Matt McCoy

Founder @ Resynthesis Global Director @ Proofig AI.

[Matt McCoy @mccoymf]

AI may make mathematical proofs abundant. But abundance is not understanding.

The real bottleneck will be verification, exposition, and whether the community can absorb results into durable knowledge.

Terence Tao’s lecture is excellent on this.

https://t.co/2dbxo3PWCc

[Matt McCoy @mccoymf]

I recently presented the concept of Augmented Integrity at WCRI. This platform update brings that concept to life.

The next stage in image integrity is not simply improved detection. It is the cultural and operational integration of explainable, structured adjudication - enabling reviewers at scale to move from identifying an anomaly to understanding what the evidence can reasonably support.

Congratulations to the Proofig team for continuing to push the envelope.

[Proofig AI @Proofig]

Image-integrity review used to need experts. Tools like Proofig AI changed that: anyone can now surface a suspected duplication or manipulation.
The harder part, judging whether a flag is a real problem or legitimate, still takes expertise.
Learn more:
https://t.co/IQQ1SoYeWj

[Matt McCoy @mccoymf]

The danger here isn’t just less money for science.

It’s the quiet replacement of peer review with political permission.

Once research has to flatter power to be funded, science stops being a truth-seeking institution and becomes another branch of communications.

https://t.co/7o4uAlkDAN

[Matt McCoy @mccoymf]

The Max Planck retractions are a perfect miniature of the problem: an algorithm may have mistaken historical republication practices for misconduct, then erased the paper rather than preserving it. Research integrity without historical intelligence becomes its own kind of distortion.

https://t.co/9tIouADvHo

[Matt McCoy @mccoymf]

The university made a “brave decision” by sanctioning Klein, and it might increase people’s trust in scientific institutions to see that even someone of his fame can be punished for misconduct. “It’s a sign that the institution works.”

https://t.co/1H4dsAeLoO

[Berkeley Lab Press @LBNLpress]

"This breakthrough will be truly revolutionary if it is able to solve unique structures of small proteins, particularly those inside frozen cells," says @berkeleylab's Peter Ercius. @amiposts @LindaFStewart https://t.co/UMom7V9CiG

[Matt McCoy @mccoymf]

Who gets to control scientific discussion when funding, policy, and government priorities become part of the conversation?

https://t.co/lxLcdDHEFc

[Matt McCoy @mccoymf]

I spend a lot of time teaching and consulting with a whiteboard marker in my hand.

This paper suggests it might be time for a change.

"Our findings suggest that long-term exposure to whiteboard markers may pose health risks."

https://t.co/1hQbzlu4kp

[Matt McCoy @mccoymf]

We've spent decades trying to detect cancer earlier.

This paper asks a different question:

Can we identify the biology that precedes cancer years before diagnosis?

Researchers identified a 14-protein blood signature that predicts lung cancer risk more than five years before diagnosis - and may help target preventative intervention.

The future of oncology may be measured not by cancers detected, but by cancers prevented.

https://t.co/AvIeMXtCBC

[Matt McCoy @mccoymf]

More than 150 mathematicians have signed the Leiden Declaration on AI and Mathematics.

The concern isn't that AI will fail.

It's that it will succeed.

When mathematical discovery accelerates, how do we preserve understanding, attribution, accountability, and trust?

https://t.co/6edNxSCIvA