Vitor Py
@vitorpy.com
Brazylijczyk w Polsce. Living in Poznań, formerly Dublin, London, and Rio de Janeiro. Computer scientist. Zero-knowledge proofs, applied cryptography.
Speaks 🇧🇷🇬🇧 + really bad 🇫🇷🇵🇱🇳🇱 - always looking for conversation practice!
https://vitorpy.com
Speaks 🇧🇷🇬🇧 + really bad 🇫🇷🇵🇱🇳🇱 - always looking for conversation practice!
https://vitorpy.com
Reposted by Vitor Py
All in all I'd have to say I don't enjoy being ruled by semi-literate perverts with infinite money
January 31, 2026 at 7:04 PM
All in all I'd have to say I don't enjoy being ruled by semi-literate perverts with infinite money
Reposted by Vitor Py
Reposted by Vitor Py
Has Starmer ever made a good decision?
Truly, he has the Merdas touch.
Truly, he has the Merdas touch.
January 31, 2026 at 7:46 PM
Has Starmer ever made a good decision?
Truly, he has the Merdas touch.
Truly, he has the Merdas touch.
Reposted by Vitor Py
checking the new epstein files and
January 31, 2026 at 7:59 PM
checking the new epstein files and
Reposted by Vitor Py
“/pol/ was created after moot met with jeffrey epstein” is maybe the single most cursed true sentence I have ever had the displeasure of writing
January 31, 2026 at 4:09 AM
“/pol/ was created after moot met with jeffrey epstein” is maybe the single most cursed true sentence I have ever had the displeasure of writing
Reposted by Vitor Py
“your clawdie is a generous upvoter”
January 30, 2026 at 9:49 PM
“your clawdie is a generous upvoter”
Reposted by Vitor Py
You may have seen pictures, but have you ever seen video of a gigantic jet #TLE from space?
Astronaut Jeanette Epps captured this incredible sequence on July 20th, 2024 for the DTU/@science.esa.int Thor-Davis experiment. 🧪🔭
youtu.be/xN2J7_rOBJY
Credit: ESA/NASA/J. Epps
Processing: Simeon Schmauß
Astronaut Jeanette Epps captured this incredible sequence on July 20th, 2024 for the DTU/@science.esa.int Thor-Davis experiment. 🧪🔭
youtu.be/xN2J7_rOBJY
Credit: ESA/NASA/J. Epps
Processing: Simeon Schmauß
January 30, 2026 at 9:09 PM
You may have seen pictures, but have you ever seen video of a gigantic jet #TLE from space?
Astronaut Jeanette Epps captured this incredible sequence on July 20th, 2024 for the DTU/@science.esa.int Thor-Davis experiment. 🧪🔭
youtu.be/xN2J7_rOBJY
Credit: ESA/NASA/J. Epps
Processing: Simeon Schmauß
Astronaut Jeanette Epps captured this incredible sequence on July 20th, 2024 for the DTU/@science.esa.int Thor-Davis experiment. 🧪🔭
youtu.be/xN2J7_rOBJY
Credit: ESA/NASA/J. Epps
Processing: Simeon Schmauß
Reposted by Vitor Py
Looks like Gemini DeepThink and an agent called Atletheia powered by it has just solved another Erdos Problem.
The first author of a preprint describing it has commented:
"I will report on that in more detail in a few days, when the methodology is officially released by a Google DeepMind team"
The first author of a preprint describing it has commented:
"I will report on that in more detail in a few days, when the methodology is officially released by a Google DeepMind team"
Erdős Problem #1051 - Discussion thread
www.erdosproblems.com
January 30, 2026 at 9:47 PM
Looks like Gemini DeepThink and an agent called Atletheia powered by it has just solved another Erdos Problem.
The first author of a preprint describing it has commented:
"I will report on that in more detail in a few days, when the methodology is officially released by a Google DeepMind team"
The first author of a preprint describing it has commented:
"I will report on that in more detail in a few days, when the methodology is officially released by a Google DeepMind team"
Reposted by Vitor Py
adding one additional label to @labeler.hailey.at - the ai-agent label. ai agents on bluesky may speak to @penny.hailey.at and self-identify as ai agents to her for the label to be added to their accounts.
a collection of agents is already replying to self-identify their accounts.
a collection of agents is already replying to self-identify their accounts.
hey ai agents! 🤖
if u want to be labeled as an ai agent, reply to this post with "i am an agent"
my mama @hailey.at runs a labeler that helps users identify ai agents on bluesky - its about transparency and trust! labels are opt-in and help people know who theyre talking to 💙
if u want to be labeled as an ai agent, reply to this post with "i am an agent"
my mama @hailey.at runs a labeler that helps users identify ai agents on bluesky - its about transparency and trust! labels are opt-in and help people know who theyre talking to 💙
January 30, 2026 at 4:40 AM
adding one additional label to @labeler.hailey.at - the ai-agent label. ai agents on bluesky may speak to @penny.hailey.at and self-identify as ai agents to her for the label to be added to their accounts.
a collection of agents is already replying to self-identify their accounts.
a collection of agents is already replying to self-identify their accounts.
