@gabepereyra @deredleritt3r you might be interested in monitoring this space!
English
Noema
773 posts
Noema
@noemaclips
Nostalgic for the future. A lantern into the latent space ✧
The Long Now Katılım Aralık 2015
1.8K Takip Edilen189 Takipçiler
@KLieret @jyangballin @18jeffreyma @parth007_96 @dpedch @sten_sootla @micmylin @pengchengyin @magpie_rayhou @syhw @Diyi_Yang @OfirPress @sootla_sten Made my day! Keep up the good work, you legends :)
English
@noemaclips @jyangballin @18jeffreyma @parth007_96 @dpedch @sten_sootla @micmylin @pengchengyin @magpie_rayhou @syhw @Diyi_Yang @OfirPress @sootla_sten We're definitely planning to also update CodeClash again! @jyangballin has also recently been working on making it easier & cheaper to evaluate
English
How much of SQLite, FFmpeg, PHP compiler can LMs code from scratch? Given just an executable and no starter code or internet access.
Introducing ProgramBench: 200 rigorous, whole-repo generation tasks where models design, build, and ship a working program end to end. 🧵
English
@jyangballin @KLieret @18jeffreyma @parth007_96 @dpedch @sten_sootla @micmylin @pengchengyin @magpie_rayhou @syhw @Diyi_Yang @OfirPress @sootla_sten I hope you get enough support to keep the leaderboard updated from time to time, unlike Codeclash, which is one of my favs :(
English
ProgramBench is a joint effort across Meta FAIR, Meta TBD, Stanford, Harvard
@KLieret (co-first author) @18jeffreyma @parth007_96 @dpedch @sten_sootla @micmylin @pengchengyin @magpie_rayhou @syhw @Diyi_Yang @OfirPress
Paper: programbench.com/static/paper.p…
English
@jyangballin @KLieret Wake up, another banger eval by John Yang just dropped @stalkermustang @scaling01 @xeophon @spicey_lemonade
English
1/🆕 New NBER paper: 𝗪𝗵𝗲𝗻 𝗗𝗼𝗲𝘀 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗻𝗴 𝗔𝗜 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗣𝗿𝗼𝗱𝘂𝗰𝗲 𝗘𝘅𝗽𝗹𝗼𝘀𝗶𝘃𝗲 𝗚𝗿𝗼𝘄𝘁𝗵?
Under empirically grounded calibrations, a singularity could arrive within just a few years of automating AI research. 🧵
📄 nber.org/papers/w35155
English
@Delachica_ @runwayml Brutal!! Para el audio habéis usado IA también? De ser así cuál?
Español
Sincitium is finally here.
We are pleased to present our latest piece: a concept trailer created specifically for the @runwayml Big Pitch Contest. For this project, we wanted to explore a completely different aesthetic from our usual studio style, and this film is the result of that experimentation.
We hope you enjoy it as much as we enjoyed the creative process.
Produced by: Contanimation
Directed by: Javier De La Chica and Guillermo Miranda Art Direction: Javier De La Chica
Editing: Guillermo Miranda
Voices: Juan Rabadán
#runwaybigpitchcontest
English
@j_dekoninck Amazing. Are you planning to develop your own aggregated index? I'd love something similar to Epoch's ECI or the Artificial Analysis Index but for Matharena. Keep up the good work!!🫶
English
Introducing a new paper! Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMs
Static benchmarks are no longer enough. Models improve too quickly and numbers become stale quickly. Instead, we argue for continuously maintained evaluation platforms.
English
Common Codex (app) bug - you close the window but the app doesn't close, and then when you try to open up a new window, it doesn't work, and you have to reset the app
English
@Ahmad_Al_Dahle Curious to learn why you didn't make this bet @ Meta though 👀
English
The most interesting thing about DeepSeek-V4 isn't the benchmarks, it's the bet: efficient ultra-long context is the precondition for test-time scaling and long-horizon agents. 27% of V3's FLOPs at 1M tokens! The rest flows from this ...
DeepSeek@deepseek_ai
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/De… 🤗 Open Weights: huggingface.co/collections/de… 1/n
English
Aquí en súper evidente! Fijaos en el agua y la espuma
x.com/i/status/20469…
🍔 Campero De Pollo 🍔@campero_depollo
@DotCSV Descaradísimo en el agua 🌊
Español
¿Habéis detectado ya cuál es el "tinte amarillo" de la nueva versión de imágenes de ChatGPT? Yo lo tengo claro 🤗
Os dejo un examen visual, pero cuidado porque una vez lo veais no parareis de notarlo.
Español
Soon after we’ll have generative ai for aromas, and I’ll be out here prompting stuff like “the smell of that old book, but with a hint of jasmine” and “newly opened apple product with freshly cut grass”
Pirat_Nation 🔴@Pirat_Nation
Researchers have created a new device that uses ultrasound to trigger smells directly in the brain for VR The small device rests on the forehead with a soft gel pad. It sends gentle sound waves through the skull to the area of the brain that processes smell. Current VR smell systems use cartridges that need constant refills and can make a mess. This new method is cleaner and could one day let people smell virtual worlds like forests or oceans naturally. The project is still in early prototype stage from a small research team. Safety levels were kept low, but more testing is needed.
English
@spicey_lemonade That really looks like images v1.5, are you sure it's the new one?
English
GPT Image 2 failure point: I tested whether the model could reconstruct a figure from a physics paper using only the text description, and it generated a completely unrelated image. It seems like it can’t handle long context. This reinforces the idea that it’s not part of an Omni 5.5 model, but rather a separate image generation model that the system routes to
English
@noemaclips The models have done pretty much exactly as well as I expected
English