AI development ecosystems are increasingly complex and challenging to maintain. Companies like Meta need to develop highly efficient systems to build, serve and improve AI models for production uses. Here's how we make this work effective + efficient
GENERATIVE AI
-
Comprehensive Survey of AI-Generated Content from GAN to ChatGPT
By
–
A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPT A great survey that provides a comprehensive overview of generative models, common models(unimodals, multimodals), and recent advances in generative AI. https://
arxiv.org/pdf/2303.04226
.pdf
… -
ChatGPT connected to Internet: major update and video test
By
–
OFFICIAL! #ChatGPT is connected to the Internet Major update that will change everything. Analysis and test in my latest video → https://
youtu.be/jIBpsJGgE-Q #GPT4 #OpenAI #OpenAIChatGPT #ArtificialIntelligence -
Broad AI Era: GPT-4 Multimodal Transformers and Path to AGI
By
–
GPT-4 and other Multitasking Multimodal Transformer based models demonstrated that we are in the era of Broad AI (Artificial Broad Intelligence, ABI) not Narrow AI and not AGI either. It will take solving for causal reasoning before we really approach AGI & ideally compressing
-
GPT-4 Demonstrates Advanced Reasoning and Common Sense Capabilities
By
–
I was reacting to the title but the paper itself is interesting. It looks at stuff that GPT-4 can do that seems to have been missing from AI systems (more common sense reasoning, grasping the elements that make an animal recognizable, and so on). It's well worth a look.
-
ChatGPT Now Connected to Internet with Safety Focus
By
–
#ChatGPT is now connected to the internet! Notice focus on safety!
-
Using ChatGPT for news publishing business automation
By
–
Hired an assistant for 20$ a month to help me build TestingCatalog news publishing business. Well… in reality, I've subscribed to ChatGPT+ to get early access to its gpt-4 model so I can share my experience over here. Inspired by #HustleGPT Follow me on this journey here
-
Transformer-Base Training Efficiency on TPU V2 Hardware
By
–
As I said above, use of Transformer-Base as proxy task *is* in So et al: "Specifically, to train a Transformer
to peak performance on WMT’14 En-De requires ∼300K
training steps, or 10 hours, in the base size when using a
single Google TPU V.2 chip, as we do in our search" -
GPT-4 adoption and future workforce implications
By
–
Quelqu’un qui n’utilise pas #GPT4 cela s’appelle un futur chômeur !