wow
TerraformLabs' legal proceedings will be completed. Claims for lost cryptocurrency have been received and compensation will be paid. Following this, US exchanges will list $LUNC and $USTC. BinanceUS and Coinbase will be among the first. The first to list wins...
#TerraClassic #LUNC @BinanceUS @coinbase
29
amitofo retweeted
一觉睡醒,好多朋友让我看 #manus,号称一款全球真正意义上通用的AI Agent,能够实现独立思考并规划执行复杂任务,并交付完整结果。听起来非常Cool,但除了很多朋友圈焦虑的要失业的声音之外,它会给web3 DeFai场景大爆发带来什么?以下,谈谈我的思考:
1)约一个月前OpenAI 推出了同类别产品Operator,AI可以在浏览器中独立完成包括餐厅预订、购物、订票、外卖订餐等任务,用户可以可视化进行监督,并随时接管控制权。
这套Agent的出现并没有多少人讨论,原因在于它是单一模型驱动,还是工具调用的那套框架,用户一想到关键决策还需要干预进行就失去了依赖其执行任务的想法。
2)manus表面看似也差不太多,只是多了不少应用场景,包括筛选简历,研究股票、购买房产等等,但实际上是背后的框架和执行系统差异,Manus由多模态大模型驱动,并创新性的采用了多重签名系统。
简而言之,AI要模仿人执行(计划-执行-检查-行动)的PDCA循环行动,将由多个大模型共同协作完成,每个模型专注于特定环节,既能降低单个模型执行任务的决策风险,又能提高执行效率。所谓“多重签名系统”其实是多模型协作的决策验证机制,通过要求多个专业模型的共同确认来保证决策和执行的可靠性。
3)如此一比较,manus的优势显然就凸显出来了,加上视频Demo里展现的一系列操作体验,让人确实有一种非凡的体验感。但客观来说,Manus对Operator的迭代创新只是个开始,还达不到颠覆性革命意义。
关键点就在于其执行任务的复杂程度,以及非统一标准用户input Prompt进入之后大模型的容错率和交付结果成功率定义。要不然,顺着这套创新,web3的DeFai场景是不是立马就可以成熟应用了?显然,还做不到:
比如:DeFai场景下Agent要执行交易决策,需要有一个Oracle层的Agent负责链上数据收集和验证,并进行数据整合分析,还要实时监控链上价格捕捉交易机会,这个过程对实时分析有很大挑战,有可能一秒前还有用的交易机会,等Oracle大模型传输给交易执行Agent之后,交易机会就不存在了(套利窗口);
这其实暴露了这类多模态大模型做执行决策的一个最大软肋,如何联网、触链调取分析Real-Time级别的数据,并从中分析出交易机会,然后进行交易捕捉。联网环境其实还好,很多电商网站的订单价格并非实时变动,不容易给整个多模态协作造成巨大动态平衡困扰,要是在链上,这样的挑战几乎无时无刻存在。
4)所以,整体上manus的出现确实会在web2领域掀起一波朋友圈焦虑,毕竟很多重复性高的文职和信息处理工种可能会面临被AI取代的风险。但让他们焦虑他们的。
这事放在web3对DeFai应用场景的推动作用我们得客观认识:
必须得承认:意义肯定重大,毕竟它提出的LLM OS以及Less Structure more intelligence理念、尤其是多重签名系统会给web3拓展DeFi和AI的结合有很大的启迪思路。
这其实纠正了大部分DeFai项目的重大误区,不要上来就想依靠一个大模型实现AI Agent自主化思考+决策等复杂目标,这在金融场景下,根本不切合实际。
真正DeFai愿景的实现需要解决单体AI模型能力上限、多模态交互协作原子性保证、多模态系统统一资源调度和支配、系统容错和故障处理机制等等复杂问题。
比如:Oracle层 Agent,负责收集链上数据和分析,并监控价格,形成有效数据源;
决策层Agent,根据Oracle喂过来的数据进行分析和风险评估,并制定一套决策和行动方案;
执行层Agent,根据决策层给出的多种方案,并考虑实际情况进行执行,包括gas费用优化、跨链状态、交易排序冲突等等。
唯有这一系列的Agent都同步强大,并有一个庞大的系统框架落定,一个真正的DeFai革命才会掀起。
Note:以下附带 manus视频Demo,可配合以上的思考查看👀。觉得内容受用,麻烦帮忙一键三连支持下,谢谢!
