2025年9月1日 星期一

How can a middle power compete in artificial intelligence? 英國的Isambard-AI, The University of Bristol's Isambard-AI, powered by NVIDIA Grace Hopper Superchips,;富士通 超級電腦 富岳加 NVIDIA公司的GPU 成新超級電腦

 富士通 原超級電腦" 富岳"加 NVIDIA公司的 GPU 新超級電腦




WWWW


An illustration of a supercomputer with red and blue cables on the outside, two strong arms flexing their muscles on the left and holding a top hat on the right with a conical flask and graphic lines bursting from an open door in the centre.

Meet Isambard AI

How can a middle power compete in artificial intelligence?

A tour of Britain’s newest supercomputer


Isambard-AI
Isambard-AI, alongside the Dawn supercomputer at the University of Cambridge, will see the UK's compute capacity increase to 23 AI ExaFLOPs - the equivalent of everyone in the UK spending 85,000 years doing what the full AIRR will do in one second.Jul 17, 2025
Jul 17, 2025 — The Isambard-AI supercomputer is made fully operational as the government unveils fresh AI plans.
Jul 17, 2025 — The University of Bristol's Isambard-AI, powered by NVIDIA Grace Hopper Superchips, delivers 21 exaflops of AI performance, making it the fastest system in the ...

The Algorithm By James O'Donnell • 9.1.25 In the end, these models are designed for scale, not fidelity. They can flatter us, amplify us, even sell for us—but they can’t quite become us.

 演算法

作者:James O'Donnell • 2025年9月1日

歡迎回到"演算法"!


我目之所及,到處都能看到人工智慧克隆。在X和領英上,「思想領袖」和網紅們為粉絲提供向他們的數位複製品提問的機會。 OnlyFans的創作者們付費使用自己的人工智慧模式與粉絲聊天。據報道,中國的「虛擬人」銷售人員的銷量超過了真人。


數位克隆——能夠複製特定人物的人工智慧模型——融合了一些已經存在一段時間的技術:與你外觀相匹配的超現實視訊模型、僅基於幾分鐘語音錄音的逼真聲音,以及越來越能夠吸引我們注意力的對話聊天機器人。但它們也提供了世界上ChatGPT(通用人工智慧技術)無法提供的東西:一種並非一般意義上的智能,但思維方式與你相同的人工智慧。


它們的目標客戶是誰?新創公司 Delphi 最近從 Anthropic 和演員兼導演奧利維亞王爾德的風險投資公司 Proximity Ventures 等投資者那裡籌集了 1600 萬美元,該公司幫助名人創建可以與粉絲聊天和語音通話的複製品。這感覺就像 MasterClass(一個由名人主持的教學研討會平台)一躍進入了人工智慧時代。 Delphi 在其網站上寫道,現代領導者「擁有可能改變生活的知識和智慧,但他們的時間有限,獲取這些知識的管道也受到限制」。


它有一個由名人創建的官方複製體庫,你可以與這些複製體對話。例如,阿諾德·施瓦辛格告訴我:“我來這裡是為了讓你少些廢話,讓你變得更強壯、更快樂”,然後興高采烈地告訴我,我現在已經註冊接收 Arnold’s Pump Club 的新聞通訊了。即使他或其他名人的克隆人未能實現德爾菲傳播「大規模個人化智慧」的崇高願景,但他們至少似乎可以作為尋找粉絲、建立郵件清單或銷售補品的管道。

The Algorithm

By James O'Donnell • 9.1.25

Welcome back to The Algorithm!

Everywhere I look, I see AI clones. On X and LinkedIn, “thought leaders” and influencers offer their followers a chance to ask questions of their digital replicas. OnlyFans creators are having AI models of themselves chat, for a price, with followers. “Virtual human” salespeople in China are reportedly outselling real humans. 

Digital clones—AI models that replicate a specific person—package together a few technologies that have been around for a while now: hyperrealistic video models to match your appearance, lifelike voices based on just a couple of minutes of speech recordings, and conversational chatbots increasingly capable of holding our attention. But they’re also offering something the ChatGPTs of the world cannot: an AI that’s not smart in the general sense, but that thinks like you do.

