AI & Maths
This is massive. A story I've been following for years. He thought he had more time but didn't. Alessandro Masullo announces a separate, purely human proof of the Riemann hypothesis. math216.wordpress.com/2026/1…
1
11
Yes. So much math built on this conjecture. Crazy that it appears to finally be proven out!!
Simple explanation of the usefulness of the Riemann proof (so far) is that it removes the conditional nature of other math theorems dependent upon it.
And, of course, is yet another mind-blowing thing achieved by super intelligence …
nitter.gq/i/grok/share/bGVnYWN5_…
1
15
Congratulations
MedChemBench is a new benchmark for AI reasoning in medicinal chemistry.
It shows the progress our frontier models have made in their ability to help scientists use AI to identify promising drug candidates.
openai.com/index/introducing…
1
10
AI & Maths-Riemann Hypothesis.
The Lean FRO and Mathlib Initiative have released a joint statement on the Lean verification of the Riemann Hypothesis.
On November 20, Levent Alpöge released a proof of the Riemann Hypothesis as a Lean library of 17 million lines of code. Over the past day, the Lean and Mathlib communities worked together to verify it. The statement certifies the proof as correct and describes how it meets all necessary verification criteria.
Congratulations to Levent on this achievement, and thank you to everyone in the community who contributed to the verification.
Full statement: leanprover-community.github.…
1
83
This will be an important Gross Margin change factor after Jensen left NVIDIA.
Aswath Damodaran: Nvidia marks its chips up 800 to 900%. That margin can't last
"It designs the chips. It gets TSMC to make the chips. So, it doesn't have any of the manufacturing facilities. It marks those chips up 800 to 900%."
"This model is rife with tension at both ends, right?"
"Remember the typical chip company has operating margins of 40%. Nvidia's operating margins are 70%."
"But the 70 cannot stay 70 because you're going to get tensions on both ends. Meta and Microsoft are working on their own chips because that way you don't have to pay the premium."
1
59
Let’s revisit Google TPU opp
Google TPU 重點更新:OpenAI 與 Anthropic 新訂單推升出貨能見度至 2029 年、v10 產品線大幅調整 (改為 Quillfish)、聯發科與 Broadcom 為供應商中之領先關注焦點
我最新的產業調查與研究結論:
1. 訂單能見度與出貨規劃:
受惠於 OpenAI 與 Anthropic 強勁新訂單,Google TPU 出貨能見度已達 2029 年,優於 Nvidia GPU 與其他 AI ASIC(多數為 4Q27–2Q28)。
(1) 自 2028 年起,v9 將成 Google 自建算力的主力方案,v8 系列(8t/8i 或其設計變種)訂單可視為來自 OpenAI 與 Anthropic。
(2) 2028–2029 年合計需求數量:OpenAI 約 12M(聯發科設計的 8t),Anthropic 約 14M(Broadcom 設計的 8i),兩者需求均呈逐年成長趨勢。OpenAI 採買斷,Anthropic 則是租賃與買斷都有。
(3) 考量供應鏈限制,2027–2029 年較合理的 TPU 出貨估算分別為 8–9M、14–15M、16–18M。市場普遍預期 AI 資本支出於 2028 年見頂,但此規劃顯示 TPU 在 2029 年仍具出貨成長動能,關鍵在於 AI 服務龍頭 OpenAI 與 Anthropic 的直接訂單。
2. v10 產品線與供應鏈調整:
v10 由原先的 Icefish 與 Merope/Bladerunner 改為 Quillfish,預計 2029 年量產,設計服務訂單由聯發科取得。
(1) Icefish 改為 v11,預計 2030 年量產,設計服務訂單同由聯發科取得。Merope 能見度較低,供應商未定,也可能取消。
