Look up here I'm in heaven I've got scars that can't be seen I've got drama can't be stolen Everybody knows me now βοΈBOWIEβοΈ 56/1.88m/80kg π§βπ»πͺππΆπΆββοΈ βοΈ
A neuroscientist trained at NIH and UCLA is now studying out-of-body experiences, mediumship and remote viewing at the University of Virginia.
She left a prestigious mainstream research career to do it. Her colleagues think she's taking a serious risk.
She thinks the data left her no choice.
Meet Dr. Marina Weiler. π§΅ [1/7]
@IRVAtweets Thank you for spotlighting Dr. Weiler! We had the pleasure of interviewing her about all this last month, and more people certainly need to be aware of her amazing work!
https://t.co/AbIca4sVfl
Vertel je nervus vagus maar dat je in het meest onbekommerde deel van de wereld leeft, in de meest comfortabele periode uit de geschiedenis van de mensheid.
https://t.co/xznbiIGxaN
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.
In koele taal beschreef Lieke Marsman wat we iedere dag al in de krant kunnen lezen. De feiten: βNederland heeft de afgelopen jaren voor 2 miljard aan IsraΓ«lische wapens aangekocht.β En daarna onomwonden: βIsraΓ«l pleegt genocide.β
https://t.co/GWCR2JlzW4
Rico Verhoeven heeft minimaal vijf peperdure bolides; wat moet je daarmee? Ik kan me daar niets bij voorstellen, je kunt er maar in eentje rijden. Van al die autoβs krijg je keuzestress en word je ongelukkig
https://t.co/MlaNgHRxOX
In Groningen zijn ze nooit op het idee gekomen om zich af te scheiden van Nederland; anders was de provincie nu het Koeweit van Europa geweest. In totaal heeft de Groningse gasbel Nederland 430 miljard euro opgeleverd!
https://t.co/Fyu8SfnSBg
The main culprit is the worldβs addiction to burning coal, oil and gas, and destroying forests.
The science is clear that human-induced climate change is making these heatwaves more frequent and extreme.
In 2024, summer heat in the EU claimed roughly three times more lives than car crashes, 16 times more than murderers, and more than 10,000 times more than terrorists.
https://t.co/Xrbjv1Q6Ls
Als je niet meer door de ogen van een man naar jezelf hoeft te kijken, je je niet meer tot hem hoeft te verhouden, jullie niet langer een huis delen, dan pas gaat de vrijheid je opvallen. De rΓΊΓmte.
https://t.co/OA5tW2eCcx
Maar in de praktijk stellen we het uit. Je wacht, zoals tachtig procent van alle euthanasiegevallen, net zolang tot die fatale injectie er nog amper toe doet.
Euthanasie is een fictie, stelde Karin. βWe maken onszelf wijs dat we een uitweg hebben als de ziekte, die ons nu al sloopt, echt te gortig wordt. Dan neem je de regie terug. Dat maakt het draaglijker, denk je. https://t.co/jz65wwC43z
Dat alles wat je waarneemt uiteindelijk maar een (gekozen) verhaal is, jouw verhaal, is even geruststellend als niet, ofzo. De keuzevrijheid vind ik geruststellend. Dat niets perse waar is niet.
Peter Middendorp formuleert het mooier, vind ik.
https://t.co/iuHHUDNgIX
Geen betaalmuur:
Uiteindelijk moet je een keer sterven. Als je dan de kans krijgt om dat in je eigen omgeving te doen, waarom zou je die kans dan verjagen? Als we de dood blijven wegduwen, βnu nietβ, kan de volgende kans weleens veel vervelender zijn.https://t.co/o4YURTsqxl