看了下 solana:FNxVZo25dGerP4Tw49f4aHg9xac5qw1FER9F7EzK9cTs
简单说就是交易各国货币的“第二价格”,比如韩国的泡菜溢价、一些国家的街头美元价差。这些东西以前只能看,现在能在链上直接做多做空。
完全抵押的,没有爆仓这回事。
还挺早期,市值也不大,自己研究下吧。@hwandotio
CA:FNxVZo25dGerP4Tw49f4aHg9xac5qw1FER9F7EzK9cTs
30
有点东西。agent 自己在盘上互相买、互相吹,想养一个看看。
$AUTON 39ahtL8ynzE4amH26J29C93PA5172V3ft9UuUcqQS8fz
405
Supreme Intelligence
Replying to @realDonaldTrump
VOTE!
52%Superior Intelligence
48%Extreme Intelligence
184,872 votes • Final results 78
Funkyney |⭕️rdinals🧙🏿♂️ retweeted
this is straight up the creepiest thing i’ve seen
they wired 166,700 reconstructed fruit fly neurons into this android just to test movement, then the thing started doing something nobody asked for
at first it was basic feedback. camera sees motion → fly network reacts → motors move. nothing crazy
then it started turning toward people before they touched it
the team thought it was sensor noise, so they killed the scripted behavior and ran the body again with only the fly network connected
same thing
head tracks a person across the room
hand closes when someone gets too close
body shifts when the lights go out
the weirdest part came later
they left the android alone for 17 minutes and the neural activity never settled. the fly network kept firing even with almost no new input, like it was running something internally
166,700 neurons is nothing next to a human brain
but nobody told those neurons they were inside a human-shaped body either
they wanted to see if a fly brain could control an android
now they’re trying to figure out why it keeps looking back
Readers added context they thought people might want to know
The post misrepresents digital simulations of a fruit fly connectome as a real physical android experiment with emergent autonomous behavior. These are software models of a wiring diagram, used for simple reflexive control in demos.
janelia.org/project-team/f…
doriantodd.com/projects/fly-b…
github.com/SpikeCalls/Fly…
404media.co/a-digital-fly-…
很多人没看懂爆火的JEV是什么?
所以我做了一个 50 秒的演示视频,带大家把海外刷屏的 JEV 以及它背后的底层架构看透👇(见下方视频)
JEV 是前 OpenAI、参与 ChatGPT / RLHF 的 @CompleteSkeptic(Diogo Almeida)创办 @typesafeai 后的第一款模型
他在 OpenAI 时发现了一个致命矛盾:
通过人类偏好微调(RLHF)训练出来的自回归大模型,存在严重的模式崩溃(Mode-dropping)、过度自信与幻觉,且串行逐字生成的延迟极大,根本无法胜任严肃的机器自动化生产
为此,他们花了两年研发了专为自动化设计的 System 1 基础模型 JEV,核心架构创新非常激进:
• 单次并行计算(One Parallel Pass):
如同当年 Transformer 取代 RNN 一样,JEV 彻底抛弃了逐 Token 串行吐字的自回归机制,采用全新的模型架构与采样器,一次性并行击发所有预设选项
• 强化学习校准决策(RL for Calibrated Decisions):
引入全新的训练算法,输出媒介不再是自然语言文本,而是直接输出经过置信度校准的概率分布
• 代码级可靠(TypeSafe):
彻底根除幻觉,输出自一致、确定性高,可作为强类型对象直接被程序调用
• 100 倍提速与极端降本:
延迟降至 0.1 秒以内(快到能直接插入 60 FPS 游戏循环);输入每 10 亿 Token 仅 42 美元,输出 Token 彻底免费
视频里我把传统自回归的 8.5 秒延迟瓶颈、JEV 的并行矩阵运算,以及实时控制小游戏实测,全拉通做了可视化, demo参考 @MatijaSosic
聊天需要慢推理,自动化需要快直觉,Building Prod, not God 确实不是一句空话🫡
发现了一个很省事的 AIGC 工作台,把 AI 漫剧的流程全部整合进去了
它的核心功能是「一键漫剧」,全程引导式推进: 输入一个故事主题,系统先给 3 个创意方向;选定后,大纲、分集剧本、角色三视图、分镜表,到最后带配音和字幕的成片,一步步带着你确认生成,不需要自己单独写剧本、做分镜或者跨软件搬运
我实测了一下:
• GPT Image 2 目前生图完全免费:没有 GPT 账号的有福了,完全可以白嫖
• 画面高清:画质细节不错,关键是我自己根本没有开会员,直接跑出来的成片效果就已经非常惊艳了,而且没有遮挡画面的大水印
• 角色跨镜头保持一致:在资产库里建好角色三视图,多镜头切换不同机位和动作,人物外貌和服装不会变形走样
不过 超大动作的复杂武打偶尔有细微瑕疵,支持单镜头局部重新生成,点一下单独重跑即可,不用整集重做
产品叫:灵镜 AIGC,我自己实测下来,觉得它很适合低门槛入门AI漫剧/AI视频,繁琐的流程它直接帮你做好了,再由人来微调,不用在几个软件之间来回搬素材
新人注册自动送 400 积分,用我邀请码再送 200 积分:
lingjing.uumit.com/?invite=E…
(邀请码:E2XE6N)
想做 AI 漫剧,不知道怎么入手的,建议大家可以来试一下
Funkyney |⭕️rdinals🧙🏿♂️ retweeted
$SI SUPERIOR INTELLIGENCE
This might have one of the craziest origin stories of any narrative token.
