Bruno Cuevas Zuviría
@brunocuevaszuviria.bsky.social
Computational biology researcher. Now working at Universidad Politécnica de Madrid. Interested in Origins of Life, early protein evolution, complexity, network analysis, synthetic biology and pathogen evolution. Bike warrior, urbanism nerd and neo-luddite.
Why sometimes Elsevieer login feels like a black hole where you cannot literally go back to the previous page? I just had to close a tab because of it
November 19, 2025 at 10:00 PM
Why sometimes Elsevieer login feels like a black hole where you cannot literally go back to the previous page? I just had to close a tab because of it
Is it me or the internet looks specially dangerous/troubled with cyberattacks, scams, and outages these last days?
November 18, 2025 at 9:17 PM
Is it me or the internet looks specially dangerous/troubled with cyberattacks, scams, and outages these last days?
Reposted by Bruno Cuevas Zuviría
New Clone-FISH paper lead by former lab member Joelie Van Beek. Way to go, Joelie! 🎉🎉🎉 @environmicrobio.bsky.social
journals.asm.org/doi/epdf/10....
journals.asm.org/doi/epdf/10....
November 17, 2025 at 10:50 PM
New Clone-FISH paper lead by former lab member Joelie Van Beek. Way to go, Joelie! 🎉🎉🎉 @environmicrobio.bsky.social
journals.asm.org/doi/epdf/10....
journals.asm.org/doi/epdf/10....
The answer from Anthropic about why they keep releasing models if they are used for harm (cyberattacks), is that they can also help us to defend from the tools that they have unleashed.
So, the path to profitability is not to solve problems, but to generate a problem and then provide the solution.
So, the path to profitability is not to solve problems, but to generate a problem and then provide the solution.
November 14, 2025 at 12:45 PM
The answer from Anthropic about why they keep releasing models if they are used for harm (cyberattacks), is that they can also help us to defend from the tools that they have unleashed.
So, the path to profitability is not to solve problems, but to generate a problem and then provide the solution.
So, the path to profitability is not to solve problems, but to generate a problem and then provide the solution.
❤️https://github.com/bnjbvr/rouille
I am waiting to see a Spanish version.
I am waiting to see a Spanish version.
October 31, 2025 at 9:58 AM
❤️https://github.com/bnjbvr/rouille
I am waiting to see a Spanish version.
I am waiting to see a Spanish version.
Reposted by Bruno Cuevas Zuviría
Join our next ProSe seminar 11th November 5PM CET and follow us @proteinstructure.bsky.social!
The next ProSE Seminar will be given by Joana Pereira @joanampereira.bsky.social (VIB, KU Leuven)
🗓️ November 11th, 5pm CET
Registration is now open: tinyurl.com/prose-seminar2
Please share!
🗓️ November 11th, 5pm CET
Registration is now open: tinyurl.com/prose-seminar2
Please share!
October 28, 2025 at 8:13 PM
Join our next ProSe seminar 11th November 5PM CET and follow us @proteinstructure.bsky.social!
I just realized that this little summer experiment that I created last year was not working anymore. Come on, little buddy! You are live again!
ruleit.streamlit.app
ruleit.streamlit.app
Rule it:Wellcome
Welcome to Rule-It! This platform was created by Bruno Cuevas and Tyma Sokolskyi to enable resear...
ruleit.streamlit.app
October 29, 2025 at 9:13 PM
I just realized that this little summer experiment that I created last year was not working anymore. Come on, little buddy! You are live again!
ruleit.streamlit.app
ruleit.streamlit.app
No wonder that people become hooked at chatbots, they are ego-inflating machines. I just asked about ideas for a synonym in my next paper, and I got this
(BTW, all the ideas were meh)
(BTW, all the ideas were meh)
October 28, 2025 at 8:30 PM
No wonder that people become hooked at chatbots, they are ego-inflating machines. I just asked about ideas for a synonym in my next paper, and I got this
(BTW, all the ideas were meh)
(BTW, all the ideas were meh)
Yo juraría que acepté recibir contenido promocional sobre colchones, no sobre emprendimiento.