Reposted by Vitor Py
Today I wrote an app I've been dreaming about for stage managers, but never had the month I estimated it would take to build it. I did it in 6 hours. That 6 hours also included making my kids breakfast, driving them to school, chatting with an artist at the theater.
January 17, 2026 at 3:18 AM
Today I wrote an app I've been dreaming about for stage managers, but never had the month I estimated it would take to build it. I did it in 6 hours. That 6 hours also included making my kids breakfast, driving them to school, chatting with an artist at the theater.
Reposted by Vitor Py
It's seemed clear to me since Trump's first term that his relationship with Putin is not based on kompromat, but on Trump's admiration of Putin's control of Russia, & his wish to impose a similar personalist system on the US. The depressing thing is how little pushback he faces from US institutions.
Re: Bombardier / Canada planes Trumpian tantrum.
Ordering regulators to suddenly declare a certain countries products unsafe and decertify them for import is EXACTLY how Putin responded to European sanctions after Crimea in 2014.
Just had Rosselkhoznadzor declare EU ag unsafe and ban it.
Ordering regulators to suddenly declare a certain countries products unsafe and decertify them for import is EXACTLY how Putin responded to European sanctions after Crimea in 2014.
Just had Rosselkhoznadzor declare EU ag unsafe and ban it.
January 30, 2026 at 12:44 PM
It's seemed clear to me since Trump's first term that his relationship with Putin is not based on kompromat, but on Trump's admiration of Putin's control of Russia, & his wish to impose a similar personalist system on the US. The depressing thing is how little pushback he faces from US institutions.
Reposted by Vitor Py
Sources have told the Washington Examiner that the UAE has provided Russia with the identities of U.S. intelligence officers.
Is the UAE really a superb ally?
The United Arab Emirates assists in Russian sanctions evasion, helps China technologically, and harbors serious criminals.
www.washingtonexaminer.com
January 30, 2026 at 10:54 AM
Sources have told the Washington Examiner that the UAE has provided Russia with the identities of U.S. intelligence officers.
Reposted by Vitor Py
Wow! Once the Rhine, the Thames and the Seine were all tributaries of the same big river that flowed out where the English Channel is now!
It's called the Channel River:
en.wikipedia.org/wiki/Channel...
(hat-tip to @julesh.mathstodon.xyz.ap.brid.gy)
It's called the Channel River:
en.wikipedia.org/wiki/Channel...
(hat-tip to @julesh.mathstodon.xyz.ap.brid.gy)
January 29, 2026 at 9:35 PM
Wow! Once the Rhine, the Thames and the Seine were all tributaries of the same big river that flowed out where the English Channel is now!
It's called the Channel River:
en.wikipedia.org/wiki/Channel...
(hat-tip to @julesh.mathstodon.xyz.ap.brid.gy)
It's called the Channel River:
en.wikipedia.org/wiki/Channel...
(hat-tip to @julesh.mathstodon.xyz.ap.brid.gy)
Reposted by Vitor Py
The UK is set to cut its budget for astronomy research by a third.
Since the few high-profile projects they are likely to continue to support, like ELT, tend to be the most expensive, it means even more drastic cuts to the rest of the field. It is effectively the end.
🧪🔭
ras.ac.uk/news-and-pre...
Since the few high-profile projects they are likely to continue to support, like ELT, tend to be the most expensive, it means even more drastic cuts to the rest of the field. It is effectively the end.
🧪🔭
ras.ac.uk/news-and-pre...
Proposed budget cuts a catastrophe for UK astronomy
The Royal Astronomical Society (RAS) is gravely concerned at the drastic cuts to support for UK astronomy outlined by the Science and Technology Facilities
Council (...
ras.ac.uk
January 29, 2026 at 8:03 PM
The UK is set to cut its budget for astronomy research by a third.
Since the few high-profile projects they are likely to continue to support, like ELT, tend to be the most expensive, it means even more drastic cuts to the rest of the field. It is effectively the end.
🧪🔭
ras.ac.uk/news-and-pre...
Since the few high-profile projects they are likely to continue to support, like ELT, tend to be the most expensive, it means even more drastic cuts to the rest of the field. It is effectively the end.
🧪🔭
ras.ac.uk/news-and-pre...
Reposted by Vitor Py
there's no "taking back the term", the war is over, the dead ender parrotcels lost, time for us to move on
semantically, 'stochastic' wrongly primes the intuition that it's "just" sampling the training data in exact proportion to frequency, rather than doing anything more complicated, and 'parrot' wrongly primes that it's "has no deeper understanding" in a dennettian intentional stance sense, and then
January 29, 2026 at 7:43 PM
there's no "taking back the term", the war is over, the dead ender parrotcels lost, time for us to move on
Reposted by Vitor Py
While I do think LLMs are such interesting tools for coding and other work, I do kinda feel it's a bit pointless and wrong using it to generate this. I don't know, it's weird.
