AI Resources
An up to date comparison of the best AI models
Last updated December 18.
There are too many developments in AI these days for me to keep an updated table here, so I’ve changed format to a short summary.
Models
The current top models are Claude 3.5, ChatGPT o1, and the new Gemini 2. The latest Meta Llama models are also good, but limited by the fact that for most of us all we can with them is basic chats. Grok 2 appears to be good too, and now available for free on Twitter. The Gemini 2 models we have access to are still experimental, so expect even better versions already in January. Hopefully soon: Claude 4.
General Tools
ChatGPT is probably still the best choice for most people who will only pay for one AI tool. It has gotten smarter recently (o1) and has more bells and whistles (advanced voice with and without video, and the new Canvas and projects functions). I still prefer Claude, they are just a little behind on shipping new things right now.
The breakout star of the year is Google’s NotebookLM which just got even better this week. If you haven’t tried it yet, you really should! Google in general is on a roll: I now recommend paying for Gemini Advanced both to use gemini.google.com with the new deep research function, and because of the integration with Docs/Sheets/Slides/Drive etc. The biggest WOW moments I’ve had with AI lately were all with the new real-time streaming AI Google offers at aistudio.google.com/live. Try both the video and the screen sharing options. This is the future.
Specialized Tools
Perplexity.ai is still best for search and adding functions all the time. I still use Otter.ai for transcribing and summarizing meetings, but it isn’t improving at the same rate as other tools. I still use Ideogram.ai to generate images (especially when I want text in the images or want to specify exact formats/ratios), but the best image generator now is Google’s Imagen 3 at labs.google/fx/tools/image-fx. I used it to generate this image of the ancient Roman goddess of Artificial Intelligence.