(2) 聯發科將是 v9(Humufish & Triggerfish)與 v10(Quillfish)的獨家供應商,2027–2029 年占整體 TPU 供應比重分別約 45–50%、50–55%、55–60%。TPU 對其營收與利潤的貢獻預計逐年增加。
(3) 自 v9 起新款 TPU 均採 Intel EMIB-T 先進封裝。考量初期量產能力,預期至 2029 年仍以 v8 系列為 TPU 出貨主力。v8 預期將是聯發科與 Broadcom 未來數年的 TPU 事業獲利成長的主要來源,因長期累積的學習曲線效益明顯,加上雙方負責關鍵零組件與材料採購。
3/ 市場關注度與供應商展望:
近期事件有助提升並維持 Google TPU 的市場關注度,關鍵設計服務商聯發科與 Broadcom 可望成為供應鏈首要焦點。
(1) Nvidia CEO 接班的不確定性,有利 Google 與客戶洽談 2028 年後的長期規劃。OpenAI 與 Anthropic 均為 TPU 關鍵客戶,其 IPO 也有助提升並維持 TPU 關注度,而聯發科與 Broadcom 是市場最容易聯想到的供應商。
(2) 聯發科無疑是未來數年 TPU 的最大贏家,身為 v9 後的獨家設計供應商,加上供應比重與 TPU 出貨成長同時逐年提升,均有利 TPU 事業營收規模擴大。高毛利率的 8t 系列出貨能見度達 2029 年,為關鍵利潤來源。
(3) 因 Broadcom 主導 8i 設計,故毛利率更高於 8t,有利該公司利潤成長。不過,需注意的是,Broadcom 未來數年的營運重點為非 TPU 的 XPU 與 AI 網路晶片訂單,預計最快 2027 年,非 TPU 營收將超越 TPU。
1
78
That will be a great move.
Replying to @zhoujinyuanfafa
目前只是华尔街分析师转述的市场猜测,并非官方确认。瑞穗证券Jordan Klein提到,有投资组合经理推测黄仁勋可能加入英特尔协助AI GPU策略,甚至接任CEO,陈立武或愿退居幕后。双方此前有50亿美元投资与合作关系,但目前尚无正式洽谈迹象。若成真,对半导体格局影响确实巨大,但现在仍属想象阶段。
1
153
Let’s reinstate Jensen.
My letter to NVIDIA’s board.
In short, it asks the board to create a special committee of three new directors, elected by shareholders at a special meeting, to decide whether and under what conditions Jensen Huang should return—and to hire no permanent CEO until that committee decides.
The current board may be right. It may be wrong. What it cannot be is impartial. All ten sitting directors are defendants in a lawsuit that faults them for, among other things, deferring too much to Jensen. In that suit, they face personal, non-indemnifiable exposure. Keeping Jensen out stands to strengthen their defenses; bringing him back may undermine those defenses. A fiduciary should decline to decide an important question on behalf of his principal when one answer stands to benefit him personally more than the other.
In other words, in deciding whether to reinstate Jensen, each current director faces two questions: (1) How will this affect NVIDIA? (2) How will it affect me personally and my own defense in the pending lawsuit? Given what is at stake, shareholders deserve a decision from directors who face only the first question.
Stepping aside is also what NVIDIA’s own Code of Conduct calls for. The Code covers potential conflicts as well as actual ones and directs that when one arises, “we immediately remove ourselves from any decision-making role in the matter.”
When the board’s chair explained Jensen’s departure, he cited “personal behavior inconsistent with the company’s code of conduct.” The board must now apply the same Code to itself.
(1/3)
Jackson Hongping Sun retweeted
President Donald J. Trump on a deal with Iran: “I’m not going to let somebody allow them to have a nuclear weapon.”
Quit agree. Not AI bubble, but valuation reset, as I have mentioned in my Mid-Term-Election result prediction, driven by:
1) Mid-Term Election Blue Wave;
2) Democrats possible data-center ban;
3) Fed aggressive 50bp hike;
4) Peak of super-high growth like that of 2nd half of FY25 & 1st half of FY26, as I mentioned on Oct 8th.
