Chenggong "Alex" Zhang

Founding Team Member/Head of Agent @Geodesiclab | Prev @UCLA

[Whisper @Whisperrrrr66]

这个skill硕博们必须收藏起来：耶鲁大学 CS PhD、多篇 Nature MI / ICML / NeurIPS 作者 Chen Liu 把他发顶会的做图方法论开源了！

57 套真实论文配图模板直接打包，覆盖柱状图、趋势线、热力图、3D 球体等。它不只是简单的 Python 脚本，而是一套完整的出版级 House Style 规范： 🎨 专业规范：严格的 Helvetica 字体、语义化配色（蓝目标、绿增量、红对比）、去除多余轴线、支持 300-600 DPI 及 SVG 矢量导出。 🤖 Skill 化配置：支持无缝接入 Claude Code、Cursor 或 Codex。装好后给它数据，AI 就能自动按这套顶会标准帮你出图，告别反复手动调参

正在写论文、急需提升图表质感的科研党千万别错过这个skill
👉 GitHub：https://t.co/4mcT2OnwZv

[Anthropic @AnthropicAI]

Memory is what allows Claude to learn from experience. In a new Engineering Blog post, we share how “dreaming” helps Claude organize what it’s learned, correct its mistakes, and find what it needs more easily the next time. https://t.co/uUEN4bNkr9

[Nikita Bier @nikitabier]

The fact he only had one extramarital affair with a state asset after leading the greatest technological revolution in history is what impresses me. For anyone else that number would be double, even triple, digits.

[Elon Musk @elonmusk]

Chinese 𝕏 is 牛逼

[Elon Musk @elonmusk]

I agree

[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

[月球大叔 @uncle_moon_x]

2027硅谷核心关键词
生物湿实验室
机器人
各种各样的数据和RL环境

[蔡子博士Chris @caiziboshi]

我不知道为什么很多人对这两个人有很大的恶意。甚至有很多人没有看她们视频，直接看面相已经下了结论。也有一些人看了视频不评价观点，只谈论她们说话的语气和表现出来的态度。

先说她们播客的名字SOWA wives

SOTA = State of the Art（最先进、顶尖水平）
在硅谷和科技圈（尤其是AI领域）里，SOTA 是非常常用的缩写，指“当前最先进的水平/技术”。这个播客刻意借用这个词，把自己定位成“SOTA Wives”，即硅谷高成就、高地位人士（科技高管、投资人等）的妻子/伴侣视角，讲她们在这个环境里的真实生活、婚姻、野心与关系。

我觉得这个名字没有问题，很有新意。

然后再说她们讲话的语气和态度，我觉得这都是非常个人化的东西，整体是比较随性，真实有自信的感觉，我也不觉得有太大问题。

再说视频的内容。

32 分半，10 月 12 日发布，目前约 93 万次观看。

1. 成功前先找伴侣：Sarah 25 岁结婚，认识丈夫时自己住在创业公寓"楼梯底下"的隔间、一事无成，丈夫后来成了被 Nvidia 收购公司的创始工程师。她们的逻辑是：成功之前结婚，对方的动机是干净的，这种信任花钱买不到；要么单身，要么在奋斗期找到那个共患难的人。

2. 婚姻与野心：两人承认婚姻改变了优先级，不再是"把一生献给工作"那类野心，更追求工作生活平衡；但反过来说，稳定的伴侣反而让人更敢冒险，因为有兜底感，不用再靠睡气垫床这类极端方式证明自己。

3. 被网暴:Vivian 因加州 AI 安全法案 SB 1047 的言论被公开抨击；Sarah 的 25 岁生日总结帖被人转发称为"心理暴力"。她们的态度是：被别人的精彩人生刺痛的人应该反省自己的人生，在网上撕别人比去看心理医生容易；争议两周就被人忘了，继续发就行。她们还对比了 TikTok 匿名喷子和 YouTube 实名评论的氛围差异。

4. YC 怎么看创始人:Vivian 说哈佛/MIT/斯坦福毕业生常被训练成"收集徽章"的人，反而不一定适合创业；她更看好"不听话"的人。还提到 YC 内部对没戏的团队有个说法叫 "palliative care"（姑息治疗），就是顺其自然让团队自己醒悟。

5. 和前男友创业:Sarah 在 MIT 疫情期间和当时男友做了 Thrift House，拿到投资和加速器录取（条件是退学拿约 62.5 万美元），两人最后选择不退学、回校读书，公司就此结束。

我感觉没啥有争议的观点,反而很多真实的分享。唯一有争议的或许：对 30 多岁单身朋友的评价：说这代价不值得、现在压力大得多，容易被读成居高临下。还有把批评者说成"该去看心理医生"。

期待未来更多视频！

[Xuheng Li @xuhengli_]

Muon now trains frontier LLMs. But how does its spectral orthogonalization shape feature learning?

