Mark Hahnel

Figshare Founder

[Mark Hahnel @MarkHahnel]

What parts of the academic knowledge creation & dissemination pipeline can we automate with AI?

https://t.co/OkltGhB5BQ

[Ai2 @allen_ai]

LLMs are no longer created w/ human data alone. They rely on other models to generate & filter data, evaluate outputs, & guide dev work.

So what is a modern LLM built on? Olmo 3 → 89 model + 183 dataset dependencies; Nemotron 3 → 273 + 560

We made ModSleuth to trace this. 🧵

[Mark Hahnel @MarkHahnel]

Ground the model, or it invents the evidence

Trusted, curated, subject-specific models are needed for academia

https://t.co/xjBVT1YrWt

[Digital Science @digitalsci]

Today! Are you joining the science sleuths?

🔗 Register here: https://t.co/RhiyBJF5ib

#ResearchIntegrity

[Digital Science @digitalsci]

The #arXiv ban on unchecked #AI in preprints is, in @MarkHahnel's words, needed & appropriate. But the underlying problem is bigger than arXiv.

General-purpose models are the wrong tool for research. The contamination, he warns, doesn't go away when the AI gets better.

🔗 Read his post: https://t.co/lMFgS3F7Jl

[Digital Science @digitalsci]

"The tools have changed beyond recognition; the intent has not changed at all." — Dr Daniel Hook, CEO, Digital Science.

In a new blog post on the @Symplectic website, Daniel discusses a 20-year-old problem that #AI is now helping to solve.

🔗 Read his post: https://t.co/YVjHoOU4Gc

[Symplectic @Symplectic]

Still dealing with “alphabet soup” in your research systems? RIMS, CRIS, IR, RDM… it adds up fast.

Watch Building a Research Engine webinar on demand to see how you can simplify your ecosystem with Symplectic Elements and Figshare.

👉 https://t.co/XowBEGYN9C

[Rory Byrne @ryrobyrne]

Operating a bio database should be much easier. To the point where each lab can run one. These should integrate with lab equipment, run automated QC on new deposits, compute a rich surface of queryable metadata, export to ML-friendly formats, and federate via shared ontologies.

[Mark Hahnel @MarkHahnel]

@mgdurrant Any way to get on trusted access program now?

[clem 🤗 @ClementDelangue]

Concentration of power, capabilities and economic wealth is the biggest risk in AI. We need open science and open-source more than ever!

[Training Centre in Communication (TCC Africa) @tccafrica]

Ongoing Figshare & TCC Africa Community Call in #Kenya: Strengthening Research Data Management & Data Repository Adoption in #OpenScience using @figshare . Speaking now is @MarkHahnel
CC @digitalsci

[Mark Hahnel @MarkHahnel]

Love this - great to see lots of green - could really do with more

The state of biological AI at a glance

https://t.co/bf6XUK2N0v

[AVB @neural_avb]

Been using the Overleaf AI Agent for latex editing the past few days... It's way more useful than I expected! 🙏🏼

[Digital Science @digitalsci]

👉 Before you attend the webinar, download the report:

"Forensic Scientometrics (FoSci) Report 2026: Understanding, Detecting, and Documenting Manipulation in the Research Ecosystem."

🔗 https://t.co/alnAJXr3kW

#ResearchIntegrity #TrustInScience

[Digital Science @digitalsci]

⚠️ This week: Join our panel of science sleuths to hear about uncovering research misconduct & manipulation. #ResearchIntegrity

We're decoding the findings of the Forensic Scientometrics (FoSci) Report.

🗓️ Thursday 11 June
🕒 3pm BST 🕙 10am EDT

🔗 Register now: https://t.co/JCWm92Ae3Z

[Mark Hahnel @MarkHahnel]

It’s a monumental shift that Tech companies can now start Pharma companies

A new era of biological compute, biosingularity and solving disease

@maxjaderberg talking about how AI is transforming drug discovery at London Tech Week

[Anthropic @AnthropicAI]

New Science Blog: Why has AI advanced faster in coding than in biology?

To agents, bio databases are like cities built before cars—maddening to drive in because they're designed for different traffic.

How do we build infrastructure agents can use?
https://t.co/PQaNQ4GRJZ

[Nathan C. Frey @nc_frey]

Opus 4.7 is as good or better than ChemDraw at interpreting NMR spectra on our evals. We're making Claude more helpful for chemists, starting with routine and time-consuming analytical tasks.

[Mark Hahnel @MarkHahnel]

On Training Data for Bio AI Models

As we advance biological foundation models, which lessons from LLM data curation transfer, and which need rethinking?

https://t.co/BPDZrJMq5d

[Mark Hahnel @MarkHahnel]

NIH Generalist Repository Ecosystem Initiative (GREI) Year 4: Reflections, Progress, and What Comes Next

https://t.co/lgqh0VTKVs

[Mark Hahnel @MarkHahnel]

Biology Open paid peer reviewers £220 and cut peer review times to five working days

https://t.co/FkO8Fl84NY