Who are they for? Delphi, a startup that recently raised $16 million from funders including Anthropic and actor/director Olivia Wilde’s venture capital firm, Proximity Ventures, helps famous people create replicas that can speak with their fans in both chat and voice calls. It feels like MasterClass—the platform for instructional seminars led by celebrities—vaulted into the AI age. On its website, Delphi writes that modern leaders “possess potentially life-altering knowledge and wisdom, but their time is limited and access is constrained.”

It has a library of official clones created by famous figures that you can speak with. Arnold Schwarzenegger, for example, told me, “I’m here to cut the crap and help you get stronger and happier,” before informing me cheerily that I’ve now been signed up to receive the Arnold’s Pump Club newsletter. Even if his or other celebrities’ clones fall short of Delphi’s lofty vision of spreading “personalized wisdom at scale,” they at least seem to serve as a funnel to find fans, build mailing lists, or sell supplements.


Via a helpful chatbot interface, Tavus walked me through how to craft my clone’s personality, asking what I wanted the replica to do. It then helped me formulate instructions that became its operating manual. I uploaded three dozen of my stories that it could use to reference what I cover. It may have benefited from having more of my content—interviews, reporting notes, and the like—but I would never share that data for a host of reasons, not the least of which being that the other people who appear in it have not consented to their sides of our conversations being used to train an AI replica.

So in the realm of AI—where models learn from entire libraries of data—I didn’t give my clone all that much to learn from, but I was still hopeful it had enough to be useful.

Alas, conversationally it was a wild card. It acted overly excited about story pitches I would never pursue. It repeated itself, and it kept saying it was checking my schedule to set up a meeting with the real me, which it could not do as I never gave it access to my calendar. It spoke in loops, with no way for the person on the other end to wrap up the conversation. 

These are common early quirks, Tavus’s cofounder Quinn Favret told me. The clones typically rely on Meta’s Llama model, which “often aims to be more helpful than it truly is,” Favret says, and developers building on top of Tavus’s platform are often the ones who set instructions for how the clones finish conversations or access calendars.

For my purposes, it was a bust. To be useful to me, my AI clone would need to show at least some basic instincts for understanding what I cover, and at the very least not creep out whoever’s on the other side of the conversation. My clone fell short.

Such a clone could be helpful in other jobs, though. If you’re an influencer looking for ways to engage with more fans, or a salesperson for whom work is a numbers game and a clone could give you a leg up, it might just work. You run the risk that your replica could go off the rails or embarrass the real you, but the tradeoffs might be reasonable. 

Favret told me some of Tavus’s bigger customers are companies using clones for health-care intake and job interviews. Replicas are also being used in corporate role-play, for practicing sales pitches or having HR-related conversations with employees, for example.

But companies building clones are promising that they will be much more than cold-callers or telemarketing machines. Delphi says its clones will offer “meaningful, personal interactions at infinite scale,” and Tavus says its replicas have “a face, a brain, and memories” that enable “meaningful face-to-face conversations.” Favret also told me a growing number of Tavus’s customers are building clones for mentorship and even decision-making, like AI loan officers who use clones to qualify and filter applicants.

Which is sort of the crux of it. Teaching an AI clone discernment, critical thinking, and taste—never mind the quirks of a specific person—is still the stuff of science fiction. That’s all fine when the person chatting with a clone is in on the bit (most of us know that Schwarzenegger’s replica, for example, will not coach me to be a better athlete).

But as companies polish clones with “human” features and exaggerate their capabilities, I worry that people chasing efficiency will start using their replicas at best for roles that are cringeworthy, and at worst for making decisions they should never be entrusted with. In the end, these models are designed for scale, not fidelity. They can flatter us, amplify us, even sell for us—but they can’t quite become us.



