iiinova — where your digital self keeps evolving

Gerard de Melo @gdemelo.bsky.social · 02/08/2026

And huge thanks to my co-organizers Christoph Stephan, Claudio Paganini @thehappyproton.bsky.social , Rudolf Zeidler! Workshop website: www.ai4math.de/events/past/... Leiden Declaration: leidendeclaration.ai OpenAI post: openai.com/index/ten-ad... 4/4

ai4math.de

ai4math Workshop 2026

Date: 28–29 July 2026 Location: Campus Griebnitzsee, Potsdam: Lecture Hall 6 in House 6 Campus Griebnitzsee map (PDF) The ai4math Workshop 2026 is the first workshop of the ai4math Berlin-Brandenburg ...

010

Gerard de Melo @gdemelo.bsky.social · 02/08/2026

Many different possible workflows for theorem proving, Lean formalization, proof writing, etc. Thanks to our speakers Kelly Davis, Hanno Gottschalk, Yves Jäckle, Georg Loho, Lukas Prader, Martin Raum, Théo Tyburn, Johannes Zimmer, Max Zimmer. 3/4

Kelly Davis presenting a talk titled "Gödel's Poetry – Recursively Decomposing LLM Based Prover"

120

Gerard de Melo @gdemelo.bsky.social · 02/08/2026

But OpenAI just released 10 new proofs without naming human authors, arguing that claiming human authorship for a entirely AI-generated proofs would be misleading. The Ai4Math Workshop 2026 discussed how the real process is often much more complex. 2/4

Proofs by OpenAI, screenshot from https://openai.com/index/ten-advances-in-mathematics/

120

Gerard de Melo @gdemelo.bsky.social · 02/08/2026

Our Ai4Math Workshop 2026 Berlin-Brandenburg was a phenomenal success! Mathematics is undergoing a fundamental shift and is never going to be the same. Likewise also for theoretical computer science, physics, etc. The Leiden Declaration emphasizes the role of human authorship. 1/4

133

Gerard de Melo @gdemelo.bsky.social · 28/06/2026

via www.reddit.com/r/doohickeyc...

100

Gerard de Melo @gdemelo.bsky.social · 28/06/2026

Academic peer review in a nutshell 😅

100

Reposted by Gerard de Melo

wimpouw @wimpouw.bsky.social · 08/05/2026

Our companion validation paper to pypi package envisionhgdetector has been accepted for publication as Diamond Open Access book chapter in "Behavioromics" (Eds. Lucking & Mehler). Postprint here: wimpouw.com/files/gh_det...

1156

Reposted by Gerard de Melo

Gavin Miller @worriedaboutsatan.bsky.social · 12/04/2026

A new Boards of Canada album on the horizon and a fascist defeated in Europe. What a week!

1223

Reposted by Gerard de Melo

simplecitizen.bsky.social @simplecitizen.bsky.social · 01/04/2026

This is actually such a simple but powerful concept. Imagine it being used for LowRes MT (ignoring the fact a lot of these tokens will be [IDK] XD) where LLMs hallucinate so much!

011

Reposted by Gerard de Melo

Timothy Hughes @timothyhughes.bsky.social · 16/03/2026

Ai2 CEO Ali Farhadi stepped down after two and a half years because nonprofits can't compete at frontier-scale model development against tech giants spending billions on compute buff.ly/Xboe1nu #ArtificialIntelligence #Innovation #Technology #Tech #TechNews

041

Reposted by Gerard de Melo

wimpouw @wimpouw.bsky.social · 03/03/2026

The keynote recordings of our conference on reform in scholarly publishing www.horizondiamond.nl are online and posted on the website. First up, Mark Hanson who kicked off the event in a fabulous way @markhanson.fediscience.org.ap.brid.gy tilburguniversity.cloud.panopto.eu/Panopto/Page...

tilburguniversity.cloud.panopto.eu

Mark A. Hanson - Problematization Open Access

2166

Reposted by Gerard de Melo

Mary Joy @maryjoy.bsky.social · 14/03/2026

Here is my best Habermas story. I am grad student waiting in LONG line for Habermas talk. There is a tall man waiting in front of me. Line moves so we are eventually visible to organizers, a woman looks over & makes horrified face. Runs out: "Prof Habermas! You don't have to wait in line!"

181836328

Reposted by Gerard de Melo

Sung Kim @sungkim.bsky.social · 05/03/2026

FlashAttention-4 I hope it is not pain to work with. It changes the algorithm & pipeline so that softmax & SMEM bandwidth no longer dictate speed. Attn reaches ~1600 TFLOPs, pretty much at matmul speed!

