Adrià retweeted
jensen, nvidia, huggingface, china, taiwan, and our new weirdo overlords
The End of an Era: Jensen Huang
Adrià retweeted
WTF
AI companies are purchasing large quantities of used and rare books. Scanning their contents to train models. Then turning the originals to pulp.
The sum of human thought, digitized and shredded.
Every book that gets scanned disappears from the physical world permanently. The knowledge survives only as training data inside a system no one can hold, browse, or resell.
What used to sit on a shelf for centuries now exists as weights in a model that might get deprecated next quarter.
Ejemplo típico de data drift, os atrevéis a evaluar si por menos estrés o por uso de IA?
Este año fue el primero en el que el examen de admisión a la UNAM se aplicó en línea.
Analicé tres carreras: Medicina, Actuaría y Ing Aeroespacial. En todas los aciertos se recorrieron a la derecha muchísimo.
Muchos han dicho que fue por trampa, aquí los datos por año 👇🧵
41
Adrià retweeted
NEW STUDY: The amino acid L-theanine, first isolated from green tea in the 1940s, is used to improve mood and sleep 🍵 Study in mice finds L-theanine suppresses metastasis of prostate cancer, the second most common cause of death in men worldwide... 🧵
Sense aire condicionat, tot bé @FGC ? M’he de portar dues samarretes per treballar
6
318
Adrià retweeted
🔴 ¡MÁS MODELOS CHINOS OPENSOURCE!
Mientras se sigue viviendo el terremoto DeepSeek, desde los labs chinos nos siguen llegando más y mejores modelos.
Hoy por ejemplo Qwen ha anunciado un nuevo modelo Qwen2.5-Max que se sitúa ligeramente por delante de DeepSeek v3 👇
The burst of DeepSeek V3 has attracted attention from the whole AI community to large-scale MoE models. Concurrently, we have been building Qwen2.5-Max, a large MoE LLM pretrained on massive data and post-trained with curated SFT and RLHF recipes. It achieves competitive performance against the top-tier models, and outcompetes DeepSeek V3 in benchmarks like Arena Hard, LiveBench, LiveCodeBench, GPQA-Diamond.
📖 Blog: qwenlm.github.io/blog/qwen2.…
💬 Qwen Chat: chat.qwenlm.ai (choose Qwen2.5-Max as the model)
⚙️ API: alibabacloud.com/help/en/mod… (check the code snippet in the blog)
💻 HF Demo: huggingface.co/spaces/Qwen/Q…
In the future, we not only continue the scaling in pretraining, but also invest in the scaling in RL. We hope that Qwen is able to explore the unknown in the near future! 🔥
💗 Thank you for your support during the past year. See you next year!
Adrià retweeted
🔴 ¡NUEVOS MODELOS OPEN SOURCE!
DeepSeek vuelve a poner todo patas arriba cumpliendo su promesa de liberar (open weights) sus modelos R1!
Modelos disponibles para todos que rinden a la par que el modelo privado más avanzado disponible, o1 de OpenAI 🔥
¡Increíble noticia! 👇🧵
Adrià retweeted
It's a bit sad and confusing that LLMs ("Large Language Models") have little to do with language; It's just historical. They are highly general purpose technology for statistical modeling of token streams. A better name would be Autoregressive Transformers or something.
They don't care if the tokens happen to represent little text chunks. It could just as well be little image patches, audio chunks, action choices, molecules, or whatever. If you can reduce your problem to that of modeling token streams (for any arbitrary vocabulary of some set of discrete tokens), you can "throw an LLM at it".
Actually, as the LLM stack becomes more and more mature, we may see a convergence of a large number of problems into this modeling paradigm. That is, the problem is fixed at that of "next token prediction" with an LLM, it's just the usage/meaning of the tokens that changes per domain.
If that is the case, it's also possible that deep learning frameworks (e.g. PyTorch and friends) are way too general for what most problems want to look like over time. What's up with thousands of ops and layers that you can reconfigure arbitrarily if 80% of problems just want to use an LLM?
I don't think this is true but I think it's half true.
Perque em quedi clar, lo de @infoLaVelada es un event que participen rics (eg youtubers), on els pobres paguen per no veure res (última graderia) i per fer més rics a aquesta gent 🤣 jo la veritat deixaria que gent amateur participés i donés una humil elecció a aquesta gent.
49
Adrià retweeted
Ultra processed foods are putting our kids health at risk:
- A study involving 1,426 children aged 3-6 over a 3-year period found a significant link between consumption of ultra processed foods and negative health markers.
- The children with the highest UPF consumption had increased levels of obesity, fasting blood glucose, waist-to-hip ratio, blood pressure and worse cholesterol status.
- The study modelled the effect of replacing 100g UPFs with minimally processed foods, which helped reduced some of these risk factors.
- These negative health biomarkers can lead to an increased risk for disease and poor health into adulthood, highlighting the need fo us to be aware of, and reduce intake of UPFs.
PMID: 38758555
PMID: 37813465
Adrià retweeted
Just so y'all know:
Every time you push code, you appear on this real-time wall at the @github HQ.

















