I have started exploring the use of ChatGPT, an AI large language model and infinitely patient conversational agent that is all the rage these days. Are others in the Forums interested in this application and in collaborating and sharing thoughts / experiences?
I intend to engage with it as my GTD coach and to get its help in reviewing my projects and next action lists from Asana. E.g. . Here is an example of the prompts I have tried: Using the GTD "Getting Things Done" methodology, does the "description" have a clear and compelling desired outcome? Are the subtasks clear "Next Actions"? What refinements can I make to either/both so that they are more compelling and clear? [paste in from Asana export]
The feedback I receive is generally helpful and on point - it feels like the support that a "coach at my fingertips" would provide:
Based on the information provided, the task title "BDC Gap Analysis" is clear and concise, but the task description could benefit from some clarifications to make the desired outcome more compelling and specific.
Firstly, it is not entirely clear what the intended outcome of the BDC Gap Analysis is, or why it is important. In order to make the description more compelling, it would be helpful to include some additional context or background information that explains the purpose of the analysis and how it fits into the broader goals of the organization. For example, you might include a sentence like: "The BDC Gap Analysis is a critical component of our annual compliance reporting process, as it allows us to identify areas of risk and ensure that our compliance program is aligned with regulatory requirements."
Additionally, the subtasks/next actions section is currently empty, which makes it difficult to assess whether the tasks are clear and actionable. In order to ensure that the tasks are clear and specific, it would be helpful to break down the BDC Gap Analysis into smaller, discrete tasks that can be completed in a specific timeframe. For example, you might include tasks such as:
Review and finalize the request list for the Adviser.
Review the pilot format for [redacted] with [redacted] and identify any necessary changes.
Conduct the BDC Gap Analysis using the finalized request list and pilot format.
Review the results of the analysis and identify any areas of risk or non-compliance.
Develop a plan to address any identified areas of risk or non-compliance.
By breaking down the BDC Gap Analysis into smaller, specific tasks, you can ensure that each step is clear and actionable, and that progress can be easily tracked and measured.
In summary, to make the BDC Gap Analysis task more compelling and clear, you should provide additional context or background information to clarify the desired outcome, and break down the task into smaller, discrete subtasks or next actions that are specific and actionable. This will help ensure that the task is completed efficiently and effectively, and that progress can be tracked and measured.
I am not an expert on AI or LLM's but in my brief experience playing around with ChatGPT, I found that it is really not very much better than Google. I didn't use it for GTD, per se. I wouldn't know what to ask it as GTD is not a problem in so much as it is a systemic process of habitual behaviors for getting clear space for your psyche. I could no more ask ChatGPT how to make me feel happy or have free space in my psyche than I could ask a random person what makes me sad or what has my attention and have that person intelligently answer.
However, I did it ask it a range of general knowledge questions, technical specific questions, and other miscellaneous questions. The key distinction being that ChatGPT necessitates that you already know the answer and really are just using it for confirmation (either that you're on the right track or that you are indeed going down a rabbit hole). Granted, it is just a research preview, the biggest issue I found with it is that its data is already 1.5 years old (which is considered "old" in the technology landscape). Is that a limitation of the technology itself or the version released or something else entirely?
I also experienced many cases where it couldn't figure out what I was asking (i.e. couldn't parse my English phrasing) and just kept throwing an error. Other times, many times, because I already knew the answer, the answer it provided was completely wrong and often had to go through numerous rounds of me following-up with "Er, I think you missed a requirement here..." before getting to a passable answer. It also never could cite its sources or explain how and where it got its information.
At a deeper level, one should be quite cautious on how "awesome" everyone is finding it, how "game changing" they are claiming it to be, and the "nothing will ever be the same" everyone says. It might become a crutch, not unlike smartphones, whereby people no longer know how to independently operate without the tool. As well, I find myself asking this basic question: "How does this make money? It was an exorbitant waste of money and therefore there needs to be a financial recuperation of investment. What is the business model for monetizing this?" I am reminded of the adage "If you're not paying for the product, you are the product". Is the business model the "addiction model" whereby the company gives you a small, free sample to start off with that gets you "hooked" only for you to then needing to keep coming back for more, of which now costs significantly more? What is the end game for it?
