if this is agi my singularity is not off to a good start
RESEARCH
-
Small Language Models: The Shift from Bigger to Smaller AI
By
–
you know how AI companies have been laser-focused on making bigger and bigger AI models? now they're focused on making them smaller and smaller! (think cheaper, more focused, more efficient.) read my latest on small language models, or SLMs:
-
Langtrace: Open Telemetry Observability Tool for LLM Applications
By
–
GitHub – Scale3-Labs/langtrace: Langtrace is an open-source, Open Telemetry based end-to-end observability tool for LLM applications https://
bit.ly/3xVsBAD #AI #MachineLearning #DeepLearning #LLMs #DataScience -
Chinese Open-Weights Model Surpasses All Previous MATH Benchmarks
By
–
Chinese open weights model that easily surpasses all previous models, both closed and open, at MATH:
-
AI for Science: Long-term Research vs Commercial Expectations
By
–
It's not all hype for sure … we've been looking at this area for last four years. http://
semanticscholar.org/paper/AI-for-S
cience-An-Emerging-Agenda-Berens-Cranmer/1a56bacaba706357229eb65231749cb9729f7c47
… http://
science.ai.cam.ac.uk But whether there's a commercial return on timeframes VCs expect is a different matter. -
AI4Science Hype Cycle and VC Funding Limitations for Complex Problems
By
–
I have a feeling AI4Science is now entering an overhyped period … AI4Sustainability is great but many sustainability challenges cannot be solved by AI and AI4Sustainability can be a distraction … Increasingly I feel VC $$ are not great at solving "wicked problems" ….
-
14-Year-Old Indian Developer Creates AI Chatbot and Patents
By
–
Kartik, a 14-year-old from Haryana, has created Robotic Vision for Humans, an Indian AI chatbot similar to ChatGPT, and has filed six patents. In addition, he writes books about coding, graphic design, and the future of technology up to 2050 and beyond. Click on the link to know
-
OpenAI Q* Strawberry Announcement Expected Soon
By
–
Mucho ruido incentivado desde OpenAI apunta a que podría haber un anuncio próximo. Algo relacionado con Q*/strawberry. No lo sé, a veces prefiero un blog post anunciando cosas que tanto jaleo por Twitter.
-
ROAM Challenge 2: LATAM Vehicle Classification with Few-Shot Learning
By
–
NEW CHALLENGE ALERT ROAM Challenge 2: LATAM Out-of-Distribution Few-shot Challenge
Develop models that can classify unusual or specific vehicle types from minimal training data, a crucial skill in environments with unique vehicular regulations -
Robot Hand Design History: Learning from Past Failures
By
–
@chr1sa it might be interesting for some of these youngsters to review hundreds of robot hand designs over the last 50 years. None of these ideas are new. Knowing where each of them failed would point to where innovations are needed. Not as good for funding/valuations by