October 27, 2025 at 3:34 PM
Yo juraría que acepté recibir contenido promocional sobre colchones, no sobre emprendimiento.
I find funny how some folks get reactive about one giving credit to other people. Who knows who said what for the first time
October 24, 2025 at 2:19 PM
I find funny how some folks get reactive about one giving credit to other people. Who knows who said what for the first time
Since I heard the idea that "surveillance used to be a side-effect; now it is the product" in @techwontsave.us with @hypervisible.blacksky.app, I cannot help finding it everywhere I read. For instance, this is the new "Claude Memory" by Anthropic.
October 24, 2025 at 8:23 AM
Since I heard the idea that "surveillance used to be a side-effect; now it is the product" in @techwontsave.us with @hypervisible.blacksky.app, I cannot help finding it everywhere I read. For instance, this is the new "Claude Memory" by Anthropic.
Reposted by Bruno Cuevas Zuviría
This is like the smoking gun of the AI bubble. I’m willing to say it is physically impossible for the LLM architecture to be turned into a product that is useful and be sold at a profit by the operator. I’m also sure Amazon is not turning a profit on LLM compute.
read this post. One of the accounting fuck fuck games in AI is fiddling the depreciation. The Nvidia cards die in 12-18mo. Amazon, Meta, Google all depreciating at 4-6 *years*. This makes *huge* difference in earnings per share.
blog.dshr.org/2025/10/depr...
(the bitcoin miners also pulled this)
blog.dshr.org/2025/10/depr...
(the bitcoin miners also pulled this)
Depreciation
Source More than three years ago, based on Paul Butler's The problem with bitcoin miners , I wrote Generally Accepted Accounting Principles...
blog.dshr.org
October 22, 2025 at 1:23 PM
This is like the smoking gun of the AI bubble. I’m willing to say it is physically impossible for the LLM architecture to be turned into a product that is useful and be sold at a profit by the operator. I’m also sure Amazon is not turning a profit on LLM compute.
Reposted by Bruno Cuevas Zuviría
Half the internet seems to be down thanks to the us-aws-east-1 outage
Perplexity, Amazon .com, Signal all down (just 3 services I tried to use)
When aws-us-east-1 sneezes, the whole world feels it indeed
Perplexity, Amazon .com, Signal all down (just 3 services I tried to use)
When aws-us-east-1 sneezes, the whole world feels it indeed
October 20, 2025 at 8:34 AM
Half the internet seems to be down thanks to the us-aws-east-1 outage
Perplexity, Amazon .com, Signal all down (just 3 services I tried to use)
When aws-us-east-1 sneezes, the whole world feels it indeed
Perplexity, Amazon .com, Signal all down (just 3 services I tried to use)
When aws-us-east-1 sneezes, the whole world feels it indeed
Reposted by Bruno Cuevas Zuviría
Are you having trouble making money from your AI super intelligence that will cure all diseases and replace all workers? Pivot to porn
Sam Altman says ChatGPT will soon sext with verified adults
Sam Altman says ChatGPT will soon sext with verified adults
ChatGPT is about to get flirty.
buff.ly
October 14, 2025 at 8:44 PM
Are you having trouble making money from your AI super intelligence that will cure all diseases and replace all workers? Pivot to porn
One week ago, I had to look up the meaning of spandrel. Now, I see this word in every place, even in articles that I had already read.
This could be a proof that we are living in a simulation.
This could be a proof that we are living in a simulation.
October 9, 2025 at 11:01 AM
One week ago, I had to look up the meaning of spandrel. Now, I see this word in every place, even in articles that I had already read.
This could be a proof that we are living in a simulation.
This could be a proof that we are living in a simulation.
Reposted by Bruno Cuevas Zuviría
How did animals repeatedly conquer land? 🌊➡️⛰️ We analysed ~1,000 gene repertoires (24M genes!) from all animal phyla to uncover how this happened. Work led by @gemmaeling.bsky.social & Klara Eleftheriadi, both first coauthors of this titanic effort!