2025年8月20日 星期三

日經新聞:人工智慧(AI)作為能自主規劃並執行任務的代理人,正迅速被納入企業經營版圖。

 Joel來談日本

人工智慧(AI)作為能自主規劃並執行任務的代理人,正迅速被納入企業經營版圖。美國半導體巨頭輝達執行長黃仁勳在2024年10月的網路節目中直言,希望未來「5萬名員工的同時,公司也能在各部門部署1億個AI」。這番談話出自全球首家市值突破4兆美元的企業領袖之口,格外引人注目。
AI已能撰寫程式、進行調查分析與製作報告,這些工作原本多由新進員工或實習生承擔,如今正逐步被取代。歐美產業界已出現縮減新人招聘、傾向優先錄用有經驗人才的趨勢,意味著傳統的「學習基礎業務」機會正在消失。這股浪潮也正吹向日本,程式設計領域的企業經營者更直言,「在多數情況下已不需要新人,因為AI能力更強」。
表面上看來,AI的進步似乎剝奪了年輕人累積經驗的空間,但實際上卻開啟了全新的學習模式。擁有超過20年程式設計經歷的吉田真吾指出,隨著AI進入開發現場,專案推進速度已大幅提升,過去新人能在資深工程師帶領下花一個月慢慢追趕,如今已不再可行。
然而,他也強調,AI也能成為最佳學習工具。透過請AI解說程式內容,新人或許能在短短三天內掌握專案全貌,並於第四天就能嘗試開發新功能並獲得資深同事的回饋。這種「高速學習」模式意味著,若能善用AI,年輕人仍可快速累積技能,甚至更快縮短與前輩之間的差距。
影像創作領域的實踐也印證了這點。創作者清水勝太認為,真正的突破在於「把自身技能與AI結合」,例如攝影經驗能幫助使用者對AI下達更精準的指令。若一味依賴AI,則容易淪為被動工具;唯有以個人專長為基礎,AI的潛能才能被充分釋放。
除了科技業,漫畫行業也率先出現基礎上色描繪工作流失的現象。漫畫家團體「うめ」成員小澤高廣指出,過去新人要出道,必須先成為資深漫畫家的助手,透過打雜學習基礎功。然而,隨著電子書與社群媒體的普及,新人能直接發表作品,不再需要走傳統的助手管道。再加上AI繪圖技術的崛起,即使繪畫技巧不足,仍能嘗試成為漫畫家。
但這並不代表新人不必學習。漫畫創作主要包含三個流程:發想劇情、繪製草圖(ネーム)以及正式作畫。小澤強調,雖然AI能協助背景繪製或成為劇情發想的對話對象,但在最核心的ネーム階段,目前尚無AI能真正取代。對年輕創作者而言,唯一的道路仍是「不斷畫ネーム」,因為這才是決定漫畫靈魂的關鍵。這也凸顯出一個共通訊息:即便AI正在重塑產業,新人若能抓住仍屬於人類的核心能力,就能找到立足點。
結論就是 AI不應被視為剝奪年輕人學習機會的敵人,而是能迫使人們把有限的精力投入真正關鍵領域的助力。過去人才養成的模式是「打好基礎、累積實務,再逐步走向高階工作」,但在AI時代,這樣的路徑已不再適用。AI讓人們能跨越部分傳統階段,直接以不同方式累積經驗,形成嶄新的成長軌跡。然而,這場變革不僅止於初階任務。
隨著OpenAI推出GPT-5等新一代模型,AI正加速滲透至專業與創意領域,連原本被視為人類專屬的創造力與共感力,也逐漸被挑戰。黃仁勳甚至描繪未來企業將同時擁有「數位與生物學基礎的員工」,AI之間能彼此對話、甚至「招募」AI。當AI被視為與人類並列的勞動力時,人類的職場定位正步入難以預測的歷史轉折期。由此可見,年輕人不僅需要適應新人時代的快速學習壓力,更需具備持續因應結構性變化的心理準備。
參考資料:日經新聞

論 AI: Eric Schmidt (2): 矽谷正逐漸與美國其他地區脫節......;Elon Musk;目前95 % AI 投資沒回收,股市跌: most generative AI initiatives implemented to drive revenue growth are falling flat

 

Eric Schmidt (2): 矽谷正逐漸與美國其他地區脫節......
苦口婆心,矽谷的AI熱遠離人民日常萬事的應用改善 (如中國農民用AI改善收成),
過熱/走火入魔/欲速不達?
加入火箭新創公司 Relativity Space 擔任C.E.O. 談 二三年後的AI之agents to action 的威力。 成立公司出書Eric Schmidt-founded Innovation Endeavors 切斷與Alphabet Inc.最後聯繫。 a new home for computer science at Princeton University將辭去Alphabet執行董事長職務 to head new Pentagon innovation board. Interview. Google: “Whack-a-mole is our life /”Google And The Search For The Future

----

Elon Musk has praised Google as having the strongest current lead in the AI race, highlighting the tech giant’s unmatched compute power and vast data resources. Speaking about the future of artificial intelligence, Musk noted that while Google is ahead today, the landscape could shift in the coming years as new innovations emerge.