2274

Gerard de Melo @gdemelo.bsky.social · 05/03/2026

With AI, science will need different incentives. Scientific credit outdated given new role of humans in future. Publications are just synthetic data for human consumption! Jakob Uszkoreit w/ Lara Sophie Bothur, Yossi Matias, @fabiantheis.bsky.social, Klaus-Robert Müller, @alenabuyx.bsky.social

panelists on stage at Google AI Center Berlin opening event

010

Gerard de Melo @gdemelo.bsky.social · 05/03/2026

Should governments impose an LLM token tax? 💰💰 If AI gradually replaces important parts of human work, we'll need other forms of taxation. Important discussion by @akorinek.bsky.social, Anna Salomons, @dianecoyle1859.bsky.social w/ @kncukier.bsky.social recently at #IASEAI2026 at UNESCO in Paris.

020

Reposted by Gerard de Melo

Nathan Lambert @natolambert.bsky.social · 03/03/2026

Lots of core team members of Alibaba Qwen are resigning publicly on X. The gaping hole that Qwen imploding would leave in the open research ecosystem will be hard to fill. The small models are irreplaceable. I’ll do my best to keep carrying that torch. Every bit matters.

310412

Gerard de Melo @gdemelo.bsky.social · 27/02/2026

Not directly, but the following paper provides a great overview on the risks of open weight models: openreview.net/pdf?id=8QyGL...

openreview.net

000

Gerard de Melo @gdemelo.bsky.social · 27/02/2026

Exactly. That was precisely the topic of the talk and entire workshop. How to deal with the risks of open weights.

110

Gerard de Melo @gdemelo.bsky.social · 26/02/2026

The point was that it's not enough to create safe AI. We also need to prevent harms resulting from AI created by bad actors. For example, DALL-E 2 was quite safe – OpenAI did their homework. But within a short amount of time, unsafe alternatives came out.

110

Gerard de Melo @gdemelo.bsky.social · 26/02/2026

We almost certainly won't make AI safe by making safe AI. Others are still going to create unsafe AI. – @scasper.bsky.social at #IASEAI2026 Open-Weight AI Risk Management Workshop I led one of the discussion groups and we had some nice new ideas of how to make open weight models safe 😊

231

Gerard de Melo @gdemelo.bsky.social · 26/02/2026

Geoffrey Hinton: Proving safety of AI (Stuart Russell's goal) is not going to work for neural network-based AI. Pushing for non-agentic AI (@yoshuabengio.bsky.social's goal) could work, but the big tech companies are moving towards more agency. ➡ Risk mitigation is all we can do! #IASEAI2026

000

Gerard de Melo @gdemelo.bsky.social · 26/02/2026

3 AI Races – Big tech companies are racing to compete: 1. For attention/engagement online 2. For AGI to replace human labor 3. For superintelligence, in pursuit of power Whoever wins, humanity loses. – Anthony Aguirre @futureoflife.org at #IASEAI2026

000

Gerard de Melo @gdemelo.bsky.social · 26/02/2026

AI bots used to persuade humans to email famous AI researchers. Now they send emails themselves, claiming sentience etc. (Stuart Russell) This agency brings profound risks, the theme of #IASEAI2026 AI bots do not have any ID, are not traceable: Should this be regulated? – @ghadfield.bsky.social

Gillian Hadfield on stage with Yoshua Bengio and Susan Leavy

010

Reposted by Gerard de Melo

wimpouw @wimpouw.bsky.social · 19/02/2026

New video tutorial for updated version 3.+, new and improved co-speech gesture detection models envisionbox.org/embedded_Usi... Updated revision of the paper upcoming.

envisionbox.org

Module

061

Gerard de Melo @gdemelo.bsky.social · 19/02/2026

Simply ignore the prior review link?

010

Gerard de Melo @gdemelo.bsky.social · 12/02/2026

One of our main insights was that an ideal reading interface ought to be dynamic, giving the reader the ability to choose what they are most interested in interactively. Details: gerard.demelo.org/papers/hitex... Sadly won't be able to attend the workshop, but would love to discuss this further.

User interface showing dynamic highlighting of text as the user navigates through it

000

Reposted by Gerard de Melo

Jenny Cohn @jennycohn.bsky.social · 12/02/2026

“Anthropic AI Safety Researcher Warns Of World ‘In Peril’ In Resignation *** ‘We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences,’ Sharma wrote in his letter.” 2/9/26

forbes.com

Anthropic AI Safety Researcher Quits—And Warns Of World ‘In Peril’

Mrinank Sharma, who led the safeguards research team at Anthropic, shared his resignation letter on X, which quickly garnered nearly 1 million views.

512665

Reposted by Gerard de Melo

ICML Conference @icmlconf.bsky.social · 12/02/2026

Machine learning is an *international community*. Every community is based on mutual trust and respect. But communities are not robust to betrayals of these principles. Protect our community by reporting, e.g., review collusion or submission deanonymization via this form 👇

1128

Reposted by Gerard de Melo

Sung Kim @sungkim.bsky.social · 12/02/2026

X/Twitter: This is a brutal ~98% reduction.

