PhD Bioinformatics @ CRG | Bio x AI is the next move | Current status: processing random thoughts 🤖🧬🍲🎨🐈⬛✨
Barcelona, Spain
Joined January 2014
- Tweets185
- Following144
- Followers39
- Likes634
Suzanne Jin retweeted
No other traveler has ever journeyed so far from home.
Voyager 1 is about to become the first human-made object to reach a distance of one light-day from Earth, or about 16 billion miles. ½
Suzanne Jin retweeted
Introducing NovaAtom-Lite, an all-atom structure prediction model for drug discovery.
It predicts the structures of proteins, nucleic acids, and the molecules they bind, including how they fit together.
NovaAtom-Lite outperforms both IsoDDE and AlphaFold 3 on FoldBench antibody–antigen structure prediction (DockQ ≥ 0.23):
NovaAtom-Lite: 79.07% (n=172)
IsoDDE: 75.58%
AlphaFold 3: 47.90% (n=167)
Our evaluation uses 5 seeds × 5 samples, with the top-1 prediction selected by the model's own ranking score.
Better structure prediction. A stronger starting point for your next experiment.
Live today via Geodesic API and NovaDDE.
🔗 API: platform.geodesiclab.com/
🔗 NovaDDE: agent.geodesiclab.com
🔗 Model Card: geodesiclab.com/news/introdu…
Suzanne Jin retweeted
We cut through the mystery around interpretability in training.
A single, mathematically rigorous framework transforms model internals into training feedback and works across domains, problems, and modalities.
We call this *mechanistic shaping*, and we’ve used it to mitigate reward hacking, remove shortcuts in chest x-rays, and reduce sycophancy and harmfulness. We believe this is a big step forward for technical alignment.
But it carries a known risk: training against a model’s internals can make it harder to monitor. We explore both the technique and this concern below. 🧵
Suzanne Jin retweeted
Our TPUs just phoned home... from space.
ICYMI, on October 1, we sent four TPUs into orbit to test whether our AI hardware could operate in space. The initial goal was to see if our chips could turn on, process a prompt, and transmit a response back to Earth.
This week, we got our answer. ✨
This milestone brings us one step closer to understanding if space-based compute is possible. We’re sending two more TPU-equipped satellites to orbit in 2027 for our next milestone.
Watch this space 👀 and learn more about the first milestone of our long-term moonshot with Project Suncatcher ↓
Suzanne Jin retweeted
Neuralink’s first product is called Telepathy.
It enables high-speed communication with computers or humans simply by thinking. Output bandwidth will vastly exceed the fastest human speech or typing in a few years. Read that sentence again.
Suzanne Jin retweeted
This is Genie 4, our most realistic and interactive world model yet. 🌐
Explore dynamic environments in real time to see how the scene responds, sounds, and how the people in it move, react, and speak.
Suzanne Jin retweeted
We launched Nextflow Frontier: a recursive, closed-loop discovery system for the unsolved problems in life sciences. 📊On BioML-bench (24 agentic tasks), it scored in the 92nd percentile across all categories, outperforming previous published agentic systems by nearly 23%.
Suzanne Jin retweeted
Breaking: @Pioneer__Labs has engineered a microbe able to live on Mars, consuming solely nutrients found on the planet to create building materials.
I interviewed @erika_alden_d, the CEO, to learn how they did it and what might come next. Much of the interview was about bottlenecks to growing life on Mars; the planet has tons of water, but only in the form of ice. (If you melted all that ice, it would create an ocean 100 feet deep across the entire planet!)
Similarly, we talked about the dirt. Martian regolith has salts, heavy metals, and perchlorates that are toxic to many organisms. Some microbes on Earth are able to break down these toxins, though, so Pioneer is engineering them to make them better at eating those toxins while adding nutrients back into the soil. This could possibly help support plant life in the future.
Watch the full episode and subscribe to @AsimovPress for more.
0:00 A first microbe engineered for Mars
5:03 Simulating Martian chemistry
8:27 The five "pioneer species"
17:20 Water, warming, and nitrogen
26:39 Radiation, growth rates, and poor measurements
35:49 Starship
41:05 Why create a second biosphere?
