The current version on ChatGPT(dot)com is still using DALLE Tool Call, but in the future it will be using the new natively multimodal abilities, which will render all of these use cases a reality.
LLMS
-
Improving AI Benchmarks: TIGER-Lab’s Enhanced MMLU Evaluation
By
–
We need to improve our benchmarks the same way we improve our models. Super interesting work from TIGER-Lab with an upgraded version of MMLU with 12k complex questions (vs. 16k for MMLU) and additional reasoning problems. – It is more discriminative among frontier models (see
-
New LLM Benchmark: Gathering Withheld Test Sets from Community
By
–
proposal: should we start gathering –from research teams and the community– new withheld test sets for a new LLM benchmark? infra/backend/leaderboard templates on the HF hub are super adapted for this so it would be very low effort to maintain it even over the long term
-
OpenAI voice model release strategy and feature integration
By
–
It is still behind a feature flag. It won't make sense for OpenAI to release it w/o a new voice model tbh
-
Technical limitations of GPT-4o interaction
By
–
Nope, only the UI. It is still a publically available 4o. You cannot interrupt it and it cannot read from the open camera yet
-
Interactive UI behavior and content generation in ChatGPT
By
–
Yes, this space in the middle works like an interactive display. What is shown there is not what ChatGPT spells but rather helpful content or subtraction of it (words, lists, summaries etc)
-
Using ChatGPT for language learning
By
–
ChatGPT got a new icon for voice conversations. I found a good use case: learning German. If you ask ChatGPT to be your teacher, it can show words and how they’re written on the display. I think it’s super useful
-
Jovu: AI Platform Generates Production-Ready Backend Code
By
–
The team at @amplication just launched "Jovu" on @ProductHunt Jovu is the world's only AI that can create production-ready backend code. Now that they support both Node.js and .NET, it's even more amazing. Do try Jovu and support them: