Steven Shen

Cofounder and CTO of CuboAi
CEO of JidouAI
ex-Circler

[Steven Shen @syshen]

Claude kills tone of startups again

[Claude @claudeai]

Introducing Claude Meetings: add Claude to your team’s video call and it builds docs, whiteboards, and other artifacts with you while you’re still talking.

Starting to roll out in beta on Team and Enterprise plans.

[Lex Fridman @lexfridman]

NVIDIA letting Jensen go is a huge mistake, for America and for humanity. It's Apple firing Steve Jobs all over again, but much bigger, given the role of AI in defining the future of our world.

I've been keeping quiet on this out of respect for Jensen's family and team, who I've met and who are amazing humans, and because I personally hate infidelity. So my heart goes out to them in this difficult personal family moment.

But I can also say that Jensen is human, and makes mistakes like all of us do. Not that I matter at all, but I've known Jensen for a long time and have gotten the chance to spend a lot of time with him, behind the scenes, away from mics and cameras. I have seen him show kindness, humility, care to everyone around him, including strangers, including from the poorest most struggling parts of society, treating all the same with the same thoughtfulness and attention, because they are fellow humans and not because he has something to gain from them. That shows rare character, to maintain humility and humanity through all of it.

Another data point is what people who know him well have said privately to me about Jensen over the years. I've never had anyone tell me anything but how awesome he is or how much they love him. Probably over a thousand people at this point. This includes hundreds of engineers, CEOs, colleagues, friends, etc.

Also, since there was some spy talk. Jensen is as proud an American as it gets. I'm in shock frankly at all of this, including the mass hysteria around it. I love America, and I think NVIDIA letting Jensen go is terrible for America.

I stand with Jensen.

Love you all ❤️

[Steven Shen @syshen]

"The intelligence is there, but stitching it together and making it consistently useful is still hard. There's a lot of opportunity for startups to help enterprises like us, or other startups, figure out how to make it all click."

— @andyfang , DoorDash co-founder, YC Startup School Taipei

[Naval @naval]

You’d think we’d have learned from Steve Jobs and Travis Kalanick that one does not just simply replace the founder…

[Steven Shen @syshen]

If NVIDIA replaced Jensen Huang, its valuation would take a big hit.

AI chips are a war that's still being fought. A company at war needs Founder Mode. It needs someone who can make the call, place big bets, and has a real instinct for the product and the customer.

Cut his pay, do whatever you want with him. But NVIDIA's CEO has to be Jensen Huang.

[Francis Davidson @FDavidsonT]

https://t.co/F3Wq6IJ6rJ

[Steven Shen @syshen]

Fable 5.2 體感上比 Opus 5.5強上不少，今天拿一個積壓許久、一直沒機會的改版工程來嘗試。從一開始原本架構的理解，直到完工，人類介入次數少，而且你是可以盲目相信模型的判斷，一路預設，最後完成品也是令人滿意

[@jason @Jason]

This short report came out in May... and everyone seemed to miss it.

Everything happens on X first... it's just a matter of if anyone sees it I guess

[Tailscale @Tailscale]

@bot can now connect to private MCP servers on your tailnet without exposing them to the public internet.

Powered by Tailscale OAuth apps, now GA, so you can build secure agent integrations, too.

Read more here: https://t.co/cmn8tpM6w7

[Steven Shen @syshen]

Cannot agree more

[0xMarioNawfal @RoundtableSpace]

DeepMind CEO Demis Hassabis:

"One person who really understands AI will outperform an entire startup team

In 5 years, the people who build with AI will run whole companies with a tenth of the headcount"

[Steven Shen @syshen]

@BensonTWN 那老黃應該早就收到警告，他家人應該也都早就知道狀況，所以目前的設定都是事先排演好的?

[Steven Shen @syshen]

@OOBE 至少她會離開立法院

[Steven Shen @syshen]

一直以來都有人會談到 AI 泡沫，但我想可能多數人都沒有想到，AI 泡沫發生的原因竟然是來自於美國的一次期中選舉

[Steven Shen @syshen]

@ZondarkAI 跟 grok bot 比起來，到底孰優孰劣？

[Steven Shen @syshen]

Notion is changing its MCP strategy to become shared memory for agents. I think that's the right call.

People don't just use multiple LLMs now. They run more than one agent. But shared memory across agents has been a pain. Google Drive and Notion can work, but they never feel natural in speed or token efficiency. Neither is an AI agent native solution.

That's why I built Notespace (https://t.co/k6Wf7E9S4k) a few weeks ago, an agent native note-taking tool.

Notion is heading in the right direction. Expect more services to follow. Humans are no longer the first-tier user. Agents are.

[Notion @NotionHQ]

New: Give all your agents one shared brain in Notion.

Now, ChatGPT, Claude, Muse, and Bot can save what they learn to the same memory and use the same skills in Notion.

Learn how to Notion-pill your agents:

[Steven Shen @syshen]

@BensonTWN 所以幣安還能信嗎 XD

[Steven Shen @syshen]

乍看之下很驚人，但這個方法不是萬靈丹。

在 LLM 上，它目前做到的是：透過分析權重加上反覆測試，找出妨礙模型發揮的特定問題並移除，讓模型原本就具備的能力表現出來。

但它無法替模型加入原本沒有的能力，所以不能當成通用的「升級手段」。至於很小的網路，它倒是能直接寫出有用的權重，不過這離實際應用在大模型上還很遠。

但的確是一個很好的開始

[Rikiya Takehi @rikiyatakehi]

We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.

We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: https://t.co/bOWCcFhCVq

[Steven Shen @syshen]

@Deiouweyol1z Claude 前陣子有出一篇文章講他們怎麼做 code review。我們沒有那麼多層，但概念類似：用不同的 AI 做多層次 review，列出優先順序，最後由人類做評斷。

有些大的功能，也會在本地端作 e2e testing 。

我想在審查上，人類做的是了解審查了哪些項目，得到什麼結果，來評估上線的風險

[Steven Shen @syshen]

很好的一篇文章，我們有類似的經歷。

我們團隊規模更小，但過去半年，送出的 PR 成長了 10 倍以上。

一開始是工程師各自使用 AI 工具。後來我們自己做了一套功能很接近 Claude Tag 的工具，把 Human-in-the-loop 轉成 Human-on-the-loop。

結果很成功：效率大幅提升，也讓更多自動化得以發生。

但當開發不再是瓶頸，客戶也更快收到更新，新的問題就來了：過去的軟體開發流程還適用嗎？還需要分工這麼細嗎？

這是我們也在思考的事。

[Sashank Gondala @sgondala2]

https://t.co/ijYNzjSiMf

[Steven Shen @syshen]

當開發速度越來越快後，開發的瓶頸開始轉往更底層，導致每間ai巨頭都在打造他們自己的git backend

[OpenAI Developers @OpenAIDevs]

We built Pando, a new Git backend for our monorepo, to handle more pushes and fetches without changing how our engineers and agents use Git.

Pando cut push latency by ~80% and now serves nearly a petabyte of Git data a day. We can grow Git capacity alongside agent usage.

[Steven Shen @syshen]

@aobashizuu_ Threads app 我刪了一年了，他是唯一會影響我心情的產品，那時候深刻感受到他有毒就退了