More reasons to check out Poe.com
The Swiss Army Knife of AI Platforms
Poe.com is the often unsung hero of the generative AI (GAI) world. Poe is a front end into a huge array of different large language models (LLMs), which are the engines that fuel the various GAI chatbots. Poe has a chat interface, just like ChatGPT, but with a huge difference. With Poe you can specify which of over three dozen LLMs you want to use. So, you can chat with Claude-3.5-Sonnet, GPT-4o, Grok, Gemini-2.0, Llama-3.1 and many others. You can create images with DALL-E, Stable Diffusion, Ideogram, FLUX, among others. It’s even possible to switch models within a conversation. Start off with Gemini doing your research, then ask Claude to help you write, and wrap up by creating a representative image in DALL-E, all within a single chat! Poe even lets you create custom bots and access custom bots created by others.
Several months ago, I declared Poe to be the AI tool to use if you could only choose one. I feel even more strongly about that recommendation today. Poe is simply fantastic. Seriously, if you haven’t tried Poe, you’re missing out.
New Affordable Pricing Plans
So, why am I writing about Poe again? Two reasons. First, they recently released a new subscription plan at only $10 per month. Poe works on a point basis. Different LLMs cost different amounts of points. For example, Poe’s Web-search model costs 15 points per message, while Claude-3.5-Sonnet (a more advanced model) costs 344 points per message. The costs are actually more complex than this since longer messages are more expensive, but this gives you the general idea. For $10 per month, you get 10,000 points per day. $20 gets you 1 million points per month. (Don’t ask me why one is per day and the other is per month.) Poe also has more expensive plans, all the way up to $100 per month for 5 million monthly points and also has a free tier, but that doesn’t give you access to more advanced LLMs. (It’s all a little confusing, but I like the range of options.)
Voice Generation: A Game-Changing Feature
What’s more exciting is Poe’s recent addition of voice generation. You can now use Poe to generate synthetic speech using voices from Eleven Labs or Cartesia. Both systems allow you to generate audio using a variety of voices. To test this, I used Poe’s Web-search model to generate a script for a solo podcast episode on recent developments in generative AI. Then I asked Poe to read the script using Eleven Labs. In a matter of a couple of minutes, I had a three minute podcast episode. I also tried using Cartesia with similar results. Poe also lets you specify voices, or re-generate the audio with a different voice using its revoice command. It’s mind-blowing.
These synthetic voices are getting better and better. In my Cartesia trial, the audio included natural sounding breath noises. Cartesia did go off the rails a bit by reading out web addresses, but I swear I could hear the exasperation in the digital voice. (I could have avoided this by stripping out the web addresses before asking Poe to generate the audio.)
Applications in Higher Education
This has all sorts of uses in higher ed. For example, you can provide a script for a lecture or tutorial, then generate the audio with Cartesia or Eleven Labs. This could be useful for team-taught online classes or for creating mini-lectures or tutorials that aren’t tied to a particular faculty member. Student affairs folks could easily create audio-based FAQ files for new students.
Easy access to AI voices also holds implications for accessibility on campus. With a bit of effort, universities can build audio versions of virtually any document or website (assuming there are no copyright issues). With the cooperation of publishers, AI voices could convert portions of textbooks into audio files. Here’s a plausible scenario. A professor uploads learning materials (such as a textbook chapter or even a detailed set of slides) into their favorite AI tool, which creates a summary based on the professor’s criteria (such as topics to focus on, depth, etc.). The professor then edits the summary to their liking, and uploads the script into an AI voice generator and in the space of a few minutes has an audio summary of the learning resource. Yes, there are intellectual property issues to consider, but hopefully those will be overcome at some point.
Soon, very soon, the limitations for these sorts of applications will be our imagination and administrative and legal barriers, not technological barriers. The tech is pretty much there. This has huge implications for accessibility, as I noted earlier, but also for equity. Basically, with AI we now have tools that can both generate speech from text, but can also transcribe text into speech, which lets anyone make their content more accessible relatively easily and inexpensively.
Making AI Tools More Accessible and Affordable
That brings me to the beauty of Poe. With Poe, you can do all of this within a single tool. Many people, including me, are concerned about the cost of AI tools. I’m afraid to even tally what I spend on AI every month, but I’m sure it’s well over $100 (probably closer to $200 or more). Many people, especially students from less advantaged backgrounds, can’t afford this. Although Poe has its limitations, it does make a wide variety of excellent tools available for as little as $10 per month in the USA. (By the way, as a test I created an audio summary of a slide deck for one of my lectures, entirely in Poe. It worked quite well.)
In closing, if you haven’t checked out Poe, I suggest doing so. Even if you like other tools more, you might want to recommend Poe to your students or to colleagues who are concerned about cost. Based on my experience, they’ll thank you.
Connect With Me
Well, that’s all for this time. If you have any questions or comments, you can leave them below or email me - craig@AIGoesToCollege.com. I’d love to hear from you. Be sure to check out the AI Goes to College podcast, which I co-host with Dr. Robert E. Crossler. It’s available at https://www.aigoestocollege.com/follow. Thanks for reading!