Treballo en coses de continguts web, seo i analítica web. Aquest és un twitter dispers: fer webs, gastronomia, tonteria... Política de no favs

Barcelona
Joined April 2007
Roger Compte retweeted
You have noticed it. ChatGPT feels dumber than it used to. Your prompts that worked six months ago produce worse results now. The writing sounds flatter. The ideas sound safer. The internet itself feels like it is shrinking. Every article reads the same. Every email sounds the same. Every answer sounds like it was written by the same voice. You thought it was you. It is not you. Researchers at Oxford and Cambridge published a paper in Nature proving what is happening. They call it Model Collapse. Here is the mechanism in one sentence. AI trained on AI-generated data gets dumber every generation until it forgets what real human data looked like. The internet is filling with AI-generated content. Blog posts. Articles. Reviews. Comments. Social media. AI companies scrape the internet to train the next generation of models. Which means the next generation of AI is being trained on the output of the current generation. Each cycle loses information. Not randomly. It loses the rarest, most unusual, most creative parts first. The researchers call these the "tails of the distribution." The weird ideas. The unexpected perspectives. The things that made the internet feel human. Those disappear first. What remains is the average. The safe. The expected. The bland. Then the next generation trains on that. And loses more. And the next generation trains on that. And loses more. The researchers proved this is not a slow decline. Major degradation happens within just a few iterations. Even when some of the original human data is preserved. They tested it on large language models. On image generators. On statistical models. The pattern was the same every time. The output converges toward a narrow, flattened version of reality that looks nothing like the original data. The lead researcher put it plainly. "Large language models are like fire. A useful tool. But one that pollutes the environment." The pollution is invisible. You cannot see which sentence on the internet was written by a human and which was written by AI. Neither can the AI that is about to train on it. And once the tails are gone, they do not come back. The damage is irreversible. This is not a prediction anymore. It is a diagnosis. The internet you grew up on was built by humans writing things no algorithm would have written. Strange, personal, imperfect, alive. That internet is being diluted. One generation of AI at a time. And the models trained on what remains are learning a smaller and smaller version of the world. Model Collapse is not a technical problem. It is a cultural one. The thing that made the internet worth reading is the thing that disappears first.
1,117
6,240
1,370
17,475
2,311,099
Roger Compte retweeted
Com a periodista, com a espectadora i com a mare de dues nenes que han crescut amb les seves obres i els seus tallers d’estiu, aquesta notícia que mai hauria volgut donar, em crea una tristor absoluta. Avui som més pobres i els nostres nens i nenes, també.
16
168
9
381
40,530
Van pujant les respostes de temes de feina amb aroma a chatgpt o equivalent. Ajuda't de la IA, però digereix, valida i selecciona el què et proporciona i currat una mica la resposta, home
1
133
Dilluns al vespre un grup de turistes borratxos van entrar, des de de dins de les bàteries (ja tancades), a la vivenda d'uns veïns. En ser descoberts, van intentar agredir a la parella, que va haver de refugiar-se dins de casa per evitar un linxament.
16
353
26
571
69,547
Roger Compte retweeted
Li ha robat la tecnologia al Cincsegons.
Rosalia i "La Perla". Kiedy niepotrzebna jest AI, ani CGI, ani efekty z "Jurrasic Park"...
12
71
2
866
36,887
Roger Compte retweeted
Ja ho diu la dita catalana, "dispara bé el trabuc qui ha menjat festuc".
3
16
183
12,292
El curioso pez murciélago de labios rojos (Ogcocephalus darwini) habita en el Pacífico, entre las islas Galápagos y las costas de Perú
2
63
2
243
5,161
Roger Compte retweeted
La Clàudia és una nena de 7 anys sociable, lluitadora, molt dolça i, sobretot, feliç. Té una malaltia rara derivada de l'alteració d'un gen que es diu MYT1L. És l'únic cas a Catalunya. A Espanya, només n'hi ha quatre més. La família demana que s'investigui 3cat.cat/3catinfo/la-malalti…
2
25
1
91
25,371
Roger Compte retweeted
🗣️ Ariadna, mare de l'única nena a Catalunya amb la malaltia rara Myt1l: "La genetista va venir amb un full ple de lletres i números, no enteníem un borrall. Serveix per posar-li un nom, però res més" rac1.cat/societat/20260226/3…
4
11
5,665
Necessari documental, nosaltres vivim una cosa semblant, amb unes altres sigles, MYT1L
⏳I d'aquí uns minuts, "Moure muntanyes" 🫂S'han unit per fer possible l'impossible: impulsar una teràpia avançada experimental per curar els seus fills #MoureMuntanyes3Cat #SenseFicció3Cat 🔗3cat.cat/3cat/sense-ficcio/
4
243
Roger Compte retweeted
⏳I d'aquí uns minuts, "Moure muntanyes" 🫂S'han unit per fer possible l'impossible: impulsar una teràpia avançada experimental per curar els seus fills #MoureMuntanyes3Cat #SenseFicció3Cat 🔗3cat.cat/3cat/sense-ficcio/
6
2
15
2,702
RIP Privacy — AI Glasses Can Now Recognize Anyone, Anywhere. A Dutch journalist just tested a pair of AI-powered glasses that can instantly identify strangers on the street. No government database. No police system. Just public data and off-the-shelf AI. You look at someone and in seconds, their name, LinkedIn, and background appear before your eyes. The scariest part? You can’t really stop it. You can ban it, regulate it, add blinking red lights… but once tech like this exists, someone will always find a way to use it. To me, this marks a turning point. We’ve officially blurred the line between seeing people and knowing them. Between being in public and being exposed. So here’s the question: When every face becomes a dataset, how do we protect the meaning of being human? #AI #Privacy #Ethics #Technology #Innovation #Data #Surveillance
947
6,144
1,053
14,591
1,528,801
La investigación de la CNN deja en evidencia la atrocidad que cometió Israel en Zikim. Atrajo a miles de palestinos a los hipotéticos centros de ayuda humanitaria para asesinarlos a tiros y después enterrarlos en fosas comunes con excavadoras. #Gaza cnn.com/2025/12/03/middleeas…
66
4,458
129
4,778
103,702
A més dels vídeos per a catalanitzar Android, Windows, iOS, macOS i el compte de Google, a configura.cat/aplicacions també teniu instruccions per posar en català: - Chrome - Firefox - Safari - WhatsApp - Telegram - X - TikTok - Spotify - Facebook Si us plau, feu-ho còrrer!
12
664
9
625
44,874
Mentre no arriba el nou de Crim, tenim ben coberta la quota de punk catalufu amb el disc dels Enemic Interior open.spotify.com/album/7mD69…
184