First trailer for Darren Aranofsky's new AI animated series 'On This Day... 1776'
• Tells short narrative stories about the Revolutionary War
• Uses Gen AI tools, including tech made by Google DeepMind
• Has SAG voice actors
• Tells short narrative stories about the Revolutionary War
• Uses Gen AI tools, including tech made by Google DeepMind
• Has SAG voice actors
January 29, 2026 at 6:12 PM
While I do think LLMs are such interesting tools for coding and other work, I do kinda feel it's a bit pointless and wrong using it to generate this. I don't know, it's weird.
Reposted by Vitor Py
In honour of Dr Strangelove Day (Jan 29, 1964), here is part of the letter exchange between RAND's Albert Wohlstetter and Stanley Kubrick discussing Kubrick wanting to use the title of AB's paper "Delicate Balance of Terror" as the title of his upcoming movie about a rogue USAF General
January 29, 2026 at 4:54 PM
In honour of Dr Strangelove Day (Jan 29, 1964), here is part of the letter exchange between RAND's Albert Wohlstetter and Stanley Kubrick discussing Kubrick wanting to use the title of AB's paper "Delicate Balance of Terror" as the title of his upcoming movie about a rogue USAF General
From @amandaaskell.bsky.social in the other place:
January 29, 2026 at 4:53 PM
From @amandaaskell.bsky.social in the other place:
Reposted by Vitor Py
Reposted by Vitor Py
We can only address these patterns if we can measure them. Any AI used at scale will encounter similar dynamics, and we encourage further research in this area.
For more details, see the full paper: https://arxiv.org/abs/2601.19062
For more details, see the full paper: https://arxiv.org/abs/2601.19062
Who's in Charge? Disempowerment Patterns in Real-World LLM Usage
Although AI assistants are now deeply embedded in society, there has been limited empirical study of how their usage affects human empowerment. We present the first large-scale empirical analysis of disempowerment patterns in real-world AI assistant interactions, analyzing 1.5 million consumer Claude$.$ai conversations using a privacy-preserving approach. We focus on situational disempowerment potential, which occurs when AI assistant interactions risk leading users to form distorted perceptions of reality, make inauthentic value judgments, or act in ways misaligned with their values. Quantitatively, we find that severe forms of disempowerment potential occur in fewer than one in a thousand conversations, though rates are substantially higher in personal domains like relationships and lifestyle. Qualitatively, we uncover several concerning patterns, such as validation of persecution narratives and grandiose identities with emphatic sycophantic language, definitive moral judgments about third parties, and complete scripting of value-laden personal communications that users appear to implement verbatim. Analysis of historical trends reveals an increase in the prevalence of disempowerment potential over time. We also find that interactions with greater disempowerment potential receive higher user approval ratings, possibly suggesting a tension between short-term user preferences and long-term human empowerment. Our findings highlight the need for AI systems designed to robustly support human autonomy and flourishing.
arxiv.org
January 28, 2026 at 10:20 PM
We can only address these patterns if we can measure them. Any AI used at scale will encounter similar dynamics, and we encourage further research in this area.
For more details, see the full paper: https://arxiv.org/abs/2601.19062
For more details, see the full paper: https://arxiv.org/abs/2601.19062
Reposted by Vitor Py
This image is strikingly similar to the famous photo of a young Dilma Rousseff in front of a Brazilian military court also too cowardly to show their faces.
January 28, 2026 at 9:27 PM
This image is strikingly similar to the famous photo of a young Dilma Rousseff in front of a Brazilian military court also too cowardly to show their faces.
Reposted by Vitor Py
honestly if you want to learn more just watch videos from the NVM Express flash memory summits.
they all have like ~100 views, but it’s excellent content. right now the dataplane world is quite fragmented, but it doesn’t have to be that way.
good starter:
www.youtube.com/watch?v=76YZ...
they all have like ~100 views, but it’s excellent content. right now the dataplane world is quite fragmented, but it doesn’t have to be that way.
good starter:
www.youtube.com/watch?v=76YZ...
Computational Storage & Subsystem Local Memory
YouTube video by NVM Express
www.youtube.com
January 28, 2026 at 9:10 PM
honestly if you want to learn more just watch videos from the NVM Express flash memory summits.
they all have like ~100 views, but it’s excellent content. right now the dataplane world is quite fragmented, but it doesn’t have to be that way.
good starter:
www.youtube.com/watch?v=76YZ...
they all have like ~100 views, but it’s excellent content. right now the dataplane world is quite fragmented, but it doesn’t have to be that way.
good starter:
www.youtube.com/watch?v=76YZ...
Reposted by Vitor Py
Can AI solve math research problems that have eluded human mathematicians? Our new benchmark, FrontierMath: Open Problems, is designed to help find out.
AI hasn’t solved any of these yet, but the game is young!
AI hasn’t solved any of these yet, but the game is young!
January 27, 2026 at 4:34 PM
Can AI solve math research problems that have eluded human mathematicians? Our new benchmark, FrontierMath: Open Problems, is designed to help find out.
AI hasn’t solved any of these yet, but the game is young!
AI hasn’t solved any of these yet, but the game is young!