Musk also used the opportunity to continue his ongoing feud with Sam Altman and OpenAI, suggesting that competition between AI companies is heating up. Despite rivalries, he emphasised that all major players, including his own xAI, are likely to thrive in the near term as the AI industry expands at an unprecedented pace.
Google’s dominance in AI is backed by decades of investment and research breakthroughs, including the revolutionary Transformer model that powers modern natural language processing. The company is now ramping up AI spending to $85 billion this year, further solidifying its position as a global AI leader. Google’s involvement with startups like Anthropic also reflects its strategy to stay at the cutting edge of artificial intelligence innovation.
As AI continues to transform industries and daily life, Musk’s comments highlight the intense competition and rapid technological advancements shaping the future. Analysts are watching closely to see whether Google can maintain its lead or if other AI innovators will rise to challenge the tech giant.

---

Media Outlet:
Fortune
Publication Date:
Description:

A report published by MIT’s NANDA initiative has found that most generative AI initiatives implemented to drive revenue growth are falling flat, reports Sheryl Estrada for Fortune. “Despite the rush to integrate powerful new models, about 5% of AI pilot programs achieve rapid revenue acceleration; the vast majority stall, delivering little to no measurable impact of P&L,” explains Estrada.

Google Now Let’s You Create Your Own Custom AI Assistants For FREE (How to build them

 Google Now Let’s You Create Your Own Custom AI Assistants For FREE (How to build them

👇)
Buried in a quiet update, Google made "Gems" - custom AI assistants - available to all free users. Not just paying customers - EVERYONE!
You can now create specialized AI bots for any job in minutes. No coding. No setup fees. No limits.
Here's what Google just unleashed:
☑ Custom AI assistants that remember your instructions and style
☑ Upload your own files as their knowledge base
☑ Works across web, mobile, Docs, Sheets, Gmail - everywhere
☑ Create unlimited specialized bots for different tasks
☑ All free. All instant. All powerful.
This isn't just another AI feature. This is democratizing AI expertise.
The game-changing part:
Instead of fighting with one generic chatbot, you can build a team of specialists:
→ A content strategist trained on your brand guidelines
→ A sales email writer that knows your pitch deck
→ A research assistant focused on your industry
→ A copywriter that writes in your exact voice
Each one laser-focused on doing one job perfectly.
Here's how simple it actually is:
2. Click "New Gem" in the sidebar
3. Write instructions like you're briefing an employee
4. Upload reference files (your docs, spreadsheets, PDFs)
5. Test it and save
That's it. You just created a custom AI employee.
The instruction formula that works:
"You are a [ROLE] focused on [DOMAIN]. Your goal is to [SPECIFIC TASK]. Use insights from [YOUR FILES]. Write in a [TONE] style, [WORD COUNT] words. Format as [OUTPUT TYPE]. If info isn't in the files, say so."
Copy, paste, customize. Done.
What makes this absolutely brilliant:
These Gems work everywhere Google does. Create one on your laptop, use it in Gmail on your phone, access it in Google Docs during meetings.
Your custom AI team follows you across every device and app.
The competitive advantage is insane:
While everyone else is prompting generic ChatGPT, you can have:
☑ A LinkedIn content creator trained on your best posts
☑ A proposal writer that knows your services inside out
☑ A research analyst focused on your industry trends ☑ A brand voice coach that keeps everything on-message
The practical impact:
Instead of spending 20 minutes explaining context to ChatGPT every time, your Gems already know:
→ Your brand voice and messaging
→ Your target audience and pain points
→ Your products, services, and positioning
→ Your industry knowledge and insights
What you should do right now:
Stop using generic AI prompts. Build your first Gem today.
Pick one repetitive task you do weekly - content creation, email responses, research, analysis - and turn it into a specialized AI assistant.
Test it for a week. Then build another one.