I am skeptical of the value, I am even more dubious of the creators.
Your experiences mirror mine. So far all of my ChatGPT atttempts have been totally worthless. It does notcie references if pushed , but in my case they were bogus and falsified and never evne existed! It missed critical new papers in the are of interest that I know about and focused on reseach that is widly discussed but inaccurate according to numerus more recent studies. The only area where it held a bit of promise was in coding a very specific task in a specific language but even then the resulting code did not runa nd needed a lot of work to get running properly.
Your experiences mirror mine. So far all of my ChatGPT atttempts have been totally worthless. It does notcie references if pushed , but in my case they were bogus and falsified and never evne existed! It missed critical new papers in the are of interest that I know about and focused on reseach that is widly discussed but inaccurate according to numerus more recent studies. The only area where it held a bit of promise was in coding a very specific task in a specific language but even then the resulting code did not runa nd needed a lot of work to get running properly.
Agree that it is not good as a "truth engine" - I pushed it on a very specific regulatory topic and it flat-out made stuff up. Including convincing references, that turned out completely bogus. "Several law firms have also published articles or alerts that confirm that revolving credit facilities and delayed draws are generally considered to be derivatives transactions under Rule 18f-4. For example, the law firm Ropes & Gray has published a client alert that notes that "revolving credit facilities and delayed draw loans are generally considered to be derivatives for purposes of Rule 18f-4" and are subject to the rule's provisions." -> NOT TRUE! Using it as a conversational agent and sounding board for feedback on my GTD lists is more helpful. Particularly with GTD, I get "stuck" with my own initial phrasing of a project, desired outcome and next actions. Having a brainstorming session with chatGPT helped me get unstuck.
Context / where I am in my GTD journey: I have improved my capture processes to where inboxes are getting to near zero daily and zero weekly. I gather my commitments in Asana, with a project corresponding to an area of focus, for each client engagement. In client calls, we screenshare my Asana list so it has some value. Where I am challenged and the next level in my practice I want to get to is that the lists themselves repel me more than they attract because there is a lot of thinking that is still incomplete. They lack compelling statements of desired outcomes and the level of detail of next actions is patchy. Having someone to talk to, in short bursts, and on demand, can help me.
With code, I found it helpful for Excel VBA but none of my little projects were very complicated and I took time to go back and forth a few times. As a beginner, working with it as a (patient) tutor has helped me learn more. What I like is pasting in macros I find in others' workbooks and asking chatGPT to explain them to me. Then asking followup questions to make sure I get it.
I've found ChatGPT an absolute game changer to be honest, to the point where Im looking at reconfiguring my staff structure due to the time it saves. They're not joking when they say this could put people out of work. This kind of things really impresses me.