amitofo retweeted
1) DeepSeek r1 is real with important nuances. Most important is the fact that r1 is so much cheaper and more efficient to inference than o1, not from the $6m training figure. r1 costs 93% less to *use* than o1 per each API, can be run locally on a high end work station and does not seem to have hit any rate limits which is wild. Simple math is that every 1b active parameters requires 1 gb of RAM in FP8, so r1 requires 37 gb of RAM. Batching massively lowers costs and more compute increases tokens/second so still advantages to inference in the cloud. Would also note that there are true geopolitical dynamics at play here and I don’t think it is a coincidence that this came out right after “Stargate.” RIP, $500 billion - we hardly even knew you.
Real: 1) It is/was the #1 download in the relevant App Store category. Obviously ahead of ChatGPT; something neither Gemini nor Claude was able to accomplish. 2) It is comparable to o1 from a quality perspective although lags o3. 3) There were real algorithmic breakthroughs that led to it being dramatically more efficient both to train and inference. Training in FP8, MLA and multi-token prediction are significant. 4) It is easy to verify that the r1 training run only cost $6m. While this is literally true, it is also *deeply* misleading. 5) Even their hardware architecture is novel and I will note that they use PCI-Express for scale up.
Nuance: 1) The $6m does not include “costs associated with prior research and ablation experiments on architectures, algorithms and data” per the technical paper. “Other than that Mrs. Lincoln, how was the play?” This means that it is possible to train an r1 quality model with a $6m run *if* a lab has already spent hundreds of millions of dollars on prior research and has access to much larger clusters. Deepseek obviously has way more than 2048 H800s; one of their earlier papers referenced a cluster of 10k A100s. An equivalently smart team can’t just spin up a 2000 GPU cluster and train r1 from scratch with $6m. Roughly 20% of Nvidia’s revenue goes through Singapore. 20% of Nvidia’s GPUs are probably not in Singapore despite their best efforts. 2) There was a lot of distillation - i.e. it is unlikely they could have trained this without unhindered access to GPT-4o and o1. As @altcap pointed out to me yesterday, kinda funny to restrict access to leading edge GPUs and not do anything about China’s ability to distill leading edge American models - obviously defeats the purpose of the export restrictions. Why buy the cow when you can get the milk for free?
amitofo retweeted
Many of you guys have been asking about optimization for ERC404 and how we plan on scaling and improving on the standard, so we thought we'd share some preliminary test results!
The following is still in development, so the numbers on release may look a little bit different - but regardless we're clearly making great strides in progress.
ERC404 V2.1
The following is data pulled from a test run of gas costs running on hardhat. We're seeing:
~28% reduction in mint gas cost
~50% reduction in transfer gas cost (by far the most important metric)
~80% increase in burn gas costs (because we're reusing token ids, this is expected)
This is a major improvement on both our current implementation as well as other attempts at optimization that we've seen from adjacent projects throughout the past week. We're extremely excited to share more details and get this iteration of ERC404 ready for production pending an up to date audit.
