai

ODIN.FUN REMINDER 🔒 1/ LOCKING & STAKING Locking (lock tokens, no yield) is code-complete, fully tested, currently going through a final AI-assisted security review. ETA: days Staking (locking + rewards) is more complex. Token creators will be able to reserve part of their supply as rewards and set custom conditions (e.g. stake X for 3–6 months to earn Y%). Code is written, preliminary testing done. No firm ETA, expect longer than locking. 🔄 2/ V1 → V2 MIGRATION Not automatic. Not 1:1. For each voted-in token: → ~10% of V1 supply migrates initially → Half goes straight into the V2 liquidity pool → Half is locked and unlocks over time → Odin.fun supplies the majority of BTC-side liquidity → Holders must actively commit during a commitment phase — the vote alone doesn't trigger migration The healthier the resulting V2 pool, the faster the locked half unlocks. A dedicated migration app is in final testing. ETA: low weeks. Vote for which V1 tokens migrate first 👉 legacy.odin.fun/roadmap 💰 3/ V1 COMPENSATION Policy unchanged: 100% of V2 profit goes toward closing the V1 deficit. Reality check: V2 is still operating at a loss, so nothing has been paid out yet. Fastest path: grow V2 volume + potentially raise outside investment earmarked for the V1 deficit. 🔐 4/ SECURITY — WHAT CHANGED SINCE THE INCIDENT → 2FA / passkey (Google Authenticator) to protect against phishing and unauthorized access → New decentralized, automated withdrawal verification flow (this is why withdrawals take slightly longer) → Continuous AI-driven security audits: frontier models running 24/7 trying to break Odin's own smart contracts 🗓️ 5/ WHAT'S COMING (1–3 months) Communities, already live. Owners earn a fee share from members' trading. Locking, days out. Staking, more work needed, no firm date. Migration app, low weeks out. And the one not fully announced yet 👀: Odin Connect, a new SDK letting any third-party site or app plug into Odin login and balances. → Connecting alone gives zero fund access → Deposits require a separate explicit approval per action → Comes with an open-source smart contract template for builders (staking games, DeFi, social apps) → The migration app itself is the first app built this way