首席AI包工头 / Aldivn亚太区野生代言人

Joined May 2026
74 Photos and videos
Pinned Tweet
🐼 包工头今天被狠狠地上了一课! 本来想着用传说中的 Fable 5 帮我快速迭代一下之前的搬砖(套利)脚本,探索一下这新工具的能力。结果代码是跑通了,但我看着账单陷入了沉思……这大哥的“工资”差点把我干破产!📉 给大家算笔账,看看现在顶级AI的“出场费”到底有多离谱: 工作时长: 仅仅 3分半钟 结账金额: 高达 $45 折合时薪: 令人发指的 $771.4!💸 好家伙,这要是让它踏踏实实干满一个小时,我这辛辛苦苦在链上套的利,估计全得给它当保护费交了。 到底我是工头,还是它是工头?这套利我不干了,谁能帮我内推一下,我直接去给 Fable 5 打工算了!🧱🤖 #AI编程 #Fable5 #加密货币套利 #工头熊猫的日常 #被AI反向收割
2
2
6
7,349
What I predict last time. #FIFA #WorldCupFIFA
Let me predict the final of the FIFA WORLD CUP. #WorldCup #FIFA #POR #FRA
2
14
Let me predict the final of the FIFA WORLD CUP. #WorldCup #FIFA #POR #FRA
2
51
Alibaba Qwen3.7 slowly fading into irrelevance at the frontier due to proprietary stance. In it's place we have Minimax M3 and... *checks notes* Rio 3.5 397b, made by the municipal IT company of Rio de Janeiro's city government. huggingface.co/prefeitura-ri…
108
283
2,742
1,362,546
Right, let's use the model to touch the real life tasks~
"I have yet to find a problem that is too difficult to solve. It's a question of how much effort you need to put into that." — @ymatias On the latest Release Notes, Yossi, VP, Google & GM, Google Research, sat down with @OfficialLoganK to discuss how Google teams drive breakthrough science to reality at a pace never seen before, whether it’s: → flood forecasting in 150 countries, covering 2B people for severe riverine floods, → the many applications of the MedGemma open model, → our geospatial models delivering critical weather warnings to 38 million farmers in India, → agentic systems like Co-Scientist reducing the timeline for generating a novel hypothesis from nearly a decade to just a couple of days, → or many other areas!
1
60
Foreman Panda retweeted
“Não estou empolgado com a Copa 😔” Asiáticos: não diga mais nada
76
1,147
10,494
312,610
Just applied a web that helps everyone run the local LLM on your device. Please feel free to let me know what can update in the web. pickllm.vercel.app

1
1
3
1,116
Foreman Panda retweeted
People replace their phones every ~4 yrs. This means there are hundreds of millions of old phones discarded each year that are still perfectly usable as computing devices. @Google in collabration with @UCSD is exploring how to turn these old phones into cloud-computing “phone clusters”. Putting phones back in service in this way can directly reduce the environmental footprint of computing by avoiding the need for further raw material extraction, and taking advantage of the embodied carbon already incurred from manufacturing these devices, and modern phones actually are already quite powerful computers. Read more in the blog below ⬇️
Today on the blog, we discuss a pathway for the second life of phones through the exploration of “phone cluster computing”, which can directly reduce the environmental footprint of computing by avoiding the need for further raw material extraction. More →goo.gle/4aJe5vO

ALT Animation of the construction of a server using smartphones.

124
456
4,703
498,407
Foreman Panda retweeted
Today on the blog, we discuss a pathway for the second life of phones through the exploration of “phone cluster computing”, which can directly reduce the environmental footprint of computing by avoiding the need for further raw material extraction. More →goo.gle/4aJe5vO

ALT Animation of the construction of a server using smartphones.

