Saw the OpenAI GPT-4o showcase video that was released today.
In it, several new features are showcased:
- Two AI bots chatting casually with one another, using voice.
- Bots being interrupted during speech and being able to carry on the conversation afterwards.
- Live video processing (recognizing what it sees) by ai bots straight from a consumer phone.
- Retains memory of “what it saw” afterwards and can tell what it saw.
- Oh, and it also can do improv singing now.
I just tested it myself and it now also easily generates Venn diagrams for any three concepts you throw at it. It also does any other kind of chart visualization of data that it has. It’s also much faster than before.
Clearly the AI now has access to Python libraries for creating charts, which is awesome and it might help me rely on Microsoft Excel or Libreoffice a bit less for simple charts.
On my wishlist for future integrations was also the ability to generate MIDI files (free beats!) which seems very close to being available as well:
The speed of innovation at OpenAI is pretty amazing. It really hasn’t been this long since we were all blown away by the talking ChatGPT4 bot.