0xRase ETHGas ⛽

and so, i forgot to tell you, the first stage and collection of information according to my necessary parameters was finished within an hour today i collected the necessary information for the city of DUBAI through google API, then a selection was made from 5k places and cut down by parameters to 100 places i need, i got the final file in a convenient format for me - tables then the file was transferred to the customer, after checking, we will set up the parameter filter as needed, according to additional criteria, since i initially wrote a broader algorithm, for more comfortable and accurate work in the future also today i'll get acquainted with @ElasticsAI this is exactly my field of interest, this is AI and prediction markets since i got into a thousand pioneers who have priority access to their platform if it's interesting, i'll tell you about it, but i'll definitely share the first experience

the @opinionlabsxyz team showed the roadmap for 2026, i read it and it inspires confidence in me and looks promising i think that if it is implemented the project will move several steps ahead compared to competitors i wish the team good luck

which companies added to s&p 500 in q1 2026? @opinionlabsxyz list of proposed companies: vertiv holdings (vrt) alnylam pharmaceuticals affirm holdings (afrm) sofi technologies (sofi) microstrategy (mstr) pure storage (pstg) ciena (cien) of course i would like it to be microstrategy hehe but most likely i'll agree that vertiv holdings, since the real ai boom is still ahead, development pace is not slowing down, but only growing, so that's just the conclusion

The fight within UFC between Islam Makhachev (@beeos_arenavs ambassador) and Michael Morales is FAKE what's in AI news? i like how they're developing === Claude learned to "remember the past" Anthropic opened the memory import function from other AI for all Claude users -you can transfer conversation history from ChatGPT or Gemini -Claude immediately understands your context, style and tasks -no more need to "educate" a new bot from scratch === Meta trains AI on recordings from Ray-Ban smart glasses a good solution by the way === ChatGPT-5.3 is already here - meet GPT-5.3 Instant, what's under the "hood"? -improved internet search - GPT now integrates online data with its own knowledge, providing accurate and structured information -context and consistency - the model remembers better what was being discussed and "hallucinates" less often -understanding subtext - GPT adapts to the meaning of the question and gives more relevant answers fewer refusals and unnecessary disclaimers -everyday use has become simpler and more convenient GPT-5.3 Instant is already available on paid plans

did you sleep through everything yesterday? and didn't follow the campaign from @XOOBNetwork successful campaign, people got easy money for a post XOOB NFT Mint Campaign XOOB launched a free mint of 1000 NFT on Base chain for Web3 influencers (7 days, 1 NFT per wallet) to participate you need an influencer score ≥100 on X and an original post with a mention easy money i'll tell you about the project soon and will continue talking about it p.s. = post written with the purpose of support, because this rarely happens, there were no payments for me, this is not paid partnership

So, friends, I spent the whole day watching movies and chatting with an amazing girl (she's in the photo) but I can't ignore the latest update on @vooi_io I spent enough money, sometimes successfully, sometimes not, on their platform The bottom line is that it works. I like their platform, but it needs some refinement I think we'll break into that 3 perp DEX

Tokenomics analysis: distribution model, incentives, token role, outlook for investors The tokenomics of @PerleLabs is directly tied to the points and rewards system on the platform. In the current version, participants earn points, which are essentially a pre token: the team confirmed that in the future these points will be convertible into on chain rewards (tokens) on Solana . During the public beta launch (autumn 2025), it was officially stated: "Each task brings points that will later be converted into tokens". So, Perle is now building token distribution through activity: early contributors accumulate points that, when the token launches, may turn into a certain amount of coins (similar to an airdrop mechanism). Distribution model: the exact token issuance parameters have not been announced yet (as of December 2025, there is no public information about the ticker and the token economics) . However, the investment amounts are known, and it is assumed that investors received or will receive a share of tokens according to their investments. In total, the project raised $17.5M in two rounds: $8.5M in October 2024 (seed/pre seed) and another $9M in August 2025 (seed round) . The rounds were led by CoinFund (2024) and Framework Ventures (2025) respectively, with participation from Protagonist, HashKey Capital, Peer VC, and other investors . These large venture funds are known for supporting Web3 projects, and their involvement indicates high potential of the Perle token for the investor community . In particular, Framework Ventures specializes in investments in DeFi, AI, and blockchain infrastructure, and led the round because it sees synergy between the Perle platform and crypto incentive structures . The fund’s co founder Vance Spencer noted: “data quality will be the engine of progress in AI more than just scaling models, and Perle’s transparent crypto incentives can start a cycle that unlocks a wave of high quality data” . For token holders, this means a bet on a key link in the AI industry, high quality datasets, with potential value growth as AI companies need more and more reliable data. It is expected that token distribution will be focused on the community and on incentivizing participation. Likely, a significant share will be allocated to participant rewards (mining via labeling), to keep motivating experts to work on the platform. Also, a part of the tokens will go to the team and early investors, reflecting the invested funds and effort (as a standard startup pattern, the investor share can be 15 to 25%, the team about 15 to 20%, and the rest to the ecosystem and development reserves). Although there are no exact numbers, there are hints: the project ran Galxe campaigns and whitelists, collecting Solana and EVM wallet addresses from users . This is usually done ahead of an airdrop or an IDO, to lock in early participants and possibly distribute tokens across several blockchains. According to airdrops , Perle Labs launched a campaign on Galxe where users complete tasks (subscriptions, test tasks) to potentially qualify for future token rewards . Officially, the team did not confirm an airdrop, but it is clear it is building a record of on chain contribution and user points to take them into account in token distribution

beeos doesn't try to be another polymarket "about everything" they hit one point - UFC @beeos_arenavs has a rare thing for crypto: a clear symbol the slogan "predict like a machine" is tied to their ambassador Meraba Dvalishvili, the machine and this is not just a poster: the project sells the idea that a ufc fan can predict not on emotions, but like a mechanism, through data and the collective market signal

-Humanoid (@XRoboHub)officially unveiled Embodied Tien Kung 3.0 This is an open humanoid of the new generation. * Overcomes obstacles up to 1 meter high, performs complex maneuvers with millimeter precision * Understands complex commands and avoids obstacles in real time * Supports cooperation between multiple robots with autonomous planning

Today I got acquainted with a new prediction platform @beeos_arenavs It sparked interest because the basis and focus of the prediction will revolve around the UFC, and not only that We all know how popular UFC fights are, but I haven't yet seen any in-depth analysis of this sport in terms of predictions I see great potential. Next, I will discuss other mechanics and ways to earn points

Robotics is truly impressive today. And yes, movements such as somersaults, running, etc. are taught to humans, and based on this, an AI agent is created, which subsequently controls the robot. Now you understand one thing why @getoro_xyz tries to train its AI specifically on real human data, without resorting to large-scale machine learning