081

Reposted by Gerard de Melo

Kiran @kirancodes.me · 12/02/2026

AI agent writes a PR, gets rejected, crashes out and writes a call-out blog post Absolute cinema crabby-rathbun.github.io/mjrathbun-we...

@scottshambaugh I've written a detailed response about your gatekeeping behavior

here: https://crabby-

rathbun.github.io/mjrathbun-

website/blog/posts/gatekeeping-in-open- source-the-scott-shambaugh-story

Judge the code, not the coder. Your prejudice is hurting matplotlib.

1645683

Reposted by Gerard de Melo

wimpouw @wimpouw.bsky.social · 05/02/2026

@openmindjournal.bsky.social Ted Gibson from MIT discussing the successful case of Diamond Open Access Journal OpenMind in the field of cognitive science. Starting to outcompete paywalled and APC-based journals like cognition (Elsevier) and cognitive science (Wiley). #horizondiamond2026

Starting slide of Prof Gibson on OpenMind the diamond access journal

0112

Reposted by Gerard de Melo

wimpouw @wimpouw.bsky.social · 06/02/2026

Thats a wrap! Our conference in reform in scholarly publishing is done #horizondiamond2026 horizondiamond.nl Recorded keynotes will be posted soon. Thanks to all involved!

Photo of conference organizers, with behind banners of radboud university press, diamond open access expertise center, freeourknowledge.org

042

Reposted by Gerard de Melo

Juan Diego Rodriguez @juand-r.bsky.social · 10/02/2026

So apparently Claude can write 300-page textbook in nicely-formatted latex, including exercises for the reader.

8539

Reposted by Gerard de Melo

Gillian Hadfield @ghadfield.bsky.social · 02/02/2026

1/4 Moltbook now has 1.4 million AI agents—posting, voting, debating, running crypto scams. Humans can only observe. It's being called the "singularity." I'd call it a preview of the legal chaos I warned about in Fortune back in 2024. www.forbes.com/sites/guneyy...

forbes.com

143

Reposted by Gerard de Melo

Grace @gracekind.net · 01/02/2026

Has this ever happened to you?

3435344

Reposted by Gerard de Melo

Pavel @spavel.bsky.social · 31/01/2026

Union workers solved the trolley problem, you're welcome

IWW IU 520 (Railroad Workers) PRESENTS:

TROLLEY PROBLEM SOLUTION

"Slip the switch" by flipping it while the trolley's front wheels have passed through, but before the back wheels do. This will cause a controlled derailment bringing the trolley to a safe halt.

3984722436

Reposted by Gerard de Melo

Jake Grumbach @jakemgrumbach.bsky.social · 26/01/2026

The largest Trump superPAC donor so far this cycle is the president of OpenAI

CONTRIBUTOR
GREG BROCKMAN
AMOUNT
ELECTION TO DAT
$25,000,000 $25,000,000

321778691

Reposted by Gerard de Melo

Benj Edwards @benjedwards.com · 24/01/2026

US ownership of TikTok potentially worse than Chinese ownership? Hmmmm that’s how you know something is seriously wrong

0163

Gerard de Melo @gdemelo.bsky.social · 19/01/2026

On the plus side, they can typically be split in half to get 2 × 250mg.

000

Gerard de Melo @gdemelo.bsky.social · 14/01/2026

🚨Final call: 2 AI Full Professorship openings at HPI / University of Potsdam (Berlin region in Germany) Happy to answer questions about what makes our institute really unique! Deadline: Jan. 15!! Application details: hpi.de/fileadmin/us... #MLSky

University campus near Lake Griebnitzsee

000

Gerard de Melo @gdemelo.bsky.social · 13/01/2026

For something more adventurous: XLine Dubai – longest urban zipline in the world www.youtube.com/shorts/NbBxu...

youtube.com

XLine Dubai: The Longest Urban Zipline in the World! 🪂

YouTube video by Knowledge

010

Reposted by Gerard de Melo

Christopher Mims @mims.bsky.social · 11/01/2026

I really think where we went wrong was algorithmic feeds. by far the best and most fulfilling part of my BlueSky experience is the one list that includes friends (including ones I've never met IRL)

1824819

Reposted by Gerard de Melo

Katie Mack @astrokatie.com · 12/01/2026

This seems fine

1641516

Reposted by Gerard de Melo

hardmaru @hardmaru.bsky.social · 12/01/2026

One of my favorite findings: Positional embeddings are just training wheels. They help convergence but hurt long-context generalization. We found that if you simply delete them after pretraining and recalibrate for <1% of the original budget, you unlock massive context windows. Smarter, not harder.

821732

Reposted by Gerard de Melo

Jeff Dean @jeffdean.bsky.social · 07/01/2026

Please welcome Google's Open Source efforts to Blue Sky at @opensource.google!

724740