Antonio Patriarca

A mathematician working as a software engineer. Currently unemployed and looking for new job directions. Previously at Amazon, Skydance Animation, Animal Logic, DNEG, Framestore...

Instagram: https://www.instagram.com/antopatriarca/
Bluesky: ‪@antopatriarca.bsky.social‬

Public Key

npub1lpmg0xu0x87kvrk0uxjma5f2asu7he3knu9wkhpyalas04cl5ucqyyvg96

NIP-05 Address

antopatriarca_at_mathstodon.xyz@momostr.pink

Profile Code

nprofile1qqs0sa58nw8nrltxpm87rfd76y4wcw0tucmf7zhttsjwl7c86u06wvqprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntznwaqz

Publishing to

relay.momostr.pink

Author Public Key

npub1lpmg0xu0x87kvrk0uxjma5f2asu7he3knu9wkhpyalas04cl5ucqyyvg96

Show more details

Last Notes

2026-09-27 12:22:37 UTC

- reply

I think the choice to speak about scientific knowledge was more of a consequence of the period he lived (and some personal bias). It is in fact quite easy to find non-scientific knowledge used to control people and many scientific facts that cannot be used in that way (knowing the anatomy of cat, the existence of photons, the crystalline structure of quartz or the composition of the universe for example). Philosophy and religion has been used to assert control since the beginning of time! In a period when humanists knowledge was considered inferior to the scientific one, power just decided to use science (often by directly controlling it or filtering the message they wanted). And some humanists seeing a loss of prestige and money obviously reacted by trying to undermine science (I don’t know about Foucault specifically but some of the others for sure). Which in many ways actually worked as in the current political climate, science denial is often more common than the opposite. Things like AI are pumped using science-fiction more than actual science and uses the prestige of mathematicians more than their knowledge. Gender politics is shaped by the denial of gender existence, racists are trying to deny all the evidence against their loved eugenics, vaccines are said to cause all kinds of issues without evidence, climate change is denied so that they can continue to profit,…

2026-09-04 12:19:57 UTC

- reply

If we speak about local models then yes. I was mostly thinking about the edge models AI companies are asking a lot of money for. But it does indeed make more sense to use local models as you suggested.
Is there however any proof this is actually a problem? Is there any statistics/study showing that attacks from this kind of people have significantly increased and that they are causing new problems we are somehow unprepared to face?

2026-09-04 10:54:52 UTC

- reply

I’m not an expert (in fact I know very little), but the cost of the attack is also a limitation you have to consider IMHO. The random guy would still need to spend a few thousand dollars of tokens for each attempted attack. At that point there isn’t much difference between using the LLM and hiring someone for doing it, paying some existing service or run scripts you find online. All things that have been available for decades.

2026-08-31 11:22:35 UTC

- reply

It was a really well written article. I think he mostly succeeded in convincing about the second fact however. In fact, while I agree they are arbitrary and ornery, I don’t think they are necessarily the most arbitrary and ornery objects of all mathematics. Particularly considering how well they are actually understood and how much regularity they show.

2026-08-09 09:54:44 UTC

- reply

As an outsider, I’m really not understanding how Springer Nature still has that much power. From what I see, it is essentially providing no value nowadays. Editing is a hit-and-miss and I’ve heard horror stories of them going all AI in some fields, causing more problems than solutions. Thanks to LaTeX, formatting of most preprints is already “good enough”. The peer reviewing is voluntary and thus not something specific to these journals. Article access is actually more limited when published on these journals than on something like arXiv (and indeed I would usually just check the preprint). If arXiv would start providing a service to somewhat certificate a “preprint” as peer reviewed and publish it in a specific section of the website as “published”, they would probably give all the value you get from Springer. What I’m missing?

2026-07-29 14:45:49 UTC

- reply

Yes, it was meant as a collective “we”, in the sense of the mathematical (but also scientific) community. There has been a lot of people who tried (and are still trying) to go against this view, but their voice is rarely heard outside of the community in my experience. This is partly because of prejudices, but also because of the way we celebrate and tell our history and achievements.