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
Independent genomic trajectories shape adaptation to life on land across animal lineages
How animals repeatedly adapted to life on land is a central question in evolutionary biology. While terrestrialisation occurred independently across animal phyla, it remains unclear whether shared gen...
www.biorxiv.org
October 7, 2025 at 7:27 AM
How did animals repeatedly conquer land? 🌊➡️⛰️ We analysed ~1,000 gene repertoires (24M genes!) from all animal phyla to uncover how this happened. Work led by @gemmaeling.bsky.social & Klara Eleftheriadi, both first coauthors of this titanic effort!
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
Me he metido en el ramal en vez de en la línea 5. Merezco perder la madrleñidad
October 8, 2025 at 4:30 PM
Me he metido en el ramal en vez de en la línea 5. Merezco perder la madrleñidad
Reposted by Bruno Cuevas Zuviría
UPDATE: The 2025-2026 list of faculty and postdoc positions in ecology and evolutionary biology is out! Be sure to check out this active and helpful community run resources! docs.google.com/spreadsheets...
ecoevojobs.net 2025-26
docs.google.com
September 19, 2025 at 9:47 PM
UPDATE: The 2025-2026 list of faculty and postdoc positions in ecology and evolutionary biology is out! Be sure to check out this active and helpful community run resources! docs.google.com/spreadsheets...
It is curious to see amazing papers that are going out now 1) ride the AI-trend, 2) use very simple ML techniques that existed 50 years ago, and 3) get amazing results.
My take: many problems are still interesting because there is not enough data, preventing the trendy data-hungry techniques.
My take: many problems are still interesting because there is not enough data, preventing the trendy data-hungry techniques.
October 7, 2025 at 8:27 AM
It is curious to see amazing papers that are going out now 1) ride the AI-trend, 2) use very simple ML techniques that existed 50 years ago, and 3) get amazing results.
My take: many problems are still interesting because there is not enough data, preventing the trendy data-hungry techniques.
My take: many problems are still interesting because there is not enough data, preventing the trendy data-hungry techniques.
in case you are wondering why I still use Word, the answer is that not everybody is aware of how infuriating is MS Word.
October 1, 2025 at 6:01 PM
in case you are wondering why I still use Word, the answer is that not everybody is aware of how infuriating is MS Word.
2025. Lots of AI and whatever, but Word keeps getting stuck on a 25 pages/8000 words manuscript.
October 1, 2025 at 6:00 PM
2025. Lots of AI and whatever, but Word keeps getting stuck on a 25 pages/8000 words manuscript.
Me meo con las recomendaciones de Linkedin. Otra razón para esquivar esta red social.
October 1, 2025 at 7:45 AM
Me meo con las recomendaciones de Linkedin. Otra razón para esquivar esta red social.
Reposted by Bruno Cuevas Zuviría
E-motos have no place in our active infrastructure. Their operators are causing problems and bringing unjustified anger and regulation into the regular ebike movement and industry.
Ugh absolute scourge… just east of the Harry Hays @tinybeaches.bsky.social #yycbike
September 26, 2025 at 1:03 PM
E-motos have no place in our active infrastructure. Their operators are causing problems and bringing unjustified anger and regulation into the regular ebike movement and industry.
What is the point of replacing auto-completions that are based on the variable / class definition (e.g. class methods) by hallucinations? Doesn't it defeat the purpose of auto-completion? The copilot just made up a "pandas.DataFrame.to_yaml" that I don't see anywhere in the documentation.
September 26, 2025 at 9:27 AM
What is the point of replacing auto-completions that are based on the variable / class definition (e.g. class methods) by hallucinations? Doesn't it defeat the purpose of auto-completion? The copilot just made up a "pandas.DataFrame.to_yaml" that I don't see anywhere in the documentation.
To me, this liquid glass design is a failure. It might look pretty to someone, but it makes everything extremely cluttered, and I am not 100% confident about this, but I can imagine that it has an impact on performance—now you need to overlay different layers, compute blurring effects, and so on.
September 26, 2025 at 8:28 AM
To me, this liquid glass design is a failure. It might look pretty to someone, but it makes everything extremely cluttered, and I am not 100% confident about this, but I can imagine that it has an impact on performance—now you need to overlay different layers, compute blurring effects, and so on.