Qinyue Zheng

I build. AI PhD Student @ETH Zürich | Prev @EPFL, @Harvard, @llama_index

[Michael Moor @Michael_D_Moor]

Running AI agents has become easy. Observing them is the hard part.

Let one run 24/7 for days and your terminal becomes a wall of scrollback no human can ever read. We don't know what it actually did, where it stalled, or whether it quietly went off the rails.
1/

[Paul Graham @paulg]

Brevity is more than politeness to the reader. Compression is understanding.

[Sam Rawal, MD @samarthrawal]

Made a Chrome extension that auto-generates Anki cards while I do practice clinical questions, to save key concepts - I can see the friction removal from this adding up across dozens of questions!

runs on Codex App Server - so no additional API costs to run the extension!

[Michael Moor @Michael_D_Moor]

What if AI-generated CT reports weren’t black boxes?
Our new agent shows every step w/ improvements in accuracy, robustness, and faithfulness.

[Michael Moor @Michael_D_Moor]

[Preprint alert] Process reward agents (PRA)

Why is it relatively easy to get LLMs to produce strong reasoning traces in math/code…
but much harder in application domains like health?

Check out our new paper & 🧵below:

[Garry Tan @garrytan]

Vibe coding

[Michael Moor @Michael_D_Moor]

Wanna do a postdoc in my group?

Check out this position on "Language Models for Pediatric Data Analysis" in collaboration with the University Children's Hospital Basel:

https://t.co/gsymxRdiVE

[Frederike Lübeck @FrederikeLubeck]

Sharing our latest work: Reinforcement Learning via Self-Distillation

We introduce Self-Distillation Policy Optimization (SDPO), an on-policy RL algorithm that turns feedback into dense, token-level learning signals.

[Andreas Krause @arkrause]

SDPO enables RL agents to learn from rich feedback (i.e., not only whether an attempt failed, but why it failed, such as error messages). Even without such rich feedback, SDPO can reflect on past attempts and outperform GRPO. SDPO accelerates solution discovery at test time!

[Andrej Karpathy @karpathy]

A few random notes from claude coding quite a bit last few weeks.

Coding workflow. Given the latest lift in LLM coding capability, like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in December. i.e. I really am mostly programming in English now, a bit sheepishly telling the LLM what code to write... in words. It hurts the ego a bit but the power to operate over software in large "code actions" is just too net useful, especially once you adapt to it, configure it, learn to use it, and wrap your head around what it can and cannot do. This is easily the biggest change to my basic coding workflow in ~2 decades of programming and it happened over the course of a few weeks. I'd expect something similar to be happening to well into double digit percent of engineers out there, while the awareness of it in the general population feels well into low single digit percent.

IDEs/agent swarms/fallability. Both the "no need for IDE anymore" hype and the "agent swarm" hype is imo too much for right now. The models definitely still make mistakes and if you have any code you actually care about I would watch them like a hawk, in a nice large IDE on the side. The mistakes have changed a lot - they are not simple syntax errors anymore, they are subtle conceptual errors that a slightly sloppy, hasty junior dev might do. The most common category is that the models make wrong assumptions on your behalf and just run along with them without checking. They also don't manage their confusion, they don't seek clarifications, they don't surface inconsistencies, they don't present tradeoffs, they don't push back when they should, and they are still a little too sycophantic. Things get better in plan mode, but there is some need for a lightweight inline plan mode. They also really like to overcomplicate code and APIs, they bloat abstractions, they don't clean up dead code after themselves, etc. They will implement an inefficient, bloated, brittle construction over 1000 lines of code and it's up to you to be like "umm couldn't you just do this instead?" and they will be like "of course!" and immediately cut it down to 100 lines. They still sometimes change/remove comments and code they don't like or don't sufficiently understand as side effects, even if it is orthogonal to the task at hand. All of this happens despite a few simple attempts to fix it via instructions in CLAUDE . md. Despite all these issues, it is still a net huge improvement and it's very difficult to imagine going back to manual coding. TLDR everyone has their developing flow, my current is a small few CC sessions on the left in ghostty windows/tabs and an IDE on the right for viewing the code + manual edits.

Tenacity. It's so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later. You realize that stamina is a core bottleneck to work and that with LLMs in hand it has been dramatically increased.

Speedups. It's not clear how to measure the "speedup" of LLM assistance. Certainly I feel net way faster at what I was going to do, but the main effect is that I do a lot more than I was going to do because 1) I can code up all kinds of things that just wouldn't have been worth coding before and 2) I can approach code that I couldn't work on before because of knowledge/skill issue. So certainly it's speedup, but it's possibly a lot more an expansion.

Leverage. LLMs are exceptionally good at looping until they meet specific goals and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria and watch it go. Get it to write tests first and then pass them. Put it in the loop with a browser MCP. Write the naive algorithm that is very likely correct first, then ask it to optimize it while preserving correctness. Change your approach from imperative to declarative to get the agents looping longer and gain leverage.

Fun. I didn't anticipate that with agents programming feels *more* fun because a lot of the fill in the blanks drudgery is removed and what remains is the creative part. I also feel less blocked/stuck (which is not fun) and I experience a lot more courage because there's almost always a way to work hand in hand with it to make some positive progress. I have seen the opposite sentiment from other people too; LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building.

Atrophy. I've already noticed that I am slowly starting to atrophy my ability to write code manually. Generation (writing code) and discrimination (reading code) are different capabilities in the brain. Largely due to all the little mostly syntactic details involved in programming, you can review code just fine even if you struggle to write it.