2026-07-29 13:02:18 UTC

- reply

I agree, but at the same time it is academia, governments and sometimes mathematicians themselves who have popularised maths as a “competitive” endeavour. Awards, IQ companies and media has also helped. We are responsible for the way the world see at math. This view as math as a competition to solve the most important theorems unfortunately existed way before AI.

2026-07-06 16:16:39 UTC

- reply

The desire to obtain something (power, knowledge, skills, …) without having to spend too much effort in it is quite universal IMHO. The trickster part is just to teach there is always a cost to pay. It is just a version of any folklore tale telling you that some desires are evil and amoral and you will end up regretting following them.

2026-05-09 19:44:43 UTC

- reply

Oh, sorry. I misread it and thought he was saying the opposite. Which didn't really make any sense to me, and that was why I asked. 😅

2026-05-09 12:28:49 UTC

- reply

I’m not particularly convinced this is how the algorithm work. At least based on my (probably a little dated now) knowledge on how these models actually work. It seems more likely it is a sort of pattern matching algorithm, following methods seen somewhere else in the the training data.

2026-05-09 12:18:02 UTC

- reply

Maybe I am missing something, but I don’t understand how having something capable of solving these kind of problems helpful to PhD students trying to learn to do it. It seems to me that they are going to remove some difficulties and maybe save time for experts, but in doing so they are removing important learning steps for students.

2026-05-04 17:53:34 UTC

- reply

For me, this would make the machine fail the Touring test as previously described in this sequence of posts. No human, not even a poet, would ever make you that many poems in different languages and on different topics they probably don't care about. Being too compliant or all knowing is a quite good sign of being non-human.

2026-04-25 21:30:31 UTC

- reply

I agree. I heard many non-mathematicians speaking about Gödel's theorem as some kind of shocking negative thing, but all the mathematicians I met either considered it an opportunity or thought it basically changed nothing for them.

2026-04-15 08:25:50 UTC

- reply

Sure, sorry. I guess I wanted to bring home the message that while some things may have been indeed done by a few evil individual, I still think there is some responsibility in the people who have elected and enabled them. But you are right, I have been asked how I could vote for someone like Berlusconi many times in my life (obviously when he was still alive) and it was annoying as I was too young for the first few elections and I never actually voted for him later. Ironically now, Americans were the most probable to ask such question as soon as they discovered I was Italian. So maybe there is also some bit of vendetta… I think that embarrassing and stupid things are the easiest to just blame on the individual. We will just say that he was a pathological liar and got dementia at the end. The fans will instead say he was passionate and direct in his speech ignoring the lies and the dementia. Berlusconi fans would have said he had sense of humour, that people were jealous and used the law to remove him from politics and that he was misunderstood and generous (this last bit connected to the women he had done favours). You will then find people saying the electors believed in his lies and they will probably cast the blame in tech bros using the social media to control the elections. So in the end no blame except for Trump and his minions/billionaires friends.

2026-04-14 22:44:40 UTC

- reply

It also somewhat helps that it takes some time for something to move from recent news to history. Particularly in history books for children. What is the most recent event taught in kid books in the US right now? Is 9/11 taught for example? The most recent event I was taught in school was probably about 40 years old already. People change in 40 years and there is at that point at least a generation “buffer” in the middle. It is your grandparents who were involved and in a period that was very different from the present.
To be completely honest, I think there are a lot of embarrassing things the USA has done before Trump and decided to somehow either ignore or pinpoint to specific individuals. For example, American concentration camps, the McCarthyism, Nixon, all the CIA operations around the world to control foreign elections, you started several wars without good reasons and you even launched bombs on countries you were not at war. Look at how your history books speak of these things and you will get d as n idea of how Trump will be told.

