Hoy un compañero de trabajo le ha puesto una transparencia en LaTeX a ChatGPT para ver si arreglaba una historia de numeración. El código sugerido eran el resto de texto de la presentación, que sólo había subido anteriormente a Overleaf de forma privada. Esto se suma a las evidencias de que ChatGPT se alimenta de repositorios privados de github.
Ahora mismo cualquier cosa en la nube de cualquier tipo es susceptible de haberse vendido a una IA generativa.
Para que lo tengáis en cuenta.
13:21:35 output message: sent to inbox https://toot.mirbsd.org/users/mirabilos/inbox (599 Unknown)
My new #fediverse home is @giacomo
Please, follow me there!
Oooooh, this by @Shamar is also good:
There is no "learning" in "artificial neural network"
[…]
Such kind of virtual machines are composed of tiny devices that are improperly named "artificial neurons" or "perceptrons", but they are simply vector reducers […]
Such vector reducers can be easily composed in a variety of topologies and programmed in a variety of ways, so that a whole "artificial neural network" can be statistically programmed ("trained", in AI/ML parlance) to approximate one of the possible translations from a vector space to a different one.
Yet, there is no "learning" ongoing: just the iterative tuning of parametric vectors to approximate a certain output. […]
[…] In other words, an "artificial neural network" does not learn anything and it's not a network of neurons or anything like that. It does not understand anything about the data and the output vectors has no inherent meaning: its semantic is always attributed by humans according to their insights about the statistical program they uploaded into the vector mapping machine.
[…]
That's why it's statistical programming: you start with a source dataset (improperly named "training set" in the AI/ML parlance) and, after a compilation process that is specifically designed for that specific virtual machine, you get a binary that such machine can run.
While such binary is not constituted by a sequence of instructions, it's still just an algorithmic transformation of the original source that, despite being expressed as cryptic matrices, still contains all the relevant features of the sources.
The software run by a vector mapping machine is not a "model", since it does not give any insight about the relations it absorbed during its statistical programming. Instead it is just a software like any other that describes a rigorous (if unknown) process that a specific machine has to follow to automatically and deterministically compute (or rather approximate) a desired output from a given input.
The quotes are from the essay published by the author under this licence (which is nōn-FOSS, but it’s still worth a read).
Copilot exposes private GitHub pages, some removed by Microsoft
Microsoft’s Copilot AI assistant is exposing the contents of more than 20,000 private GitHub repositories from companies including Google, Intel, Huawei, PayPal, IBM, Tencent and, ironically, Microsoft.
#microsoft #copilot #ai #privacy #surveillance #enshittification #tech #github #softwaredevelopment
"Mientras que la matemática pertenece al dominio de la mente humana y es comunicable, cualquier concepto que pertenece a la matemática es información. Como tal, también pertenece al campo de la informática.
Lo contrario no es cierto: los errores están muy presentes en la informática pero no se dejan ver en el mundo de las matemáticas.
Como consecuencia, las matemáticas son parte de la informática, y no al contrario.
Aunque esta afirmación puede parecer una herejía al principio, no debería sorprendernos demasiado. Podemos observar que la informática altera cada aspecto de la vida humana. Este fenómeno tiene una explicación sencilla: transforma la forma en la que la humanidad piensa colectivamente. La información pertenece al dominio de la mente humana y, al compartirse con otros miembros de la comunidad, construye la cultura de dicha comunidad. Después, la cultura vuelve a la mente de los nuevos miembros de la comunidad a modo de información en un bucle infinito."
— Giacomo Tesio, Ekaitz Zárraga: ¿Qué es la informática?, p. 21
The red means it was removed in case you're not sure about github, the term sell has been entirely removed from the FAQ page https://www.mozilla.org/en-US/firefox/faq/
Satire is dead folks.
"Many organizations – big and small – still lack access to basic safety resources, hindering innovation and putting users at risk." And enter Google and OpenAI to the rescue. Obviously.
Google and OpenAI are going to help YOU not put users at risk folks.
Google, a brief history
2004: Don’t be evil
2018: OK, be evil but at least don’t kill people
2025: Fuck it, kill people
https://www.independent.co.uk/business/google-drops-pledge-not-to-use-ai-for-weapons-b2692533.html
#Google #BigTech #AI #SiliconValley #USA #capitalism #fascism
The self-destruction of the OSI has no end.
“Get ready to vote for OSI’s board of directors in 2025. […] Please have a look at the Data Governance in Open Source AI […]”
Seems like only candidates who are in favour of their “OSAID” (badly named, as it violates Open Source standards) are even considerable anyway, and voters are directed with leading messages months in advance.
Oh well, they are in the USA after all.
Demistificare e porre le domande giuste sono i due obiettivi di questa critica sistematica e multidisciplinare a ChatGPT, il più noto tra gli strumenti di intelligenza artificiale prepotentemente entrati nelle nostre vite. E se da un lato è necessario decostruire le narrazioni tecno-entusiaste che prospettano un avvenire radioso, dall’altro appare cruciale stimolare una riflessione più consapevole sul tipo di mondo che stiamo costruendo quando accogliamo – senza farci troppe domande – la rivoluzione tecnologica in atto.
Il 31 gennaio esce in libreria la nostra seconda novità dell'anno "Critica di ChatGPT" di Antonio Santangelo, Alberto Sissa, Maurizio Borghi.
scheda del libro:
https://www.eleuthera.it/scheda_libro.php?idlib=608
leggi la prefazione:
https://www.eleuthera.it/files/materiali/Santangelo_sissa_borhi_chatGPT_indice_prefazione.pdf
007: la spia che insegnava sarebbe soltanto un film.
Pare che la preoccupazione di @aisa sull'articolo 31 del #ddlsicurezza sia un "allarmismo ingiustificato".
Ma se fosse vero che l'articolo 31 del #ddlsicurezza non rende obbligatoria, per le università e gli enti di ricerca, la collaborazione con i servizi, perché introdurre una norma che non cambia nulla rispetto a quella vigente e ostinarsi a difenderla, magari evocando usanze non solo "statunitensi" ma "asiatiche"?
Anche Tomaso Montanari, che pure interpreta la norma in discussione in senso restrittivo, almeno per le università, suona bizzarramente allarmista, al punto di scomodare Orwell
Oggi ho discusso con gli studenti del problema della liceità del regicidio nella Metafisica dei costumi di Kant - tema di scottante attualità nell'ultimo decennio del XVIII secolo. Se gli studenti sapessero che io posso (o devo?) spiarli, che il mio rettore può (o deve?) farlo, ricerca e didattica diventerebbero più libere e sicure, o meno?
La Commissione UE è accusata di aver trasferito a Meta, negli Stati Uniti, senza le adeguate garanzie previste dal GDPR i dati personali di un cittadino che ha effettuato l'autenticazione con Facebook
A book about writing a toy operating system for RISC-V in 1000 lines of C.
ho costruito una cucina ikea