Slopacolypse. I am bracing for 2026 as the year of the slopacolypse across all of github, substack, arxiv, X/instagram, and generally all digital media. We're also going to see a lot more AI hype productivity theater (is that even possible?), on the side of actual, real improvements.

Questions. A few of the questions on my mind:
- What happens to the "10X engineer" - the ratio of productivity between the mean and the max engineer? It's quite possible that this grows *a lot*.
- Armed with LLMs, do generalists increasingly outperform specialists? LLMs are a lot better at fill in the blanks (the micro) than grand strategy (the macro).
- What does LLM coding feel like in the future? Is it like playing StarCraft? Playing Factorio? Playing music?
- How much of society is bottlenecked by digital knowledge work?

TLDR Where does this leave us? LLM agent capabilities (Claude & Codex especially) have crossed some kind of threshold of coherence around December 2025 and caused a phase shift in software engineering and closely related. The intelligence part suddenly feels quite a bit ahead of all the rest of it - integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability.

[Qinyue Zheng @QueyJ]

@samarthrawal Will VLMs get closer than LLMs?

[Qinyue Zheng @QueyJ]

🤔

[Alex Veremeyenko @alex_verem]

🚨 RAG is broken and nobody's talking about it.

Stanford just exposed the fatal flaw killing every "AI that reads your docs" product.

It's called "Semantic Collapse", and it happens the moment your knowledge base hits critical mass.

Here's the brutal math (and why your RAG system is already dying):

[Yulun Jiang @YulunJiang]

Meta-RL Induces Exploration in Language Agents

Exploration is essential for LLM agents, but how can we train an agent that actively explores?

Introducing 🌊LaMer, a general Meta-RL framework that enables LLM agents to explore and learn from the environment feedback at test time.

Key idea:
(i) Cross-episode training to encourage exploration and long-term reward optimization
(ii) In-context policy adaptation via self-reflection, allowing agents to update behavior from feedback without gradient updates

LaMer trains the agents to explore in the early episodes and adapt to the environment in the subsequent episodes, leading to:
(i) More diverse trajectories of trained agents compared to RL baselines
(ii) Stronger performance and test-time scaling
(iii) Better generalization to hard or out-of-distribution tasks

📜Paper: https://t.co/BqpO0AR6d8
💻Code: https://t.co/ZmO06VOXWZ

Had fun collaborating with @LiangzeJ and his advisor @DamienTeney, and huge thanks to my amazing advisors @mariabrbic and @Michael_D_Moor

[Michael Moor @Michael_D_Moor]

🚀 New PhD position in my group:

https://t.co/mrBQOkmy4v

We’re hiring a doctoral student at ETH Zurich (located in D-BSSE, Basel) to work on medical reasoning.

🌍 Fully funded PhD as part of the EU-funded Marie-Curie project "MLCARE".

Our group has access to high-end GPU clusters, we are embedded in the life science hub of Switzerland in Basel, are involved with the ETH AI Center and SwissAI projects. For more details on the position, check out the link.

📩 Pease share!

[Qinyue Zheng @QueyJ]

@fchollet Insight is just one of the possibilities

[Sam Rawal, MD @samarthrawal]

Med friends -- I'm releasing 🔵 Pearls, an app I built for capturing clinical pearls while on rounds or studying.

Once you add a pearl, an AI agent will automatically fetch additional context, latest guidelines, and augment the pearl.

Other features: free sync between devices, auto-tagging, study mode.

More broadly, I'm exploring some ideas on how LLMs can improve the ability of residents to absorb and retain higher amounts of information; this is one small practical exploration of those ideas.

[Saibo-Creator @SaiboGeng]

🚀 Excited to share our latest work at ICML 2025 — zip2zip: Inference-Time Adaptive Vocabularies for Language Models via Token Compression!

Sessions:

📅 Fri 18 Jul 
- Tokenization Workshop

📅 Sat 19 Jul 
- Workshop on Efficient Systems for Foundation Models (Oral 5/145)

[Qinyue Zheng @QueyJ]

Thanks @qdrant_engine for featuring MIRIAD! 🫶

Always a great team effort with @salmanabdullah_

🚀 5.8M+ medical QA pairs to support grounding LLMs in reliable medical knowledge! Check it out it’s fun!

 📂Dataset: https://t.co/NeoqY8Wsio
 🪐Demo: https://t.co/nRU3D7OdHA

[Qdrant @qdrant_engine]

Researchers at @ETH_en and @Stanford released an open dataset of 5.8M+ long-form medical QA pairs, each grounded in peer-reviewed literature and designed for RAG. 🚀

The pipeline:
▪️ Source: 900K+ full-text medical papers (S2ORC)
▪️ QA generation via GPT-3.5 with a three-stage filtering process (regex, Mistral-7B classifier, human-in-the-loop)
▪️ Embeddings generated and indexed in Qdrant for scalable dense retrieval

The dataset is available on @huggingface🤗 with full code for embedding, indexing, and RAG setup.

👉 Full story: https://t.co/JunG8aWS8O

[Kyle Corbitt @corbtt]

Big news: we've figured out how to make a *universal* reward function that lets you apply RL to any agent with:
- no labeled data
- no hand-crafted reward functions
- no human feedback!

A 🧵 on RULER

[Qinyue Zheng @QueyJ]

@samarthrawal This is awesome