2026-04-14 20:41:47 UTC

- reply

History books for children do not contains such details. They will probably just mention the president used social media and memes for official communication often causing embarrassment in allies and citizens. Then if they want they will present an example. As someone who was born in a country with an embarrassing past, the way it was handled was by simply separate the present from the past, as if it was done by different people. Obviously that is only going to happen after the fall of the embarrassing regime. And people are all just going to behave as if they were always at the right side

2026-04-14 17:51:21 UTC

- reply

why would/should they say anything different from describing how embarrassing it all is?

2026-03-28 17:25:10 UTC

- reply

Ah, ok, you are talking about the stuff at the bottom of the page, which I still haven't looked into.

2026-03-28 17:07:54 UTC

- reply

I am quite new to these, so maybe I am misunderstanding things. But I thought parity is related to the quantum state and not specific to a particle. So, for example, an s orbital would have parity 1 and a p orbital would have parity -1. An electron can clearly live in both. I also think there are experiments that can measure parity.
What is specific to particles is the symmetry or asymmetry under the exchange of two or more of them (with the proof essentially being their different multi-particle statistics).

2026-03-07 17:49:54 UTC

- reply

yes, I agree. In the painting the mirror seems to be showing a different scene than the non-mirrored part. It is only from the explanation of what we are seeing that the interpretation of the two customers start to appear.

2026-03-07 17:36:31 UTC

- reply

I would in fact say she doesn’t seem to be serving the man. A more plausible version is that she served the man already and she is looking at the next customer, the viewer/painter/photographer.

2026-03-03 14:53:53 UTC

- reply

Out of curiosity, why Italian?

2026-02-24 06:56:42 UTC

- reply

Obviously, the greatest questions are the ones which doesn't have an answer, yet? And the quickest questions are the ones from people who understood absolutely nothing and thus stay silent.

2026-01-08 09:41:16 UTC

- reply

As a software engineer, the main message I receive in this world is to make things fast, sacrificing on quality, ethics and everything we have learned. So your message somehow triggered my hatred for all the vibe-coding messages I see around me. I understand the art world is different, and I do actually enjoy making my own private, very imperfect art. This is why I had mixed feelings; it triggered something deep in me, even if I understood and agreed with your message with my head. I suppose I never realised I am starting to have almost PTSD from all of this.

2026-01-07 17:36:54 UTC

- reply

I have mixed reactions to the message. I agree that doing something by ourselves is an act of resistance, but I feel that doing things imperfectly is completely on brand with the current AI slop. Quality, accuracy, precision, ... do not seem to have any role in the AI world the tech bros are trying to create.

2025-11-17 16:25:14 UTC

- reply

Even if he knows nothing, it may still matter as the investments are based on hype and people may assume he knows something and do the same out of fear, causing the eventual collapse.

2025-06-01 21:21:58 UTC

- reply

Just to make myself clear, what these people do is not math or physics for the simple reason that they are not following the rules of physics or mathematics. Even worse, they consider themselves superior to the game itself. It's like someone saying they are great chess players but always play by their own rules. The reason I say it is a shame is only because they have a big imagination and, if it wasn't for their psychosis, they could have been great science fiction writers or maybe even scientists.

2025-06-01 15:37:48 UTC

- reply

It is a shame however these people are convinced of crazy stuff because they do have a lot of imagination and dedication.

2025-06-01 15:26:41 UTC

- reply

Wow! Was this some clever plan to give crackpots a place to put their stuff so that no human needs to look at them?

2025-05-05 16:58:11 UTC

- reply

probably, as in some famous literary example, only the answer was provided, which we already knew to be 42, but we still need to find the problem this is the solution to.

2025-02-07 20:44:08 UTC

- reply

As someone exposed to the Noether normalization lemma more than this theorem, I initially asked myself how that lemma would help physicists. Then I remembered there were other important theorems with her name. She was truly an amazing mathematician and we need to celebrate her more.