I played with GPT-2 a lot, and when GPT-3 came out, it was remarkable how much it improved. Even if GPT4 is only 3x more powerful, having seen the evolution before, I'm super excited by the potential
LLMS
-
Building a Better AI Assistant: Whisper, ChatGPT, VALL-E Stack
By
–
The perfect recipe to build a Siri or Alexa 10x better! 1. Whisper will listen to what you say.
2. ChatGPT will create a response.
3. VALL-E will read the response with the natural voice you want in seconds! We're close to JARVIS! -
Thank AI Models to Help Them Learn Better
By
–
This is going to sound like a joke, but it isn't. Machine learning models like ChatGPT benefit from positive reinforcement. It helps them learn when their responses are accurate and effective. So say THANK YOU to your favorite bot. You're helping it get a little smarter.
-
OpenAI’s Superior PR Strategy Against Google’s LLM Advantage
By
–
Google was ahead of OpenAI in large language models, but OpenAI played the PR game much better.
-
Copilot Limitations: Model Access and Customization
By
–
Technically, Copilot, but there’s no setting of Copilot where it just gives you direct access fo the model or lets you choose other models.
-
Foundation Models Transform Legacy Workflows at Virtual Summit
By
–
Join us online at the Foundation Model Virtual Summit, presented by @SnorkelAI
. Jimmy Lin, #NLP product lead for @SambaNovaAI
, will discuss use cases where foundation models are transforming legacy workflows as well as a look to the future. https://
snorkel.ai/event/foundati
on-model-virtual-summit/
… #ai -
ChatGPT Preferred for Coding Queries Over Alternatives
By
–
I like chatGPT's responses to coding queries better (more detailed, cleaner UI, and self-explanatory).
-
Cerebras Launches AI Model Studio for Transformer Training
By
–
ICYMI, Cerebras launched the AI Model Studio, a cloud-based pay-by-the-model computing service that makes training large transformer models fast, easy, and affordable. Read more in this Forbes article: https://
hubs.li/Q01xxQDx0 Sign up for a free trial: https://
hubs.li/Q01xxX3x0 -
Instruction-tuned Models vs Creativity: When Older Models Excel
By
–
I love OpenAIs instruction-tuned models. They make it so easy to get what you want, quickly. But sometimes their outputs are just too “down the middle”. For some applications where creativity is key, older, non-instruction-tuned models can actually perform far better.