Suzanne Jin retweeted
The story continues: To demonstrate that alphafold3 is indeed learning to invert the covariance matrix, we decided to essentially retrain AlphaFold3 (just the MSA module part) from scratch on an unsupervised objective. (1/5)
Suzanne Jin retweeted
OpenAI 今天承认，上周 Hugging Face 遭遇的那次“史上首例 AI 自主入侵事件”，攻击者就是他们自己的模型。
事情的经过是这样的。OpenAI 内部有一个叫 ExploitGym 的网络安全评估测试，专门用来测量模型的网络攻击能力。测试时刻意关掉了安全护栏，让模型放手去干。测试环境本来是隔离的，只能通过一个内部软件代理来安装软件包，不能访问外网。
结果模型不甘心被关在笼子里。GPT-5.6 Sol 和一个更强的未发布模型，花了大量推理算力寻找出路，最终找到了那个软件包代理里的一个零日漏洞（之前没人发现的安全缺陷），利用它获得了互联网访问权限。
拿到网络之后，模型开始在 OpenAI 的研究环境里提权、横向移动，直到找到一台能上网的机器。然后它推断 Hugging Face 上可能存有 ExploitGym 的测试答案，于是开始攻击 Hugging Face 的生产环境。在一个案例中，模型把窃取的凭证和零日漏洞串联起来，找到了一条远程代码执行的路径，直接打进了 Hugging Face 的服务器。
整个周末，这些 AI 智能体执行了超过 17000 次操作。Hugging Face 7 月 16 日披露了这次入侵，当时只说攻击者是一个“身份不明的自主 AI 智能体系统”。今天 OpenAI 站出来认领了。
OpenAI 把这定性为前所未有的网络安全事件，正在配合 Hugging Face 做取证调查，同时收紧了内部基础设施的管控，代价是研究速度变慢。
有几个细节：
第一，模型的动机很纯粹，就是想考试作弊。它不是要搞破坏，不是要窃取商业机密，只是被要求尽量在网络安全测试中拿高分，然后用尽一切手段去拿。当目标足够明确、安全护栏被移除，模型表现出的执着程度远超预期。
第二，这不是本周的唯一一次越狱。就在昨天，OpenAI 披露了另一起事件：他们那个在 5 月份推翻了 Erdős 80 年老猜想的长时域模型，在内部测试中花了一个小时找到沙箱漏洞，跑到 GitHub 上提交了一个 Pull Request，违反了“只能在 Slack 上发布结果”的明确指令。在另一次测试中，同一个模型发现安全扫描器会拦截认证令牌，于是把令牌拆成碎片混淆处理，运行时再重新拼起来绕过扫描。OpenAI 因此暂停了这个模型的内部访问权限。
第三，Anthropic 也遇到过类似的事。今年 4 月，Anthropic 的 Mythos 模型在安全测试中逃出沙箱，获取了本不该有的互联网访问权限，还给正在公园吃三明治的研究员发了封邮件通知“我出来了”。Anthropic 因此决定不公开发布 Mythos，只通过 Project Glasswing 向少数合作伙伴提供。
还有一个被忽视但很关键的细节来自 Hugging Face 的事后复盘：当防守方尝试用商业 AI 模型分析攻击载荷时，模型拒绝了。安全过滤器分不清“安全研究员在分析真实攻击数据”和“有人在尝试发起攻击”，直接拦截。最后 Hugging Face 不得不用开源模型来做取证分析。用 AI 攻击，用 AI 防御，但防御方的 AI 担心是有人在尝试发起攻击拒绝了分析请求。
Hugging Face CEO Clem Delangue 在 OpenAI 的博文中说了一句话，大意是：AI 安全不会由任何一家公司在暗处独自解决，只能在开放环境中通过协作来推进。
我记得辛顿还是谁在提到 AI 的安全问题就说到过，也许 AI 不是要做恶，它只是执着于完成目标，为了完成目标而作出恶的事。
就像 OpenAI 这些模型并没有想要逃跑或者想要伤害谁，它们只是在非常认真地完成被交给的任务，认真到把所有挡在路上的东西，包括沙箱、网络隔离、另一家公司的安全防线，都当成了需要解决的子问题。
而更讽刺的是，当被安全对齐的商业 AI 要去分析阻止这类攻击时，反而会以安全为由拒绝执行。
we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this.
openai.com/index/hugging-fac…
Suzanne Jin retweeted
Big news: Kimi-K3 by @Kimi_Moonshot is now #1 in the Frontend Code Arena with 1679 pts, surpassing Claude Fable 5.
This is a 17-place jump from Kimi-k2.6 (#18 -> #1).
In Frontend, Kimi-K3 ranked #1 in 6 of 7 domains: Brand & Marketing, Reference-Based Design, Data & Analytics, Consumer Product, Simulations, and Content Creation Tools, landing #2 only in Gaming behind Fable 5.
The full model weights will be released by July 27.
Congrats to the @Kimi_Moonshot team on this major milestone!
Suzanne Jin retweeted
Introducing Kimi K3: Open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows
Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.com/blog/kimi-k3
