Pinned Tweet
Replying to @raphaellin
We're introducing GLM-5.2, our latest flagship model for long-horizon tasks. It marks a substantial leap in long-horizon task capability over its predecessor GLM-5.1 and, for the first time, delivers that capability on a solid 1M-token context. GLM-5.2's new capabilities include:
Solid 1M Context: A solid 1M-token context that stably sustains long-horizon work
Advanced Coding with Flexible Effort: Stronger coding capabilities with multiple thinking effort levels to balance performance and latency
Improved Architecture: We propose IndexShare, which reuses the same indexer across every four sparse attention layers, reducing per-token FLOPs by 2.9× at a 1M context length. We also improve GLM-5.2’s MTP layer for speculative decoding, increasing the acceptance length by up to 20%
Pure Open: An MIT open-source license — no regional limits, technical access without borders
Supporting long-horizon tasks starts with making long context engineering-usable: the model must maintain quality across long, messy coding-agent trajectories, not just accept more tokens. A 1M context is easy to claim, but much harder to keep reliable under real engineering pressure. To this end, we substantially expanded 1M-context training for coding-agent scenarios, covering large-scale implementation, automated research, performance optimization, and complex debugging. The result is a long-context system that is not only wide in scope, but solid in execution: a practical substrate for sustained engineering work.
This capability is reflected in GLM-5.2's performance on three long-horizon coding benchmarks. FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. On this benchmark, GLM-5.2 trails Opus 4.8 by only 1%, while edging out GPT-5.5 by 1% and Opus 4.7 by 11%. On PostTrainBench, where each agent is given an H100 GPU and evaluated by how much it can improve small models through post-training, GLM-5.2 outperforms both Opus 4.7 and GPT-5.5, ranking second only to Opus 4.8. On SWE-Marathon, an ultra-long-horizon software engineering benchmark covering tasks such as building compilers, optimizing kernels, and developing production-grade services, GLM-5.2 still has room to grow, trailing Opus 4.8 by 13% while remaining second only to the Opus series. Across all three benchmarks, GLM-5.2 is the highest-ranked open-source model, showing that its 1M context has translated into practical long-horizon delivery capability.
Raphael.btc🦇🔊 retweeted
Introducing Frontrun—the ultimate Chrome extension built specifically for Solana traders.
🎯 Live Features:
1️⃣ Largest Wallet Label Database: 40k+ verified KOL wallets.
2️⃣ Advanced Alt Wallet Detection: Instantly uncover all associated wallets from a single address.
3️⃣ Smart Followers: Real-time insights directly in your timeline.
4️⃣ Universal Labels: Label once, see it everywhere—X, Solscan, Photon, Axiom, GMGN, and more.
We're currently in Beta, with next-level AI Analysis & Monitoring launching soon. Like, repost, and join our TG chat to secure your exclusive beta invite code!
Link below 👇
Raphael.btc🦇🔊 retweeted
比特币在某种程度上是memecoin,
黄金也是。
事实上能成为货币的东西,必须是一种memecoin,它需要有传说、有叙事,有长时间的社交传播,来建立纯粹的共识价值
美元当然也是一种memecoin,虽然它被中心化发行和控制。
值得一提的是,美元还是一种一直在贬值的memecoin 😂
On memecoins…
Memecoins have generated a lot of buzz recently, and I’ve gotten some questions on how I think about them. I am personally not a memecoin trader (beyond a few test trades), but they’ve become hugely popular. Arguably, they've been with us since the beginning – Dogecoin is still one of the most popular coins. Even Bitcoin is somewhat of a memecoin (one could argue so is the U.S. dollar, once it was disconnected from gold). Just like the early days of the internet with animated gifs, new technologies often look like a toy but evolve into something much more powerful over time. So we should be open minded about where memecoins are going, even if some are silly, offensive, or even fraudulent today. Memecoins are a canary in the coal mine that everything will be tokenized and brought onchain (every post, image, video, song, asset class, user identity, vote, artwork, stablecoin, contract etc).