Gas bad 😇
amitofo retweeted
✨ We have a mint date ✨
🗓️ February 22 🗓️
💫 Where 💫
@MEonBTC Launchpad
Be ready to cast 🪄
amitofo retweeted
Amber买了几百万u @Pandora_ERC404
,他们很牛,尤其现在ftx的钱能拿回来,爽了。
Amber和 @animocabrands 是一伙,animoca是投资链游和元宇宙最大的VC,和币安关系没得说。
Animoca和9gag是天天香港约饭的,他们的meme上币安,我觉得404上币安我不会觉得奇怪。
我只是在推理,不预测价格。
Thank you Emmy for showering me with 1,125 Diamonds and getting me to shamelessly tweet this out. If you're a Solana OG head to magiceden.io/rewards to claim the Diamonds you've earned
67
amitofo retweeted
brc420 协议方,发布 bitcoin L2 解决方案 Merlin Chain 。
看了一下官宣内容,主要是在几个方面入手:
🌟保证安全和原生体验
组合了cobo和 particle 的方案,保证用户的L1资产是非托管,并且直接使用L1的钱包就可以操作;
🌟更丰富的生态dapp
带来了包括 defi 和 游戏的多个生态 dapp 应用,保证用户在L2上有更多的使用场景;
🌟抓“刚需”
可以直接在L2上铭刻各条L1链的铭文,这个是当前铭文生态最刚需的,资产创造。
🌟隐藏信息
其实在 L2 解决方案出现前,420 团队就已经铺垫许久,从围绕 bitmap 做生态建设(带来链上3万多地址社区),到发布一系列资产:蓝盒子,音乐盒,各种DLC,蓝宝石等。
这些资产的用户由于 L2 带来的“图币转换”需求,都会直接转化为 L2 的“刚需”用户,这是多达几万人的冷启动活跃用户。
为什么是“刚需”?
图币互换的新颖双向赋能和双向流动性,会给bitmap和420资产持有者,更多的赚钱效应,这就是刚需。
这也是为何我之前如此看好420资产,以及即将上线的 L2 梅林链 Merlin Chain。
这是互联网产品团队的打法,真的理解什么是“刚需”,什么是“社区”,什么是“流量”。
真的理解怎么解决最多人的“刚需”问题,让这条L2真正有用,而不是制造又一条无人使用的“vc链”。
激动的宣布我们的比特币二层方案 - Merlin Chain。
🎁转发英文原推 Bitmap * 5
Merlin Chain 将基于比特币一层的原生资产、协议和产品,继续创新,释放其潜力,Make Bitcoin Fun Again。
Merlin Chain 将桥接多个比特币一层的原生资产至二层,使用 @Cobo_Global 的 MPC 解决方案来确保所有资金非托管且安全。Merlin 的整体解决方案将经过多家公司审计并开源。
与其他比特币二层方案不同,Merlin Chain 将继续支持比特币原生钱包和地址。用户可使用 @unisat_wallet 或 @okxweb3 在 Merlin Chain 上转移、授权和接收资产,得益于我们合作伙伴 @ParticleNtwrk 的“账户抽象”功能。
Merlin Chain 上线时将有多个 Dapp 加入,包括 DEX @izumi_Finance、跨链桥 @mesonfi @PolyhedraZK,以及比特币原生资产应用 Bitmap Game 和 Bitmap War @BitmapWar,还有 BRC-420 原生 DeFi 应用 @mner_club 等。Merlin 还将支持其他链的稳定币及成功证明自己的 Dapp 团队。
最后,Merlin Chain 致力于成为所有比特币用户的“起点”。你可以轻松地在 Merlin Chain 上铭刻或铸造包括比特币在内的其他一层资产。
创新在此,革命在此。让我们一起Make Bitcoin Fun Again。
Replying to @web3cryptoMia
We have been saying that the #brc69 protocol is the most potential and indispensable recursive inscription protocol in a bull market, but many people don't actually know the technical implementation logic inside. So what is recursive inscription and what is so good about the 69 protocol?
⭕️ First, what is recursive inscription?
Simply put, in the past, when we engraved inscriptions, if we had to engrave 1000 images, we needed to upload these images one by one to the Bitcoin network and then generate them. This process was not only very labor-intensive, but also limited by the block capacity of the Bitcoin network. The size of each image had to be limited to a small range to control the cost of engraving, which would sacrifice the quality of the images.
The recursive inscription method instead extracts all the features from these images. The project only needs to upload the feature components of the NFT and cleverly combine them together. For example, a 10K collection may only require 200 components, including mouths, ears, noses, and so on. This greatly saves the cost of being on-chain.
⭕️ Secondly, what are the advantages of the brc69 protocol?
BRC69 further compresses the volume of NFT inscription (mint) based on recursive inscription, making it more user-friendly for participants.
It modularizes the implementation of recursive inscription, achieving a further reduction of 90% in the block space occupied by the entire collection.
