Alban Ramette

Microbial ecology - genomic epidemiology - multivariate statistics

[Curiosity @CuriosityonX]

Milky Way is not only moving at 600 km/s, but also “flapping its wings” kinda like a butterfly🦋

[Tansu Yegen @TansuYegen]

Awesome👏

[Chris Laub @ChrisLaubAI]

MIT just published a paper that quietly explains why LLM reasoning hits a wall and how to push past it.

The usual story is that models fail on hard problems because they lack scale, data, or intelligence.

This paper argues something much more structural: models stop improving because the learning signal disappears. Once a task becomes too difficult, success rates collapse toward zero, reinforcement learning has nothing to optimize, and reasoning stagnates. The failure isn’t cognitive, it’s pedagogical.

The authors propose a simple but radical reframing. Instead of asking how to make models solve harder problems, they ask how models can generate problems that teach them.

Their system, SOAR, splits a single pretrained model into two roles: a student that attempts extremely hard target tasks, and a teacher that generates new training problems. The catch is that the teacher is not rewarded for producing clever or realistic questions. It is rewarded only if the student’s performance improves on a fixed set of real evaluation problems. No improvement means zero reward.

That incentive reshapes everything.

The teacher learns to generate intermediate, stepping-stone problems that sit just inside the student’s current capability boundary. These problems are not simplified versions of the target task, and strikingly, they do not even require correct solutions.

What matters is that their structure forces the student to practice the right kind of reasoning, allowing gradient signal to emerge even when direct supervision fails.

The experimental results make the point painfully clear. On benchmarks where models start with zero success and standard reinforcement learning completely flatlines, SOAR breaks the deadlock and steadily improves performance.

The model escapes the edge of learnability not by thinking harder, but by constructing a better learning environment for itself.

The deeper implication is uncomfortable. Many supposed “reasoning limits” may not be limits of intelligence at all. They are artifacts of training setups that assume the world provides learnable problems for free.

This paper suggests that if models can shape their own curriculum, reasoning plateaus become engineering problems, not fundamental barriers.

No new architectures, no extra human data, no larger models. Just a shift in what we reward: learning progress instead of answers.

[Alban Ramette @RametteA]

Microbial Genomics @MicrobioSoc
"Evaluation of Oxford Nanopore Technologies workflows for genomic epidemiology of outbreak-associated bacterial isolates in the clinical setting"
ONT vs. Illumina for bacterial outbreaks? The protocol matters.
📄 https://t.co/GLFXLd0UDU

[AI Research Tools 👨‍🎓 🧬🧪 🐬 🔬 @airesearchtools]

Accelerating Scientific Research with Gemini: Case Studies and Common Techniques

[bioRxiv Bioinfo @biorxiv_bioinfo]

PlotGDP: an AI Agent for Bioinformatics Plotting https://t.co/vjfELV03YL #biorxiv_bioinfo

[Hasan Toor @hasantoxr]

🚨BREAKING: Google just dropped another hit!

It's called PaperBanana and it generates publication-ready academic illustrations from just your methodology text.

No Figma. No manual design. No illustration skills needed.

Here's how it works:

A team of AI agents runs behind the scenes
→ One finds good diagram examples
→ One plans the structure
→ One styles the layout
→ One generates the image
→ One critiques and improves it

Here's the wildest part:

Random reference examples work nearly as well as perfectly matched ones. What matters is showing the model what good diagrams look like, not finding the topically perfect reference.

In blind evaluations, humans preferred PaperBanana outputs 75% of the time.

This is the recursion we've been waiting for AI systems that can fully document themselves visually.

Waitlist’s open, Link in the first comment.

[Alban Ramette @RametteA]

A new method called Cartesian Projections of Asymmetric Distances (CAPASYDIS) is introduced. It helps visualize the relationships between massive datasets of genetic sequences or "seqverses" by calculating asymmetric distances.

[Alban Ramette @RametteA]

Unveiling the Terra Cognita of Sequence Spaces using Cartesian Projection of Asymmetric Distances https://t.co/1yYaFTAqDb

[Alban Ramette @RametteA]

Unveiling the Terra Cognita of Sequence Spaces using Cartesian Projection of Asymmetric Distances https://t.co/1yYaFTAqDb

[Eric Topol @EricTopol]

"Why does the brain need billions of neurons to process 10 bits/second? Why can we only think about one thing at a time?"
by @jieyusz and @mameister4 @Caltech @NeuroCellPress https://t.co/5QMHIXoX4H

[Dr Emma Hodcroft @firefoxx66]

One of the reasons I've been quiet on Twitter for a few months is because I've been working on something big - something I'm really happy to share now!
https://t.co/m7KHsR3H8c

A new, open-source database for pathogen sequence sharing. 👐🧬🌎

1/12

[Alban Ramette @RametteA]

Open position: Bioinformatics / Biostatistics Staff scientist in Clinical Microbiology https://t.co/gWOBHB8hKq via @LinkedIn

[Alban Ramette @RametteA]

Component specific responses of the microbiomes to common chemical stressors in the human food chain https://t.co/YUt9ZMXfAO is out... Chemical stressors decreased microbiome diversity in soil and animal, but not in other microbiomes.

[Patrick Hsu @pdhsu]

Is DNA all you need?

In new work, we report Evo, a genomic foundation model that learns across the fundamental languages of biology: DNA, RNA, and proteins. Evo is capable of both prediction tasks and generative design, from molecular to whole genome scale.

[Alban Ramette @RametteA]

NASCarD: A Rapid, PCR-Free Method for SARS-CoV-2 Whole-Genome Sequencing in Clinical Samples https://t.co/7ehRSqXsjc #mdpipathogens via @Pathogens_MDPI a new approach allowing low amount of nucleic acids to be sequenced while selectively enriching for sequences of interest.

[Max Roser @MaxCRoser]

How many animals get slaughtered for meat every day?

My new @OurWorldInData article is now online:
https://t.co/pbzFBxpqNf

[Alban Ramette @RametteA]

@cyklee Have a look at:
https://t.co/y1L2VOH2CJ

[Reuters @Reuters]

Exclusive: Aspartame, one of the world's most common artificial sweeteners, is set to be declared a possible carcinogen next month by a leading global health body, pitting it against the food industry and regulators https://t.co/7Jv2XiJmAU

[Paul Graham @paulg]

I asked a friend who's a criminal defense lawyer what he'd change about the world to fix the most problems. He specializes in death penalty cases, so he's seen a lot. I expected him to say drugs, but he said if he could change one thing, it would be to make people better parents.