Write me a 3 day seminar on learning the basics of GTD for a group of 20 beginners
Day 1:
Topic: Introduction to GTD and Setting up your system
Duration: 8:00 am – 5:00 pm
8:00 am – 9:00 am: Welcome and coffee break
9:00 am – 9:30 am: Introduction to Getting Things Done (GTD) and its benefits
9:30 am – 10:30 am: Understanding the 5 stages of GTD
10:30 am – 11:30 am: Exercise: Identifying your areas of responsibility and capturing your commitments
11:30 am – 12:30 pm: Lunch break
12:30 pm – 1:30 pm: Collecting and processing your incoming information
1:30 pm – 3:00 pm: Organizing your system and creating lists
3:00 pm – 4:00 pm: Exercise: Setting up your GTD system using a tool or application of your choice
4:00 pm – 5:00 pm: Q&A and wrap-up
Day 2:
Topic: Reviewing and Mapping Out Your Goals and Projects
Duration: 8:00 am – 5:00 pm
8:00 am – 9:00 am: Coffee break and review of Day 1
9:00 am – 10:00 am: Understanding the importance of having clear goals and projects
10:00 am – 11:00 am: Exercise: Mapping out your long-term and short-term goals and projects
11:00 am – 12:00 pm: Lunch break
12:00 pm – 1:00 pm: Defining outcomes and next actions
1:00 pm – 2:30 pm: Organizing your projects and tasks
2:30 pm – 3:30 pm: Exercise: Applying the GTD method to a real-life project
3:30 pm – 4:30 pm: Dealing with procrastination and distractions
4:30 pm – 5:00 pm: Q&A and wrap-up
Day 3:
Topic: Mastering Your System
Duration: 8:00 am – 3:00 pm
8:00 am – 9:00 am: Coffee break and review of Day 2
9:00 am – 10:30 am: Reviewing and updating your system
10:30 am – 11:30 am: Exercise: Practicing your weekly review
11:30 am – 12:30 pm: Lunch break
12:30 pm – 2:00 pm: Advanced tips and strategies for GTD
2:00 pm – 3:00 pm: Q&A, feedback and wrap-up
Note: The exact timing of each topic and exercise may be adjusted to suit the group’s pace and needs.
For Day 1, give me a detailed breakdown of the topic areas
Sure, here's a more detailed breakdown of Day 1:
Topic: Introduction to GTD and Setting up your system
Duration: 8:00 am – 5:00 pm
8:00 am – 9:00 am: Welcome and coffee break
- In this time slot, the facilitator will greet the attendees, introduce themselves and explain the logistics and objectives of the seminar. The attendees, on the other hand, will have a chance to meet and network with each other.
9:00 am – 9:30 am: Introduction to Getting Things Done (GTD) and its benefits
- The facilitator will provide an overview of GTD, explain its principles and the benefits of implementing the GTD method in their lives.
9:30 am – 10:30 am: Understanding the 5 stages of GTD
- The facilitator will explain the five stages of GTD and how they work together to provide a holistic approach to productivity.
10:30 am – 11:30 am: Exercise: Identifying your areas of responsibility and capturing your commitments
- Attendees will be asked to identify their areas of responsibility and list down their commitments to each one of them. This exercise will help them understand the importance of capturing their commitments and how it fits in with the GTD system.
11:30 am – 12:30 pm: Lunch break
- Attendees will have a break to grab lunch and network with their peers.
12:30 pm – 1:30 pm: Collecting and processing your incoming information
- The facilitator will teach the attendees how to collect and process their incoming information to avoid feeling overwhelmed and stressed.
1:30 pm – 3:00 pm: Organizing your system and creating lists
- Attendees will learn how to organize their GTD system and create lists that fit their specific needs and preferences.
3:00 pm – 4:00 pm: Exercise: Setting up your GTD system using a tool or application of your choice
- Attendees will have the chance to practice setting up their GTD system using a tool or application of their choice. This exercise will allow them to learn and implement GTD in a way that makes sense to them personally.
4:00 pm – 5:00 pm: Q&A and wrap-up
- The facilitator will answer attendees' questions about what they have learned so far and provide a summary of the day's topics. They will also set the stage for Day 2 by previewing what attendees will be learning during the next day of the seminar.
<!-- endblock -->
For the Introduction session, what are the 5 most common questions people may ask?
There are various questions that attendees may ask during the Introduction session of a GTD seminar. Here are five common questions that people may ask:
1. What exactly is GTD, and what are its benefits?
Attendees may ask the facilitator to give a clear and concise explanation of GTD and outline its main advantages. They may want to know how GTD can help them become more productive and efficient in their personal and professional lives.
2. How does GTD differ from other productivity systems or methods?
People may be curious to know how GTD stacks up against other productivity systems that they may have tried or heard about. They may want to know what makes GTD unique and why it is worth learning.