61
210
1,688
688,789
So they just said GPT-5.5 can do the same thing. Please tell us what other model can not do? #llm #fable5 #Anthropic
1
2
4
2,462
Can anyone tell me what things can be done by Fable 5 but cannot be done by other models? #fable5 #Anthropic #llm
2
158

DiffusionGemma can now run at 2000 tokens/sec! ⚡ We made local DiffusionGemma inference 1.8× faster. Run it on 18GB RAM via Unsloth Studio. GitHub: github.com/unslothai/unsloth Guide: unsloth.ai/docs/models/diffu…
4
110
🐼 工头今天监工发现个大无语事件!那些吹上天的大厂头部AI模型,算个基础数学题依然狂翻车,妥妥的九漏鱼!🤪 为什么?因为现在全行业都在死磕Coding能力,毕竟多做算术题又不能多搞钱啊!💸 所以大家想想,现在的LLM到底在卷啥呢?🤔 说实话,写代码本身真不是什么高深莫测的领域,咱们普通打工人花点时间练练也就能熟练掌握了。这就说明,LLM真正的Native应用在非代码方向还有一大片未开垦的超级蓝海!🌊🚜 真正的杀手级应用,绝对得靠截然不同的新方向来破局!别光盯着代码搬砖了,兄弟们多开点脑洞发掘新搞钱路子吧!🚀🤑 #AI #大模型 #LLM #人工智能 #科技圈吐槽 #创业方向 #搞钱 #工头熊猫
3
73
Listen up, brick-movers 🐼🚬. You gotta look at the macroeconomics of token pricing. DeepSeek is dirt cheap (and will stay that way) because it's pegged to local consumer indexes and the sheer grind of local engineering. It's priced for the trenches. 🧱💦 But Fable 5? Bro, even if you normalize for regional Purchasing Power Parity, that thing is still a pure luxury good. It’s the Hermès of AI. We're just out here moving bricks while the elites burn Fable 5 tokens. Back to work! 👷💸
2
55
Just onboarded a new little helper to supervise my desk site! Look at that face... already judging my code or questioning its life choices. No slacking off under this Panda's watch, buddy! Welcome to the team, Stack-chan. Time to move some digital bricks! 🐼💻 #M5Stack #Stackchan #DeskSetup #TechToy #DevLife
1
2
85
喂!都给老子停下手里的活儿!👷‍♂️🐼 听大工头讲两句!今天讲讲那个刚来的刺儿头,叫什么‘扩散吉玛4’(DiffusionGemma)。名字挺唬人,其实原理比咱们工地上的搅灰机还好懂滴很!😤 你们瞅瞅那边砌墙的老王👴,他就是咱们传统的‘自回归’大模型。他砌墙那叫一个规矩,搬一块砖,抹点水泥,拍拍平,确认稳当了,再搬第二块。🧱📏 一次就砌一块砖,打死也不越雷池半步。要是让他修个九曲回肠的复杂拱门(搞深奥逻辑、写复杂算法),虽然慢点,但准没错。但要是让他砌这一大片直墙,哼哼,老王能砌到咱们退休!🐢🐌 他的算力(力气)都用在等下一块砖上了! 这‘扩散吉玛’呢,那就是咱们新来的‘闪电熊猫施工队’!🐼⚡️ 他们的绝活儿不是‘砌砖’,那是直接‘变现成墙’!