Originally started learning to code from Sentdex Took a CS class in my first year at college, failed miserably… it was totally useless so I wasn’t paying attention Sentdex actually made it practical and interesting
@mattshumer_
-
Unified Voice AI Model Architecture vs Pipeline Systems
By
–
Kind of. More like the voice assistant systems we see today, but a singular model, rather than a pipeline of speech to text, plus LLM generation, then text to speech, with interruption handling built in. Who knows? Maybe the gpt2-chatbot models are just the language output
-
LLMs for Product Analysis: An Underrated Application
By
–
Using LLMs for product analysis is criminally underrated
-
Voice Technology Enhances AI Use Cases and Enables New Applications
By
–
Agreed. Voice won’t be right for everything, but it will improve *some* existing use-cases and enable new ones.
-
OpenAI Voice Assistant: End-to-End System Breakthrough Expected
By
–
If OpenAI really is releasing a voice assistant tomorrow, it’s highly likely to be a true end-to-end system. It’s the natural way forward… a somewhat-better multi-model system won’t be enough to really wow anyone. End-to-end with low latency would be a breakthrough.
-
Open Source Usage Rights and OSS Philosophy Debate
By
–
Huh? He didn’t steal, and didn’t claim to invent it. The whole point of OSS is so that others can use it. This take is way off-base.
-
Anthropic Haiku Rate Limit Increase Request
By
–
Hey @AnthropicAI just hit our Haiku rate limit in prod Any chance we can get that bumped up a bit? Thanks! cc @alexalbert__
-
Training AI Models on Synthetic Roleplay Input Data
By
–
Have you tried training a model on actual RP inputs to act as the user? Write a couple hundred of them yourself, it’ll go pretty far.
-
LLaMA 3 Model Outperforms GPT-4 on Classification Tasks
By
–
A few days ago, a team released a model/product (trained on LLaMA 3) that beat GPT-4 at classification tasks for training data generation. Anyone remember the name?
-
Testing Prompts for Evaluating LLM Reasoning and Instruction Following
By
–
What reasoning/instruction following prompts do you use for testing new LLMs? I want to test them on a new model I'm training.