Productivity Doesn't Mean AI Silo-ing
Sharing the Love with AI Workflows Rather Than Being Exclusive with Multimodal Tools
Image created with ChatGPT 4o
Why I’ve Been Pigeonholing Myself (and You)
For the past few months, I have been preparing and administering a bootcamp with Steve Hargadon (see the plug below). this bootcamp is focused on promoting professional productivity related to AI tools. As you probably can guess, it focuses on collaboration with AI tools rather than automation. It also promotes the idea that the human using the AI tool is ultimately responsible for its output and the product created from that output.
As with any practitioner, Steve and I have our favorite AI tools for different uses. Our preferences differ from each other. Steve likes Perplexity, and I like ChatGPT for most cases. I have also enjoyed ChatGPT-4o’s multimodal abilities. I demo’d analyzing and visualizing datasets with ChatGPT in a tutorial for the bootcamp and was very impressed with its abilities and the fact that it could do more than I asked it to. I appreciated ChatGPT’s role-playing technical communication skills. I also looked at Claude’s ability to create engaging scripts.
However, I focused on ChatGPT almost exclusively in my work. I personally prefer it, and its ability to switch from AI tools and combine their products is truly impressive. It may not democratize creativity, but it certainly makes creativity more time-accessible and has an easy-to-use interface.
These abilities, and my tendency to remain with what I find comfortable with new technologies, led me to pigeonhole myself and you, my readers. When talking with Steve, I found that Perplexity can analyze data (and visualize data, although this costs more money) much better than ChatGPT can. I also found out that while Claude cannot create images, it can analyze images and PDFs and retain their information in easy-to-navigate Projects. This does not require the creation of a Custom AI tool (as OpenAI does).
Bursting Out of Said Pigeonhole
There are multiple reasons that I do not like Perplexity. Mostly, because it claims to primarily be a search engine but does not follow the robots.txt rules. Furthermore, it makes things up. This type of hallucination is fine when you are strictly an LLM. However, when you tout your tool as a “conversational search engine,” there is a different type of expectation.
Those issues aside, Steve demonstrated that the ability of Perplexity to assess data far surpassed GPT 4o.
There are other bots (and manual functions) that outperform GPT 4o in other areas:
Ideogram makes better images.
Text-to-Speech makes better voiceover recordings.
Microsoft and Google Designer make better slide presentations.
Any manual process makes better PDFs if what you need on them is complex.
Unless are already deep in the process with ChatGPT, it will be just as easy to select spreadsheet columns and ask for a visualization from Excel as it will be to prompt ChatGPT.
I suppose what I am trying to say is that just because you can use AI for something doesn’t mean that you should. Also, just because an AI tool can do something does not mean that you should only use that tool for that function. Be well-rounded and experiment with many tools. In my trainings, I try to emphasize that I use ChatGPT and others use other tools, and that is fine. But in my own practice, I should probably be more well-rounded.
AI Siloing
Ofttimes when we have work projects or educational programs, even if they are interdisciplinary, we tend to sequester ourselves into groups based on some mutual defining characteristic. We only associate with members of that group, and the instructors have to cajole us to actually communicate with people. This term is called “siloing,” and it happens in virtually every group of which I have been a member. The worst occurrences are when we create silos of one, where no one talks. I hate that.
Well, we can do that with AI tools as well. We only use one tool for a task, and we shut ourselves off from learning about others, or at least experimenting with them. This is often done with multimodal tools. After all, it can do all of the steps in a project. So why shouldn’t we just use one tool for the whole process?
Why We Should Use Multiple Tools For the Same Task
When I talk about using multiple tools, I don’t mean having five text generators create text for the same product (although that has its benefits). If you know what tool you would like to use for a particular aspect, then use it.
What I am talking about is using Claude for the text, Ideogram for the image, Bark for the voiceover, Perplexity or ChatGPT for data analysis and summary, and Capcut for the video creation. And this is just assuming that these tools have been vetted by you. Don’t use a tool just because everyone is using them. Experiment with them and decide which tool is most effective and productive for you and your team. If you truly think that a multimodal tool is best at all of the steps, then use it. But make sure that it is actually the best choice.
References
Something else I’ve been doing wrong is that I have been linking to materials and not referencing them. I have rationalized that because you can click the link and see the material, you don’t need to see the entire reference. While this is appropriate for many blogs, this is an academic blog and as such needs a References section.
Marchman, D. M. and T. (2024, June 19). Perplexity is a bullshit machine. Wired. https://www.wired.com/story/perplexity-is-a-bullshit-machine/
T., D. W. (2024, June 20). Perplexity ai under fire for unethical practices. LinkedIn. https://www.linkedin.com/pulse/perplexity-ai-under-fire-unethical-practices-diana-wolf-torres-fsfjc/
The Plug (Again)
As this is a free blog, and as I have a consulting business that I run, I assume that you will forgive me for advertising briefly for a commercial function.
The first session of this happened last week, and the next two occur on the next two Fridays! In coordination with Steve Hargadon of Learning Revolution, I am hosting a bootcamp on professional productivity with ChatGPT and other AI tools! You can register without feeling like you need to attend all three sessions, because recordings will be available forever afterward.
$149/person, $599-999/institutional license
For more information scan the QR code below, or go to this link:
https://www.learningrevolution.com/professional-productivity