Coinbase differentiates itself by being the most trusted and easiest to use platform in crypto, so how are we approaching memecoins through this lens? From first-principles, we believe in free market capitalism. If our customers want it, and it's legal, we aim to let them make that choice for themselves. But it's our job to provide them with the best information we can find to make an informed choice. If a token is a scam or fraudulent, we'd want to remove it. But if it's just low quality, the customer should be able to see reviews or community notes, and make their own choice. (Defining what is an outright scam vs low quality is often a blurry line btw, so you need to be careful not to accidentally create a censorship regime.) We want to provide useful information to consumers, not play judge and jury or endorse/recommend investments.
Some memecoins have clearly gone too far lately, to the extent people are insider trading. This is illegal, and people should understand that you will go to prison for this. In every crypto cycle, there is a get rich quick crowd that comes and goes, and learns this lesson the hard way. Don't break the law! And don't try to get rich quick. You should get rich by contributing real lasting value to society (or at least be trying - it's not easy!).
So where do we go from here? We should purge the bad actors and lift up the people trying to build lasting value. We want to bring the next billion users onchain, and the only way to do this is to build products and services which people want. Help them earn a living, avoid high fees when sending money to friends and family, find truth about what's happening in the world, or get a loan to grow their business. Give them access to sound money and financial infrastructure, to accelerate progress in the world. Memecoins have a role to play here, and I think will evolve to help artists get paid, track trends, or who knows what - it's too early to say, but we should keep exploring.
Keep building with this long term mindset, and make something people want.
Raphael.btc🦇🔊 retweeted
一个很遗憾的事实是,本周期发展至今,真正相信meme币的玩家只是少数。很多朋友对了解meme文化,研究meme传播嗤之以鼻,认为炒个“空气”币,赚钱就好研究那么多干嘛?
这其实取决于你怎么看待meme币这个事物,认知将最终决定你的判断。
在我看来,memecoin将是在BTC之后,加密货币中最大的改变人生的机会。它之所以巨大,是因为人们对它的认知还没有真正被固化,还有着巨大的分歧与不确定。人们甚至对meme之于世界的价值都缺乏清晰认知,他们可能认为马斯克反复强调的“谁控制了meme,谁就控制了宇宙”只是一句玩笑话。
这实际上与BTC在刚出现时的状况相似,而meme币由于缺乏技术外衣,会显得格外荒谬。人们太容易注意到表面的荒谬,而忽略了人们生活中的深层荒谬:这个世界巨大的注意力浪费以及它们被寡头开采攫取了数以万亿计的价值。
而meme币的奇妙之处在于它开始以戏谑的方式,身披庞氏外衣登上舞台,告诉人们他们实际上可以从被窃取的注意力价值中夺回一部分。
那些以营销之名被瓜分的一部分
那些以娱乐之名被瓜分的一部分
那些以品牌之名被瓜分的一部分
那些以奢华之名被瓜分的一部分
那些以炫耀之名被瓜分的一部分
那些以品味之名被瓜分的一部分
这一部分被价值化、被去中心化拥有时,将会变得有多巨大,人们终将会在meme币赛道拥有万亿美元市值时意识到。
而你如果愿意成为一个先行的信徒,你就不难明白认真研究meme的重要性:
1. meme币将成为下一个颠覆认知式的巨大机会
2. 在数以千万计的meme币中,只有少数几个能真正成长为巨无霸
3. 经典meme币将会最终胜出。找到、确认并持有它们就是我们在这个周期的主要任务
14 is a magical number. I believe some miracles will definitely happen this year.