Since the quantity of NFT inscription is the data of the entire collection, compressing the volume of this inscription can greatly reduce its size.
In summary, #brc69 serves you, and #recursive inscription serves the project. Recursive inscription reduces the cost of being on-chain, not needing to upload 10,000 images but only needing 200 components. This is not related to users but to the cost of uploading. The 69 protocol saves the cost of users' minting, which relates to your gas fees.
Of course, regardless of whom it serves, it is all for the BTC ecosystem, reducing the on-chain computational load and saving the money of miners, which ultimately saves money from our pockets.💰
From a practical perspective, when a bull market truly arrives, if the traditional way of engraving inscriptions is used, the gas fees will be very high, and many newcomers will not be able to participate due to the high gas fees. However, with the use of recursive inscription technology and brc69 technology, more than 90% of gas fees can be saved. Therefore, the popularization of this technology is inevitable. This is also why we are deploying #69frogs on the 69 protocol, to better promote brc69.
1
213
🔮 I survived The Trials of @TheWizardsOfOrd 💫
trials.wizards.art/
🧙♂️ WIZARD LEVEL 373.5 🔥
#TheWizardsOfOrd #Block78
75
LONG LIVE CURSED INSCRIPTIONS~!!@Wolfof_Bitcoin @blesstiger
Happy Jubilee! Let us play for the souls of cursed inscriptions, and rejoice in their vindication.
CURSED INSCRIPTIONS ARE DEAD. LONG LIVE CURSED INSCRIPTIONS.
1
95
amitofo retweeted
什么是Vindication? BRC20成功升级 皆大欢喜?
Vindication是昨天新发的一个PR,新增一个charm. 可以理解为把【按Jubilee前的规则会被认定为诅咒铭文, Jubilee后被Bless,直接刻为正常铭文,赋予正常编号】的铭文都打了个标签【Vindicated】。
而按照当前Domo, unisat, okx等各方达成的共识,brc20会忽略所有的Vindicated铭文,把他们视为无效,和此前的诅咒铭文一样,这个角度来说,brc20相当于没升级。
- Vindication是否识别,到底影响什么?
使用Pointer进行的Batch inscribing,此前受制于诅咒铭文的设定没有被普及开,而在Jubilee升级后本来应该大杀四方(但因为最新的vindication设定,它暂时不会被brc20接纳)。
这个批量铭刻有多种模式,最高级的模式是same-sat batch。就是在单笔交易中,刻1000个甚至更多个铭文,在同一个聪上。
docs.ordinals.com/guides/bat…
Separate-outputs可以大幅节约网络费,并且最快可在一个区块内完成所有铭文铭刻;
而Same-sat不但能节约更多的网络费,还能进一步节约预留聪的花费,减少utxo集膨胀的问题。
本身是个非常有意思的升级,但根据brc20最新的共识,Vindicated的铭文们暂时并不会被承认。
这个省钱省空间的刻铭文方式,和brc20无关。
但就像我前两天说的,其他号称要优化brc20的协议,已经提前开始布局了。CBRC20协议已经能承认和识别由i1, i2等编号结尾的非i0铭文,它们显然已经做好了拥抱Jubilee的准备,这也是CBRC20在自家协议里写到的优点之一【大幅优化网络费成本】。
-----------------
有人说这是个皆大欢喜的结局,brc20有足够的时间去准备,待稳定后迎接新版本,迎接Vindication。
实际上,Jubilee是个在10月就已经公开的事件,甚至有人在11月底找domo正面提出了这个疑问。一个多月以后的现在,Jubilee前不到一周,才来主张从长计议,通过打补丁的方式变相延后升级,着实有些欠考虑。
Just entered a giveaway with 100 airdrop spots for @BendDAO on @okxweb3, come and join!nitter.cf/BendDAO/status/1742943…
@0xfengxun
@BaleDavid4
@xin_yang666
📢 New Partnership
@okxweb3 x $BDIN 🍮
We're thrilled to collaborate with @okxweb3, a leader in the #Web3 space! Stay tuned for more details.
okx.com/web3/marketplace/lau…
81