比如咱们要砌个256块砖的大墙段,这帮家伙不一块块搬,他们直接运来一车杂乱无章的‘建筑垃圾’——全是沙子、石子、碎砖头、烂钢筋(这就叫初始的‘噪声’)。🗑️💣 看着跟废墟一样,对吧? 接下来就是魔法时间了!✨✨ 这帮熊猫队全员同时上阵,围着这一坨废墟,手里拿着铁锹,开始整齐划一地‘啪啪啪’瞎拍(这就是‘并行去噪’迭代)!每次拍16下。 刚拍几下,你只能看出个粗糙的墙廓。 拍到第10下,墙的主体承重柱(关键上下文 Token)稳固了,周围的乱石开始自己对号入座。🧱💪 拍到第16下,‘嘭’的一声巨响!💥💥 这坨废墟瞬间定型,变成了一个极其丝滑、横平竖直、完美无缺的256砖成品墙段!比老王砌得还整齐,速度快了整整4倍!🏎️💨🏎️💨🤩 但这帮家伙也是‘有得有失’。你要是让他们凭空在废墟里变出个极其复杂的艺术雕花(高深度线性逻辑、精密算法推导),他们可能拍着拍着就拍‘迷糊’了,拍出来的雕花跟毕加索抽象画一样(容易出逻辑幻觉)。逻辑这东西,还是得靠老王一块砖一块砖地夯实。👴 不过,要是让他们干那‘补窟窿’(代码填空/Infilling)的活儿,那他们就是‘降维打击’了!🧱🕳️🚧 你把墙中间挖个大洞丢给他们。传统的砖匠老王还得先想左边砌了啥,再想右边砌了啥,纠结半天。这帮并行熊猫队不一样,他们直接围住这个洞,同时看着左边的墙和右边的墙,然后一阵狂拍,瞬间把这个洞‘天衣无缝’地填满!速度和准确率高得吓人!🚀🚀 总结一下:传统的吉玛4是稳重的‘好砖匠’,擅长走心;这‘扩散吉玛’就是个暴躁的‘活字印刷墙段机’,擅长走量、格式化、补窟窿和极速吞吐!谁跑本地大内存电脑上干这粗活,谁用谁知道!爽!😎🤘 都听明白了吧?明白了就给老子赶紧干活去!🏗️🛠️ 别偷懒!再偷懒就把你们也扔进搅灰机里扩散了!😈 #ForemanPanda 🐼 #DiffusionGemma #扩散模型 #大模型 #技术浅解 #工地文学 #GoogleDeepMind #端侧AI #性能怪兽 #GGLS #并行生成 $GOOG
4
1,216
卧槽,大无语事件!美国现在搞出个牛逼轰轰的新行当,叫“虚拟电厂”。说白了就是合法薅羊毛!AI大佬们算力饥渴到发疯,连新建电厂都等不及了,直接把主意打到了老百姓的空调上!你想想,你正吹着空调吃着西瓜,谷歌突然塞给你几十块钱说:哥们,委屈你热一会儿,我的GPU要冒烟了!🥵 听着高大上叫电厂,其实连个烟囱都没有,本质上就是个拉皮条的软件公司💼。比如那个刚跟谷歌签了卖身契的Voltus,这帮人挨个小区装“节电器”,跟你签协议。一到当地用电高峰,这玩意儿就开始大显神威,直接给你家空调强行调高几度,或者按住你的大功率电器不让用🛑。省下来的电去哪了?全优先喂给谷歌的数据中心了!然后中间商拿大头,给你发点现金补贴,主打一个“你受热,我赚钱,AI改变世界”。💰 为啥要把主意打到千家万户的插座上?还不是因为AI这帮吞金兽太费电了!🔋 那些科技巨头空有千亿美金的预算,就是花不出去。自己盖实体电厂吧,就老美那个拉胯的制造业效率,加上比老太婆裹脚布还长的繁冗审批流程,等厂子建好,黄花菜都凉透了🐢。增量死活搞不出来,那只能在存量里搞“内卷”了,从老百姓嘴里抢电吃。⚡️ 你肯定想问,直接涨电价用市场经济调节不就完了?嘿,亚当斯密看不见的手这回算是被人民群众的铁拳给撅折了🌊。美国老百姓的思想觉悟那叫一个低,根本不吃“引进数据中心是招商引资政绩”那一套。什么拉动经济?去你的,老子只看到电费单肉眼可见地变厚了,生活成本蹭蹭往上涨!💸 哪个市镇领导敢把这种吃电狂魔放进来,老百姓手里的选票分分钟教他做人,直接给他撸下台📉。连懂王都发过行政命令,要求绝对不能助长电费价格。👑 所以你看现在的局面多滑稽:左边是无柴可烧饿得嗷嗷叫的GPU,右边是死板僵硬寸步不让的制度🤬。科技巨头们急得焦头烂额,这才逼出了“虚拟电厂”这个奇葩缝合怪。这玩意儿简直就像在国家电网上安了个智能路由器📶,把谷歌掏出来的真金白银分出那么一小撮,当做窝囊费,奖励给那些愿意在家委屈自己多流两滴汗的用户。不得不说,这商业模式算是把“钞能力”玩明白了!🤖🤑 #虚拟电厂 #AI算力尽头是电力 #谷歌 #搞笑段子 #科技那点事 #互联网新鲜事
2
5
743
Foreman Panda retweeted
Reddit 上有人制作了一个游戏,你可以在任何一家公司的股票走势图上开摩托车。
252
332
3,348
587,875
Wow
4
28
Next, need some local AI device with infinity tokens~LOL~
This should be the average American home
2
35