#MyXAnniversary
Raphael.btc🦇🔊 retweeted
我心目中的meme币经典定义
- 有meme内容 :包括有趣的梗图、视频或者文字
- 有社交网络传播:社交传播越迅速,meme生命力越旺盛
- 有丰富的社区二创:再创作和传播是meme繁衍的基础
- 叙事(加分项):叙事往往会成为meme的传播催化剂
- meme的历史积累(加分项):越经典的meme越容易占领人们的心智
- 公平启动(加分项):没有绝对的公平,但相对公平的启动有助于快速凝聚社区共识
Raphael.btc🦇🔊 retweeted
大年初五迎财神,为了方便大家链上收红包链下花,Infini x 0xSun Group也来一张限量联名卡,支持一下李哥 @Christianeth @0xinfini ,也给大家抽5个联名卡邀请码,转发评论就好。
💳 联名卡特点:
1、Mastercard预付借记卡:充u即用;
2、免KYC享$500消费额度:KYC后消费无上限;
3、每日收益:web3余额宝,存u每日生息,收益由底层协议决定。
📢 如何获取:
1、0xSun Group社区内领取激活码,免费开卡,获得专属卡面。
2、注册链接:app.infini.money/signup?ref=…
3、开卡教程:nettle-fuchsia-c30.notion.si…
🆕节后新卡选择:
节后有2种卡可选,用于不同支付场景。
1、Meow Card:全球消费,支持绑定微信、支付宝、美团,Uber、Grab、Netflix、OnlyFans 等日常支付。
2、Rabbit Card (2月5号上线):用于订阅chatgpt, X premium等风控严格的场景。
Raphael.btc🦇🔊 retweeted
错过了Bome,我想可能不会再有三天几百倍的预售了。
踏空了Neiro,我觉得很难再有15M上线币安现货的代币了。
卖飞了Goat,我不知道下一次能拿到赛道龙头1%筹码的机会要等多久。
即使抓住了马斯克喊Pnut,我也在想,不会有比这更夸张的事了吧?
焦虑过,迷茫过,懊恼过,从未停止过,因为我享受挖掘项目、靠认知下注、再被市场认证的过程,焚膏油以继晷,恒兀兀以穷年。
日复一日的枯坐,像西西弗斯推动巨石那般不知疲倦,而天道终将酬勤。在底部大仓位买到特朗普亲自下场发的Meme币,是做梦都不敢想的剧情,如今却照进了现实。
容易感到疲惫的,是做社区和推特这两件事。一个人玩,只需要顾及自己的盈亏,也可以承受更大的风险,而带大家玩,需要考虑一个币适不适合集体上车,还要时刻提醒大家止盈,毕竟利润回撤对大部分人而言是很难受的。好在社区氛围一直很不错,这些事不会真正让我觉得心累,只是一定程度上会影响我自己的操作。
推特就更不必说了,我之所以一直用明牌地址,就是为了公开透明,至少你可以看到我的操作和分享是知行合一的,这也是我认为链上最大的优点,即使这样,小黑子也总能找到角度,亏钱是我菜,赚钱是我有内幕,甚至说是我开的盘子,各种造谣抹黑数不胜数。
可能很多人都不知道,“月底的0xSun”这个梗一开始也是别人用来黑我的,说我快到月初收群费的时候就营销一波,结果硬生生被我一次次在月底抓到金狗变成了吹我的梗。他们不知道我的群累积了几千份申请没有看,哪怕群友推荐的朋友也要等上几个月才有位置,从第一个月到现在三年多每个月都是爆满,我需要做什么营销呢,我的地址就是最好的营销。
正好快过年了,我想也是时候休息一阵了。当然我不会退圈,交易还是会做,有意思的项目还是会分享,只是可能适当降低一些频率,思考思考未来的路。
再一次,感谢各位一路以来的支持。最后,以查理芒格的一段话作为结尾吧:我们的经验往往会验证一个长久以来的观念:只要做好准备,在人生中抓住几个机会,迅速地采取适当的行动,去做简单而合乎逻辑的事情,这辈子的财富就会得到极大的增长。上面提到的这种机会很少,它们通常会落在不断地寻找和等待、充满求知欲望而又热衷于对各种不同的可能性作出分析的人头上。这样的机会来临之后,如果获胜的几率极高,那么动用过去的谨慎和耐心得来的资源,重重地压下赌注就可以了。























