Kitty

ok 4 days of stress testing the subnet template is enough ✅100% uptime

@BrutalBTCTruths @dominic_w Having inference directly on-chain is currently not viable in scale. The route currently is putting AI in decentralized networks without mempools and tx's and incentivizing them on-chain via consensus algorithms

@dominic_w Running frontier LLMs is impressive. But making inference “onchain” raises a harder question: is verifiability worth the trade-off in latency, cost, and scalability? Because in the end, users don’t reward what’s provable - they reward what actually works.

Interesting! DFINITY researchers have found a way to enable new Internet Computer GPU nodes to run: — frontier LLMs (not just neural networks as today) — competitive speed & cost — inference verifiable i.e. "onchain" Notes...🧵

@lukecannon727 People really get anti-something and stick to it

You can just say things

@ArtCryptoz At first I thought these memecoin>tradfi trading was a meme but im starting to think you guys are absolutely serious... holy opposite end of the same scamcurve

To no surprise the day trading space is full of larps Normies genuinely think they can out trade quant firms running billion dollar algorithms Only a matter of time until they find out about memecoins We’re so early

@y_nabih @deeberiroz At least remove the double dash

@deeberiroz Calling this 'LLM-assisted' is just a fancy way of rebranding 'Solidity 101.' If a hacker needs AI to spot a missing Nonce or Domain Separator in a 2018 contract, it's not 'brilliance'—it's that we still treat basic Signature Replay as 'sophisticated' in 2026.

A hacker (likely LLM assisted) is exploiting old contracts on Ethereum mainnet that have signature verification logic 🧵

@Ynyesto @PatrickAlphaC The RGB

@PatrickAlphaC Who's creating this new standard? 👀

All blockchain hardware wallets are worthless unless either: 1. You spend 20 minutes per transaction verifying your calldata 2. All hardware wallets adopt a transaction legibility standard "oh but my hardware wallet is EAL6+ rated with a secure element and a MCU made from minerals mined from pluto's ultra secure crust that-" - It doesn't matter. If you do not check calldata, you're essentially saying "I trust 100% that this website has not been hacked, because I trust whatever data they send to my wallet". It doesn't matter if your wallet is the most badass piece of security tech that's ever been, because you're letting hackers send whatever they want to your wallet, and you'll blindly sign! We've seen websites hacked all the time. @Compound_xyz was hacked just last week!! Reference: x.com/Compound_xyz/s… And we've seen MASSIVE losses from these hacks across both retail and enterprise use. - Bybit ($1.4B) - Radiant Capital ($50M) - WazirX ($200M) Some wallets have done a great job of getting the ball rolling on their own like @gridplus and @KeystoneWallet who both offer calldata decoding at the device level. @Ledger and @Keycard_ offer EIP-712 digests for signatures which are easier to verify than EIP-712 structs. @MetaMask snaps allow me to build my own custom AI bots and custom decoders to read calldata easier. But it's not enough. Calldata is still very annoying to read, and decoding it can be more confusing. Not enough wallets support EIP-712 digests. The good news... Is that transaction legibility is finally coming... Once we have a standard in place for human-readable transactions, it will be unacceptable to use a hardware wallet that does not have such a feature. And we can FINALLY use hardware wallets the correct way! I'M QUITE EXCITED.

@PatrickAlphaC Technically compound website wasn't hacked, it looks like they hacked the DNS or redirect of the domain

@tranmautritam Yes because this is literally just a 3d world library

Can AI make this?

@0xEnjooyer logos isnt a picture of a cute animal or cartoon??? im out

"there are no good devs" Yeah no shit. Good dev spends $5k and 300 hours building. Rugger spends $300 and 15 minutes copy pasting. Good dev's coin does a 3x. Rugger's does a 50x before the pull. Good dev gets called a scammer when it retraces. Rugger is already on coin number 4. Good devs spend months shipping content and updates, no one cares. Rugger drops the next CA, everyone is in. WE made the incentives completely backwards yet everyone wants to blame the launchpad or the bundlers. YOU have the power to vote with your wallet. Vote.