5) NVIDIA black swan event-potential closure of the central bank of the whole AI industry & Tighter AI Regulation;
I expect there will be a technical bounce next week but S&P 500 will not go back to its previous high until all the above concerns are fully cleared out.
1
31
Great founders-Congratulations!
Anthropic’s co-founders on the story so far. Told by Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Chris Olah
The making of Anthropic
Anthropic’s co-founders on why they started the company, what they’ve learned, and the work still ahead.
1
11
Seeing what everyone's building with Gemini 4 Neon has been electric.
Starting with a literal neon world: @HarshithLucky3 used Neon on high thinking in @Antigravity to build this glowing 3D Earth.
More Neon builds that caught our eye 🧵
Jackson Hongping Sun retweeted
To the people of Venezuela — our war is over.
It’s time to build peace together.
Jackson Hongping Sun retweeted
Entanglement usually gets top billing in quantum computing, but quantum contextuality, the idea that properties don't have fixed values independent of how you measure them, may be the deeper reason quantum computers can outperform classical ones.
Today on the cover of @ScienceMagazine, we put contextuality to work on our Willow quantum processor:
- won the "impossible" magic square game 98.3% of the time (classical limit: 88.9%)
- measured a 5.59 on state-independent contextuality tests (classical bound: 4, quantum max: 6)
- beat classical strategies in multiplayer games up to 71 qubits
- ran a landmark shallow-circuit algorithm at scale for the first time - up to 105 qubits
These tasks also catch subtle hardware errors random benchmarks miss, a sharper ruler for building better quantum systems.

ALT Quantum systems have “strange” features that are key to quantum algorithms. Researchers highlighted the benefits of quantum contextuality by playing a game that is impossible to always win using classical resources alone. By executing a sequence of measurements on a (superconducting) quantum processor, represented by the cubes, they achieved results that defy classical limits, illustrated by the impossible bow tie figure. Illustration: Nicolle Fuller, SayoStudio
Let’s see this technical bounce here in 7000 S&P 500 level.
BTIG Krinsky: S&P 500 is near a bottom, but breadth still shows the final low is not there yet @seekingalpha
1
11
1) 🔔 ANTHROPIC ($ANT) DAY ONE: Popped, then faded
Anthropic opened on Nasdaq today in the largest IPO in history: $81.3B raised from 980M shares at $83, beating SpaceX's $75B in June. With the $26B private placement, the total is ≈$107B.
1) 🚀 ANTHROPIC IPO: Priced at $83, ABOVE the range
Anthropic has reportedly priced its IPO at $83 a share**, above the $70–78 range. Shares start trading Thursday, Nov 19, on Nasdaq under the ticker **$ANT.
📊 What the pricing tells us:
1️⃣ Above the top of the range: $83 is about 6% above $78. Reports say demand was several times the number of shares on offer
2️⃣ About $1.5 trillion fully diluted, one of the biggest IPOs ever
3️⃣ About 41% above the last private round (≈$58.9/share at ≈$964B)
4️⃣ $26B private placement alongside the IPO, from Google, Microsoft, NVIDIA, AMD, Broadcom, Micron and TSMC
9
1
97
⚠️ Research only, not investment advice.
#Anthropic #ANT #IPO #Claude #AI #RSI #TAM #Google #SP500 #Nasdaq #LargestIPO #Valuation #IPOEXPRESSAI #Investing
— Jackson | IPOEXPRESS.AI
1
55
👀 What to watch: whether it holds $83, the first earnings call, gross margin, and the big unlock around mid-May 2027.
1
7
📌 My take: My fair value is ≈$77/share. The IPO came in at $83 and the stock closed at $97, about 26% above my number. I said in the video that the real risk was a first-day pop, and that's what happened. If the RSI story plays out, Anthropic can grow into this price. If it doesn't, ≈$901B of fixed compute commitments and a year of lock-up expiries won't wait. Above $77, you're paying for 2036, not 2027.
1
10



























