Tiago Peixoto

Tiago Peixoto

@tiago.skewed.de

Statistical mechanic, secular Bayesian. Prof. of Complex Systems and Network Science @ Goethe University Frankfurt Head of the “Inverse Complexity Lab”. @invcomplexity.skewed.de https://skewed.de/lab (Personal account)

4.9K followers747 following838 posts

Pinned

Good news everyone! 🎉 The new version 3.0 of graph-tool is just out with major improvements! See below. graph-tool.skewed.de graph-tool is a comprehensive and efficient Python library to work with networks, including structural, dynamical, and statistical algorithms, as well as visualization. 1/N

The whole vote count in Brazil (pop 213M) took only a few hours, where the majority flipped twice mid count. Imagine this dragging on for days and weeks like in the US.

Stunning comeback from Lula in just a few weeks! I was hopeful but not very optimistic. Interesting how the projections got it right even as Lula was trailing behind with 30% of the votes counted! Now we get to have a little more hope for humanity, and ananother 4 years of basic sanity.

I don't mind people doing this for their own purposes, but what is the point of *publishing* it? What is the meaning of scholarship and authorship if you can't precisely answer for technical details?

💀Adina Feinstein PhD 💀@afeinstein20.bsky.social

interestingggg.... paper on the arXiv tonight of a new planet candidate found with and written up entirely by AI, including the caveat, "The author has not independently checked the analyses and does not have the expertise to verify every technical detail." arxiv.org/abs/2610.28575

If you typeset it, you get one step closer to nirvana.

Richard McElreath 🐈‍⬛@rmcelreath.bsky.social

Once again grateful that one can get TeX source from arXiv papers, so that I didn't have to typeset Ramanujan's mad formula for pi myself (just an intro slide in a talk I am giving tomorrow)

I think people really should understand — especially young researchers — that if you use LLMs to draft most of the text of your paper, then: 1. It is often entirely obvious that you did just that. 2. The quality of the generated text is often terrible, and it hurts to read it.

Tiago Peixoto reposted

Approximating the Lorenz attractor with a chaotic leaky water wheel. The orange dot is the centre of mass of the water. It traces (a projection of) the familiar attractor as the wheel spins seemingly randomly left and right.

At ~$0.07/kWh for electricity alone, $22M buys ~310 million kWh. A human spending 100W on average consumes ~64,000 kWh in 75 years. So $22M corresponds to the lifetime metabolic energy of roughly 4,800 people. This is brute force, not intelligence.

Tiago Peixoto@tiago.skewed.de

Some humans were already fooled by Eliza-like programs decades ago. I don't think we stand much of a chance against data-driven brute force at scale mascarading as intelligence. Most people can't intuitively grasp what $22M of computing power means. The AGI propaganda will be difficult to beat.

Some humans were already fooled by Eliza-like programs decades ago. I don't think we stand much of a chance against data-driven brute force at scale mascarading as intelligence. Most people can't intuitively grasp what $22M of computing power means. The AGI propaganda will be difficult to beat.

Trevor A. Branch@trevorabranch.bsky.social

Millennium math problem solved by OpenAI may have "regurgitated" competitors unpublished work... also they spent more than $22 million in computing power in one week to scoop them. techcrunch.com/2026/09/08/o...

Tiago Peixoto reposted

And you have to ask yourself: why was it worth $22M to them? Because to them math is just their most successful PR bid, (since the claims of AGI aren't sticking). They are betting that the (false) claim "We solved math, so we can solve everything" will gain them way more than $22M in market share.

Trevor A. Branch@trevorabranch.bsky.social

Millennium math problem solved by OpenAI may have "regurgitated" competitors unpublished work... also they spent more than $22 million in computing power in one week to scoop them. techcrunch.com/2026/09/08/o...

“Almost everyone who ever lived was born recently, because population grew exponentially, so a random birth is far more likely to fall near the present.” According to this logic, either we are all outliers or our individual presence is this date is evidence for the end of the world coming soon.

Julia M. Rohrer@dingdingpeng.the100.ci

Fun little game in which a human life is randomly drawn for you. I ended up a girl, born in the Yangtze Basin in 670 CE. Died in infancy at 6 months 🥲 anyhumanever.com

An AI “hallucinating” is still pretty much science fiction, since it implies an internal experience. A much more accurate terminology would have been “AI confabulation”. But alas, we are stuck with bad terminology, such as sampling from a fitted LLM being called “inference”. 🤦‍♂️

William Gibson@greatdismal.bsky.social

The idea of AI “hallucinating” is one I wish I’d had!

Tiago Peixoto reposted

People of Clacton! Show your support for me by putting a bin outside your house, one day this week, of a colour of Tendring Council's choice. This will be unequivocal proof of the success of my campaign.

Die brasilianischen Präsidentschaftswahlen finden nun im Oktober statt, und Bolsonaros Sohn kandidiert und liegt in den Umfragen knapp auf dem zweiten Platz...

Christoph Bautz@christophbautz.bsky.social

Was für ein Signal der Hoffnung: Die Entwaldung des Amazonas - und damit der Lunge der Erde - ist fast gestoppt. Und das in einem Moment, wo das komplexe Ökosystem kurz vor einem irreversiblen Kipppunkt steht. So sieht konsequente und effektive Umweltpolitik aus, von der wir lernen können.

Tiago Peixoto reposted

Tiago Peixoto reposted

People rave about AI doing everything from proving theorems to finding bugs in software, etc, but by far the biggest boon to my productivity has been the ability to add my signature to PDFs directly from Firefox and Thunderbird with a few mouse clicks. What a time to be alive.

It's hard to express how weird and cool it was to think about virtual reality and the internet *before* it existed, while the USSR was still standing (even in the future predicted in the book). Reality turned out to be so much lamer; even if our computers have way more than 3 MB of *HOT RAM*.

Richard McElreath 🐈‍⬛@rmcelreath.bsky.social

ok I looked up the different editions and I am pretty sure this is the cover of the used paperback I had. so like 1985. you young people have no idea how cool this book was

Tiago Peixoto reposted

My investor told me he doesn't really see the moat in my plan. I explained that's the point. The glacis protects the base of the wall and obscures the ditch entirely. By the time the assault party reaches the counterscarp with fascines, they will be committed and under crossfire from grapeshot

Tiago Peixoto reposted

Replying to @tiago.skewed.de

I’ve managed to run a simple test on a 2M nodes network, PPBlockState for a quick evaluation: 13 h => 15 minutes (16 cores). I’m speechless, congratulations for this milestone! I’m leaving for holidays with other models running, I’ll come back with lots of things to do! 🙏🙏🙏