New paper: on factual recall, Muon learns facts in a more balanced way and is far less slowed by softmax saturation.
🔗https://t.co/hPm5VI0SG2
[1/8]

[Geodesic Intelligence @Geodesiclab]

Excited to partner with @GenScript to launch the Beyond Binding Challenge! 🧬

AI-designed antibodies need to go beyond binding affinity to become real therapeutics.

Let's push the boundaries of AI-driven antibody design and turn computational breakthroughs into real-world impact.

Up to $2.5M in awards. Stay tuned!

[OpenAI @OpenAI]

We’ve acquired Orchestra Bio, a platform that helps biopharma teams coordinate the complex process of drug development.

Together, we’ll build on Rosalind, our AI models and tools for life sciences research, to streamline those operations and help new treatments reach patients sooner.

Welcome @ashokaraj_ and team!

https://t.co/vi51OhznQn

[Chenggong "Alex" Zhang @ChenggongAlex]

At @Geodesiclab, we are also building our NovaLab as well as Wet-Lab.
https://t.co/WjwHWdAYz9

[Google Research @GoogleResearch]

We are bringing agents to the wet-lab bench at Google Research to navigate complex biological landscapes. Learn how we are grounding agents in physical lab robotics to run fast closed-loop molecular design cycles, ultimately discovering single-digit nanomolar protein binders for cancer targets →https://t.co/iPesR5MXCd

[Chenggong "Alex" Zhang @ChenggongAlex]

Check out my old colleague’s amazing new work about understanding On-Policy Distillation from theorist perspective.

[QIWEI DI @di_qiwei]

OPD forgets less than SFT, yet breaks under teacher mismatch and on long rollouts. Why? Our answer: they optimize different targets. With multiple teachers, SFT (forward KL) learns their arithmetic mixture; OPD (reverse KL) learns their normalized geometric mean. 🧵

[Sitan Chen @sitanch]

🚨 Thrilled to be co-organizing a Simons Institute semester-long program on diffusions and flows with @joanbruna, Sinho Chewi, @KonstDaskalakis, @ArnaudDoucet1 in Fall '27!

The call for Simons postdoctoral research fellowships for the program is out now, more info below ⬇️ 1/3

[月球大叔 @uncle_moon_x]

遇到真诚的人，自动切换到真诚模式，努力沟通，好好做事
遇到不真诚的人，自然就要切换到打哈哈状态，避免被坑

结果就是自然而然，物以类聚人以群分

甚至不同人对一个人的评价可以天差之别

[lily zhang @lily_gpupoor]

launching 𝐒𝐎𝐓𝐀 𝐓𝐕: 𝐚 𝐧𝐞𝐰 𝐩𝐨𝐝𝐜𝐚𝐬𝐭 𝐰𝐢𝐭𝐡 𝐥𝐞𝐚𝐝𝐢𝐧𝐠 𝐦𝐢𝐧𝐝𝐬. we'll kick off a series of conversations at COLM this week.

for episode 1, @Madisonkanna and I sat down with @gneubig, CMU professor, OpenHands co-founder.

we talked about his new agents course at CMU, which AI agents will still be around in five years, why he is an agent-forward professor, and more.

[Chenggong "Alex" Zhang @ChenggongAlex]

Check out this most efficiency way with watermarking for AI Safety.

[Gang Wen @GangWenYale]

Thrilled to see textGrain launch! So many ideas and iterations led us to its sampling-diversity perspective, and a huge team effort brought it to life. Proud of what we’ve built together. Our paper will come out about a week. Happy to chat, stay tuned!

[Freda Duan @FredaDuan]

AI Drug Discovery Is Becoming a Bottleneck Trade

I get excited when an industry starts going through a real regime change. AI drug discovery (“AIDD”) increasingly looks like one of those moments.

Two things are happening:
1/ Upstream: the frontier AI labs are piling in.
2/ Downstream: The supply chain is clearly moving.

Supply-chain checks suggest the upstream picks-and-shovels of discovery and early preclinical R&D are starting to feel the increase in experimental volume.

DNA → protein → assays → sequencing → automation → preclinical testing

Names across that stack include $TWST , @GenScript , $ILMN , $TXG , lab-automation vendors and CROs.

$TWST expects triple-digit percentage growth in AI-enabled drug-discovery orders in FY26, and another year of triple-digit order growth in FY27.

@GenScript's AIDD business doubled YoY in 1H26. Its current platform advertises industrial-scale validation of 4,000+ designs/day, with integrated sequence-to-data workflows. Our channel checks suggest the ramp is moving even faster: roughly 8,000 designs/day currently, with a path toward ~16,000/day by YE26.

---
Why does AI drive more wet-lab demand?

