Interesting article, makes me want to consider Notion again. BUT what I really want is: to be able to ask a question and be given an answer informed by all the specific notes I’ve taken in the past. Ideally all the documents I’ve saved as well. Is that too much to ask??
Awesome, thanks for the video. Key sentence, though: "we're starting with individual files, but over time we'll include folders and your whole Dropbox collection." So....not yet
Thanks for this article Chantal- I like how you summarised the different AI use cases. Thanks to the article I tried Claude and Perplexity. I really like both and from now on I'll use Claude instead of Bing. I found Claude more succinct for the exercise (redoing my cv) I tried today. I moved away from Chat GPT to Bing at the beginning of the year after reading a couple of Ethan Mollick's training articles.
I'm a Patreon of somebody who uses notion, I also digressed my way into watching people who got hog deep into this, using an app designed to backup you brain. Can't find my notes though, may be on my desktop at home. (It's called obsidian https://obsidian.md)
I suspect you can do this with Google if the web site is open, I did that with the Google assistant, though this may only be relevant to people with a Pixel phone as it was a different assistant voice than the full fat web version so it may be on device. Essentially I asked it to narrate a long web page, 40 minute read. Which it did with aplomb.
That said I've been asking natural language questions of Google for years as I was reading their blog where they announced that they were trailing functionality, etc.
I remember reading that you could interact with documents having a conversation with them, though I forget what that was.
Personally my favourite Generative AI tools are chatbots. They force me to really think, to not be passive, and not rely on the crutch of copy paste. You have to be creative and in the moment. As such I think of hallucination as creativity as opposed to error.
As Hayao Miazaki said, "Reality is for people who lack imagination"
Chantal - thanks for the article which is interesting. I’m with Azeem and have moved to Claude! Whilst advocating using Generative AI tools at work the main challenge put to me is how much energy is used every time you ask a LLM a question compared to using a search engine. I know a lot has been written about the amount of energy it takes to train LLMs and therefore the carbon footprint associated with them but have you come across anything that compares the amount of energy consumed to answer a query in Claude, ChatGPT or Bard? Or produce an image in mid journey, leonardo ai, or dalle-2? And how this compares to other daily activities like sending an email, 1 hours Teams call, making a cup of tea or coffee, flying to see a client?
Generative AI is a process. If you're looking for a given thing, a productivity hack, I think you're "doing it wrong" in my view at least the journey is it's own reward.
Interesting article, makes me want to consider Notion again. BUT what I really want is: to be able to ask a question and be given an answer informed by all the specific notes I’ve taken in the past. Ideally all the documents I’ve saved as well. Is that too much to ask??
That would be my absolute dream as well
My understanding is that is what Dropbox’s Value Proposition is with Dropbox AI and Dropbox Dash https://youtu.be/7Zh6PrC-lGo
Awesome, thanks for the video. Key sentence, though: "we're starting with individual files, but over time we'll include folders and your whole Dropbox collection." So....not yet
Definitely the dream
Thanks for this article Chantal- I like how you summarised the different AI use cases. Thanks to the article I tried Claude and Perplexity. I really like both and from now on I'll use Claude instead of Bing. I found Claude more succinct for the exercise (redoing my cv) I tried today. I moved away from Chat GPT to Bing at the beginning of the year after reading a couple of Ethan Mollick's training articles.
Thanks for the kind words Margaret! You're not the first to tell me Claude has been more useful, I may have to change my habits.
Thanks for sharing this Margaret, so happy to hear!
I'm a Patreon of somebody who uses notion, I also digressed my way into watching people who got hog deep into this, using an app designed to backup you brain. Can't find my notes though, may be on my desktop at home. (It's called obsidian https://obsidian.md)
I suspect you can do this with Google if the web site is open, I did that with the Google assistant, though this may only be relevant to people with a Pixel phone as it was a different assistant voice than the full fat web version so it may be on device. Essentially I asked it to narrate a long web page, 40 minute read. Which it did with aplomb.
That said I've been asking natural language questions of Google for years as I was reading their blog where they announced that they were trailing functionality, etc.
I remember reading that you could interact with documents having a conversation with them, though I forget what that was.
https://theresanaiforthat.com/ai/docu-talk/ There's one for pdf's too
Personally my favourite Generative AI tools are chatbots. They force me to really think, to not be passive, and not rely on the crutch of copy paste. You have to be creative and in the moment. As such I think of hallucination as creativity as opposed to error.
As Hayao Miazaki said, "Reality is for people who lack imagination"
Thanks for sharing! I definitely have to give obsidian a second chance, too
Chantal - thanks for the article which is interesting. I’m with Azeem and have moved to Claude! Whilst advocating using Generative AI tools at work the main challenge put to me is how much energy is used every time you ask a LLM a question compared to using a search engine. I know a lot has been written about the amount of energy it takes to train LLMs and therefore the carbon footprint associated with them but have you come across anything that compares the amount of energy consumed to answer a query in Claude, ChatGPT or Bard? Or produce an image in mid journey, leonardo ai, or dalle-2? And how this compares to other daily activities like sending an email, 1 hours Teams call, making a cup of tea or coffee, flying to see a client?
Generative AI is a process. If you're looking for a given thing, a productivity hack, I think you're "doing it wrong" in my view at least the journey is it's own reward.
Thank you for the kind words, Ben. I'm not sure about the comparative carbon footprint, and I'm not sure the data is available given OpenAI's lack of transparency. According to this article (https://carboncredits.com/how-big-is-the-co2-footprint-of-ai-models-chatgpts-emissions/#:~:text=As%20reported%20by%20the%20MIT,the%20more%20energy%20it%20needs.) training a model can emit 5x the lifetime of an average passenger car.
I haven't found any compelling data on the energy consumed to answer a query, however. I would be interested if you or someone else found some.