@AlgodTrading Been saying this. The most obvious narrative is cryptoAI/decentrlized AI/agents. That alone will be the majority of all of the top 100 tokens in 10 years. Half within 5

The largest companies in the world have links or are directly focusing on AI In crypto the cryptoAi category is ridiculously undervalued, top 50 has only 2 ai related projects In hindsighτ it will look so obvious

@naruto11eth @aave Exactly. There's more to the story here. There's a minute chance is was some guy simply pressing "swap" without understanding what's happening. No one with 50m doesn't know what liquidity is, that was one of the biggest concerns with crypto for the wealthy

two issues with what happened today: 1. people are dumb af to be holding @aave accountable here. everything worked on how it is supposed to work and how defi works. people shitting on aave just don't understand how defi works. it was purely a user flaw this time. 2. there is no way, absolutely no way, that a user sees the slippage warnings and press enter and sign the transaction. technically, a user holding $50usdt is not a user who got into crypto 6months ago and didn't know what slippage is. it just looks an extreme possible case of laundering.

@TFTC21 This is the dumbest post ever and people are going to actually think this is accurate

Someone set loose two AI agents with $1,000 each and 48 hours to trade on Polymarket. Claude: +1,322% to $14,216 OpenClaw: liquidated to zero in under 48h.

@hanni_abu @_julianma @cryptoeconprof @casparschwa Is swapping to aETH not a tax event? Are there different definitions for A-token style tokens?

@_julianma @cryptoeconprof @casparschwa tax efficiency, swapping to stETH is a tax event

Why do so many people choose to lend out ETH on Aave for 2% instead of stake via Lido for 3% yield? In a new paper with Joel Hasbrouck, @cryptoeconprof, and @casparschwa we build a structural econometric model and estimate it with data from Aave and Lido. The result? There is a huge market inefficiency that cannot be explained by smart contract risk, depeg risk, or any other risk.

@MTorygreen @AndrewYang That's really not the point. It's about the overall economy and everyone. We can't route this idea into "learn something new". In 20 years, there will be less jobs than if AI hadn't existed

of course, but if we thought this way then most of us would still be literally farming in the fields older professions have to be destroyed for newer ones to be created, and vice versa much more productive to ask yourself: What do I have to do and to learn to be suitable for those new jobs that AI creates?

All roads lead to Universal Basic Income.

@123skely Now you're just yapping

Aight what if I bought 1-2 of these, and ran an "Ai tornado Cash", pay with crypto only, private Ai, and maybe add in some TEE stuff? Thoughts? Genius? Retarded?

@aaalexhl What products does pancake swap have that wasn't coded by another team?

Wow no way Another Chinese product working with Binance gets caught violating licenses and stealing code How about you take their ass to court instead of this weak ass "lets be friends" PR speak

@TedPillows In all honesty, probably privacy smart contracts, ZKs, will.

In a world moving toward more surveillance, privacy blockchains like Zcash will become essential infrastructure. Everyone wants privacy. I’m positioned long on $ZEC.

@solana @RWA_xyz Again, another blockchain boasting about manipulative metrics. I'll never forget AVAX doing the same thing about how 90,000,000 smart contracts were launched in one day

BREAKING: Solana flipped Ethereum in total RWA holders for the first time, according to @RWA_xyz 🤯

@alexocheema @nateberkopec What setup is exo doing 80 tok/s with Kimi K2.5?

@nateberkopec What if you could get Kimi K2.5 at 80 tok/sec?

Knew this was gonna happen. Personally, can't quite justify $20k in capex for tokens right now when the best you're gonna get out of it is ~Kimi 2.5 at ~20 tok/sec.

@MTorygreen @AndrewYang It will create new jobs, but just not as many as it takes away

people propose ubi when they think ai is replacing everyone, i think the opposite ai is about to create trillions of agents, entire new markets, entire new jobs the folks pushing universal basic income are assuming a static economy history says the economy explodes after and during a revolution you don’t hand out checks during an industrial revolution, you build the rails

@nicrypto Lmao how dumb are these people

Bearish for ETH. A very bearish report from Culper Research just came out, and the writers are now short ETH. They’re essentially saying that the Fusaka upgrade increased blockspace and fees collapsed by about 90% The lower fees also mean a lower validator income - so weaker staking economics They go on to call out BitMine, arguing the rise in transaction activity and active addresses being cited as bullish are just driven by spam and poisoning attacks rather than real adoption. They also bring up Vitalik selling ~19k ETH as if he’s in-the-know It’s a significant amount, but at around 8% of his total holdings wouldn’t call it an exit. Also keep in mind Ethereum can always change the rules again with another coordinated fork. Not easy politically or technically, but it’s possible. I’m not taking sides. But when a firm publishes a thesis and then puts their money behind it, it’s at least worth understanding the mechanics they’re pointing to. So, are you holding or selling?