1/ AI makes hypothesis generation almost free → way more shots on goal.
The bottleneck is moving from expert-driven design to biological validation.

2/ AI models need continuous experimental feedback — and both good and bad data are useful.
Traditionally, only the highest-conviction A+ candidates might get pushed into expensive validation. With AI, even the B/C candidates can be valuable because failed experiments generate training data.

@GenScript has said its sequence-to-binding workflow can return data in 4–7 days, and that faster cycle times matter because AI models depend on continuous experimental feedback.

@Anthropic is a clean example. @claudeai designed 1,320 protein binders. @adaptyvbio converted those digital sequences into DNA, expressed the proteins and tested binding. Only 354 actually bound. And the 966 failures are not wasted. They are useful negative labels: what does not express, what does not bind, what has poor affinity. Those results help train the next model iteration.

3/ Wet labs are no longer just making drugs. They are making training data.
$TWST / @GenScript increasingly look like biological data foundries.

$TWST explicitly talks about generating model-ready data from AI-designed sequences. In some workflows, the customer may care less about receiving the physical protein than about getting structured experimental results back into the model.

Traditional drug discovery asks: “Does candidate X work?”

AI drug discovery also asks: “What can this experiment teach the model?”

---
TAM of AIDD

If AI is simply a better R&D tool, the relevant spending pool is the $300–400B of annual global pharma R&D. If AI meaningfully increases the number of viable drug programs, the opportunity is larger because it expands downstream demand for DNA synthesis, protein production, assays, and preclinical work.

Near term, we can also size demand from AI-company spending. If Anthropic reaches $80B of ARR in 2026 and spends just 1% on AIDD, that alone would imply ~$800M of annual investment.

---
Trade setup

This is a trade that could have long legs. It’s hard to really stop working until PhaseI/II results (2028+)

It smells a lot like the bottleneck trade we just saw in semis: GPUs → HBM → networking → power/cooling. In biology, It basically follows the drug discovery process downstream: AI models → designs → DNA/protein → assays → preclinical capacity.

After the upstream picks-and-shovels, animal testing could become the next bottleneck. Monkey prices are already near prior highs and CRO capacity is tight. AIDD pushing more candidates into preclinical development would only add demand.

?? But clinical trials are still the bottleneck?
This is the biggest pushback I keep coming back to. No matter how fast discovery becomes, drugs still need to go through preclinical → Phase I → Phase II → Phase III → approval. You still need patients, time and capital.

But that doesn’t mean the bottleneck trade won’t work. More viable candidates — especially with higher success rates — still means more demand throughout the development process.

And who knows: clinical trials themselves may eventually be optimized by AI.

?? What breaks the trade?
Near term, the picks-and-shovels trade breaks if experimental budgets stop growing, AI-generated designs don’t translate into useful wet-lab hits, or capacity catches up too quickly.

Longer term, the thesis breaks if AI drugs look great in discovery / Phase I but fail at normal rates in Phase II/III. That is why Phase II matters so much.

?? Milestones
Late 2026–2027: first Isomorphic-designed drugs enter human trials; more AI-native programs move into IND-enabling work / tox.

2028–2030: clinical trial results start telling us whether AI-designed drugs actually perform better than conventional drugs.

Calling all the "bottleneck bros". :) @jukan05 @zephyr_z9 @aleabitoreddit @ParadisLabs

+++
More comprehensive analysis: https://t.co/vaTuiyyYB6

[Geodesic Intelligence @Geodesiclab]

What does an AI-native workflow for protein binder design look like in practice?

In our latest post, we walk through a pilot design run using NovaDDE:

Given a target, the AI agent analyzes the design problem, helps select an epitope and design approach, submits the job, and guides the computational evaluation of the resulting candidates.

We also share some of the directions we’re exploring as we continue to develop the workflow.

A closer look at NovaDDE and where we’re taking it next:

https://t.co/I9eOefsOP6

[Chenggong "Alex" Zhang @ChenggongAlex]

5/5

📐 The proof uses a localized Eluder dimension https://t.co/Gxrhz3szUx for the excess log-loss class, separating informative local rounds from rare non-local ones.

[Chenggong "Alex" Zhang @ChenggongAlex]

1/5

🚀 New paper: Saturation-Insensitive Dueling Bandits with General Function Approximation

Can we learn from pairwise preferences without paying exponentially for BTL saturation? Answer: YES!

@ChenggongAlex @xuhengli_ @di_qiwei @WeitongZhang @QuanquanGu

🔗https://t.co/WASIeCfYza

[Chenggong "Alex" Zhang @ChenggongAlex]

4/5

🔥 The key insight: two-arm regret matters even if we ultimately care about ONE arm.

Small two-arm regret forces BOTH queried arms to be near-optimal ⇒ a small reward gap ⇒ unsaturated, informative comparisons.

Small one-arm regret alone gives no such guarantee.