ChatGPT Aims to Become a One-Stop Shop, While Claude Goes Global
New product rollouts shake up the essential AI toolkit I recommended just 2 weeks ago
In my last post just two weeks ago, 4 must-have AI tools for knowledge workers, I highlighted four tools I thought almost every knowledge worker should be using:
Claude for writing, creative, or strategy tasks that require more than 3000 words of context and don’t live internet access
ChatGPT for any other writing, creative, or strategy tasks that don’t require live internet access
Perplexity Copilot for live web research
AI voice-to-text tools such as Fathom.video and Inflection’s Pi chatbot (a bit of a cheat category, I recognize)
But I started out the whole post by saying “If you feel overwhelmed by keeping track of what’s happening in the AI tools market, you’re in good company. It’s literally my job and I still feel constantly behind.”
There’s a name for the feeling this constant state of change induces: AI vertigo.
So I guess it shouldn’t be surprising that since then I already have 3 major updates to share on the very same use cases I highlighted for those four tools. So here we go:
Claude expands from 2 countries to 95! Huge news on Claude (the ChatGPT alternative that has the much larger context window and allows you to attach files): Last week it rolled out to 93 more countries (with the notable exceptions of the EU, Canada, and of course China).
—> Try it out for free at claude.ai.
You might not need Perplexity Copilot anymore (??!?) In my previous post I said “Perplexity Copilot is essentially an alternative to Google… [it’s] kind of what Bing and Bard promised to be, but Perplexity Copilot is way better… at a lot of stuff.” Well, after that post, I got access to the new iteration of ChatGPT’s new web browsing feature, and it’s a big deal. (Much better than it was when the 1.0 version was briefly available earlier in the year!) I'm still putting it through the paces, so stay tuned, but I think Browse with Bing might supplant more than half of my Perplexity Copilot use.
Here’s how you use it:
Log in at chat.openai.com
Click on GPT-4 in the middle right of your screen (and upgrade to ChatGPT+ if you haven’t already)
Mouse over GPT-4 (or top if you’re on mobile, but don’t click if you’re on a computer). You should see a drop-down menu with some of these options:
Select Browse With Bing (you won’t see Advanced Data Analysis or Plug-Ins unless you turn on in your settings — but you should have Browse with Bing and DALL-E 3 now1)
In this chat, ask something that requires live web access, like “have there been any recent changes to non-profit tax law?” or “what are the current most highly-recommended AI music generation tools?” (Note: Sometimes, with borderline cases, you need to prompt ChatGPT explicitly to browse the web.) The answer will include footnotes linking to the source for each piece of information.
You can now have conversations out loud with ChatGPT. In my previous post, the fourth “tool” I named a bit of a cheat: “AI voice-to-text” tools as a general category rather than a specific tool. One big use case of transcripts for meetings does still require a specific tool like Fathom. But I also suggested chatting out loud with Inflection’s Pi chatbot as a use case in this category. Since that post, you can now hold a conversation out loud within ChatGPT on your phone. Here’s how:
Download ChatGPT’s app onto your phone
Go to Settings → New Features on the mobile app
Opt into voice conversations.
Then, return to your home screen and tap the headphone button located in the top-right corner of the home screen.
Start your conversation!
If you’re a verbal processor, chatting out loud with a chatbot can be a great way to noodle, reflect, or problem-solve without needing to find a person to talk with. You can also use this feature to take a voice memo in an interactive kind of way, or give long-winded, rambling instructions to ChatGPT to write something as you think about it out loud.
ChatGPT has also rolled out other big new features in the last few weeks, notably DALL-E 3 and Vision. You can read the basics in their announcement post: ChatGPT Can Now See, Hear, and Speak. More on those soon.
I hope to write a whole post soon on ChatGPT’s new Vision and image generation capabilities.