3. Is GTD hard to implement, and what are the common challenges?
Attendees may be worried about the difficulty of implementing GTD in their daily lives, and they may want to know what the common pitfalls are and how to overcome them. They may also ask the facilitator for tips on how to stay motivated and on track.
4. Can GTD be customized to suit individual needs or job roles?
Attendees may be interested to know if GTD is a one-size-fits-all system or if it can be adapted to their specific needs and work environments. They may ask the facilitator for advice on how to customize GTD to suit their circumstances.
5. What are some success stories of people who have implemented GTD?
People may want to hear about real-life examples of people who have benefited from using GTD. They may want to know about success stories and the impact that GTD has had on people's productivity, stress levels, and overall quality of life.
<!-- endblock -->
Now that may not be the best GTD seminar you've ever seen devised, but given that I asked it 3 questions and went from a blank page to the above in under a minute, I don't think you can deny that's impressive.
Other things I've used it to create recently
- Outlined our next staff away day, including ice breakers, workshops, outcomes - even a decent halal lunch spread
- Created a Job Advert, Job Description, interview questions and presentation topic for a recent hire
- Drafted key metrics for job performance for an underperforming staff member
- Outlined our company's Health and Wellbeing Strategy
- Created a skills matrix for Board Performance
In each case I still had to edit and refine what it produced. You can't just copy and paste the lot. But its a damn site easier editing someone else's content than starting with a blank page of your own. In each case above, it took a piece of work that would take between 1 and 4 hours each, and got me there in under 20 minutes. I probably saved an entire day of work.
There's a few other interesting uses.
- When I take one to one notes with my staff, they're always pretty rough round the edges. I've found I can just write "Tidy up these notes" at the top of the notes, send the entire page of text to ChatGPT and it'll return them all tidied up and expanded out.
- It'll do Twitter campaigns, telling you all the Tweets you need to make on what date up to a launch
- It can summarise long emails or the exec summary of reports that arent exec enough
- It can make Microsoft Power Automate flows
It presents an odd challenge to my GTD system, since it isn't always obvious what will work well with AI and what won't. Its mostly replaced my @mindmap and @outline tasks tasks, since it excels at going from a blank page to an initial list of items very quickly, but there are other times when its not as clear cut. Some @research tasks for example fit well and some don't, its trial and error. So currently, I double tag, and have an AI tag as well as my traditional tag. Hopefully, I'll eventually get a handle on what works best and what doesn't/
I can't wait to have an AI GTD coach that can help review my system in a conversational format, then update my tools directly. So much of my struggle is the maintenance of the lists themselves. I've searched every day for a way to converse with ChatGPT as a GTD coach, then have it update lists, calendar appointments, and scrub emails with me. I'm convinced it will be a huge help
I largely agree with the sentiments here, though I am more optimistic about ChatGPT's output (Bing has already incorporated fact checking guardrails), it will continue improving.
I use an app called Scriptable on iOS that is basically a fully-featured JavaScript environment. Within that I've built a number of "apps" including a task manager that sits on top of iOS Reminders, which is my GTD brain.
Some uses of ChatGPT I've tested out:
Suggest a context list for a task (based on existing lists)
Suggest rephrasing of a task title to shorten and clarify
Generate subtasks for a task based on its title, description, and parent list
"Motivate me" - basically a prompt like "I've been procrastinating on task with title (x), motivate me to get started on it"
Related resources - as pointed out, the source data for ChatGPT is not current, and it can't (possibly an artificial limitation) connect to the internet. BUT it can generate search queries which can be converted to Google links!
Another big area that I want to explore is to (almost) fully automate the inbox processing process. Not to actually take actions on Inbox tasks, but to suggest them based on "knowledge" of GTD and my guidance. Especially if you're writing very clear task titles, this process is largely mechanical. ChatGPT should be capable of determining whether a task is already a next action or not (i.e. should it be moved to a NA list or is it a project), and it is very good at guessing the appropriate NA list.