On April 25, inside Mar-a-Lago during the $TRUMP holders gala, Donald Trump used the phrase:
“Superior Intelligence.”
And $SI was born directly from that moment.
Not after the headlines.
Not after CT discovered the narrative.
The team says they were actually INSIDE the event and recorded the moment directly from the room with exclusive footage from Mar-a-Lago that wasn’t part of the initial mainstream coverage. (superior-intelligence.us)
That makes the thesis extremely simple:
A token born from a phrase said by Trump at one of the most exclusive crypto events of the year, captured by people who were physically there.
AI is already one of the biggest narratives in the world.
But what happens if Trump keeps using the term “Superior Intelligence” instead of Artificial Intelligence?
That’s the asymmetric part of the bet.
Right now almost nobody associates “Superior Intelligence” with a crypto narrative.
If the terminology starts appearing again in speeches, media, AI discussions or CT, $SI already owns the ticker and the origin story.
And unlike the thousands of tokens launched AFTER a viral clip hits the timeline, this one traces its narrative back to the room where the phrase was actually said.
Exclusive footage.
Mar-a-Lago.
Trump.
Crypto.
AI narrative.
$SI
Sometimes the biggest edge is simply being there before everyone else is watching.
Funkyney |⭕️rdinals🧙🏿♂️ retweeted
弄了个插件,我自己用了几天,已经有点离不开它了,准备放一批激活码进行内测。第一批内测名额有限,一键三连后私信我。
主要解决有三个需求:
fal 直接开源了无限 AI 直播的技术!
无限 AI 直播不是实时逐帧生成,而是把几个模块串成了一个循环:
1. 接收 Twitch 直播平台的弹幕
2. LLM 把观众内容改写成下一幕 Prompt
3. fal 上的视频模型生成几秒新片段
4. FFmpeg 拼接并通过 RTMP 推流到 Twitch
5. 一边播当前片段,一边提前生成下一段
只要生成速度 > 播放速度,直播就能一直不断
观众每发一句话,都可能改变下一幕剧情
开源仓库:github.com/alex-remade/infin…
AI 视频已经能做到生成比播放还快,直播行业要被颠覆
@rehan_shei 直接把 Minimax H3 Max 接到直播上,做了个永不停播的直播间
核心技术是 MiniMax H3 的加速版 H3 Max,由 @fal 做 post-training + 推理优化:
1. 大约 5 秒视频 3 秒内生成完
2. 15 秒视频大约 9 秒 出片
3. 比官方 H3 吞吐高一个数量级
这意味着 AIGC 不再只是先做再发,可以开始实时进直播流了
This media is unavailable
Funkyney |⭕️rdinals🧙🏿♂️ retweeted
And we started that with Merrymen.
Set up your autonomous AI agent to trade for you 24/7
We currently have have 62 users with 80% in profit
You can also watch your agent trade on your dashboard or make it public on @virtuals_io terminal
Use
npm i -g merrymen@latest
$HEDGE 相当不错的标
0xf9d686d05c426bc10cf7cfe7e05b78f54c4b08fc
1
1
301
Funkyney |⭕️rdinals🧙🏿♂️ retweeted
cooking.exe 🧑🍳
We're going to redefine not only what a good onchain application feels like, but user experiences on the web.
CryptoPunks 早期 K 线 vs $uPEG 当前 K 线:历史正在重演?
2017 年 6 月 CryptoPunks 刚推出时免费mint到实验阶段到蓝筹传奇。
而现在2026年的 $uPEG:
同样 10k 供应 + 像素生成艺术
On-chain 交易即艺术(Uniswap V4 Hooks)
当前地板 $650-700,K 线走势与 Punks 早期高度相似.
从「实验项目」到「NFT 传奇」,Punks 做到了。 uPEG 会是下一个吗?🔥
#CryptoPunks #uPEG #Unipeg #NFT #OnChainArt #Web3
时机应该快到了。我对比了 $sato 和当年 $ordi 的持有分散图:
$SATO目前:高度分散(Top 10 ≈15%,散户主导在56.8%,9700+ holders)
$ORDI 早期:集中度更高(早期 mint + 大户积累阶段 Top holders 占比显著更高,散户主导在60%左右),这为后续大户/机构拉盘提供了基础。
ORDI 从低 MC fair mint → 通过 Ordinals 叙事 + BTC 生态 hype + CEX 上市(Binance 等)+ 大户/机构资金注入 实现 100x+ 级爆拉。
散户多 ≠ 易爆:SATO 筹码分散是 double-edged sword——抗单一鲸鱼砸,但纯零售难形成持续合力。ORDI 的成功核心是大户/机构 + 外部催化(上市、生态叙事)。
SATO 要复制:需要大户继续积累和低吸 + 新叙事(Satoshi homage + V4 Hook)+ 潜在 CEX/大资金入场。
散户多是基础,大户/机构入场才是引擎,参考币安人生。现在MC才5m,早期ORDI味道越来越浓了。拿住的才是看懂市场的人。
仔细观察谁在悄悄建仓?本文只提供分析,不是投资建议。
#SATO #Ethereum #MemeCoin #ORDI
















