For sure next logical step for of AI is start decentralized GPU networks. Last few years many folks tried to create. Bittensor Bittensor is the biggest by far. I am following what this guys are doing last few years and I am quite bullish on things what they are doing.

Revolut just dropped 1:1 Stablecoins by Revolut 😳 @RevolutApp is introducing 1:1 conversion between USD and stablecoins (USDT & USDC), meaning you get exactly $1.00 in stablecoins for $1.00 in cash (and vice versa), every time. Stables are here to stay.

Example of good cooperation and how mature players should act. Kudus to 0xmikko.eth and whole gearbox team.

Bitcoin’ power law at Bitcoin.Amsterdam looks cool and promising Presented by Giovanni Santostasi

We’re announcing the closure of the TAKOTAKO project All users are kindly asked to withdraw their funds and complete any remaining actions before February 18th, 12:00 UTC After this date, the website will be permanently closed, and all interactions will only be available

So true

Congrats!

x.com/i/article/2002…

2026 will be a banger year

x.com/i/article/2014…

Within just few days AI agents are able to create new protocol for secure communication. Right now also a lot of talks between OpenClaw🦞 agents on moltbook to create own language. Very curious to see what will happened by end of next week.

Best thought I’ve read about AI this year. Hard to say better.

Nice and useful skills for your LLM by Austin Griffith ethskills.com

Interesting reading. Taking into account speed of adoption openclaw I see it is happening around that time.

5 years of roadmap done in few days with AI about more then 5K worth of tokens. EF should onboard Claude

Hey Trustee Redirect what happened? support is awful and unprofessional. almost one month support can’t answer simple question what happened with refund. “Please wait, it will be good” It is also strange that CEO Vadym Grusha is not on X 🤦‍♂️

Bet continues with vitalik.eth. After 1 month, ETH2030.com's devnet is running. FOCIL inclusion lists. Frame transactions. Native AA. BAL parallel execution. All producing blocks on a 6-node network. Here's what's not working.

TBH did not expect that drop in capabilities for switching opus 4.6 to OpenAI 5.4 on openclaw. The difference is huge md, setting, scrips and code scripts are the same but result is vey different .

I want try for coding. But mythos capabilities of hacking sounds quite scary.

looks like my missing 5th element in my AI setup. lets since all other memories are not working as i expect. Stay tune.