This is the right 'thing' at the right time for me. Thanks for this. Last fall, the platform I work on for my company had an "AI Days" thing that I attended and (having already been steeped in AI since 2021...) a lot of it was rehash - but the one takeaway that resonated with me is that with the pace of AI development, an Annual Review is no longer sufficient, and Quarterly is the way to go.
I have a related question. Within this context of personal planning and goal setting, do you use LLMs or other AI tools, and to what extent? I've recently been experimenting with custom Gems to assist me in prioritizing work and tasks based on my long-term goals, so I'm wondering whether you've been doing something similar. If so, do you find it helpful? Do you have any tips/tricks to share?
I do use LLM tools for a lot of the data collection and processing (plus vibe coding cool dashboard), and then also use LLMs as a second opinion sometimes, but I make sure to allocate a bunch of time for non-LLM reflection and intention setting as well.
This was something I was struggling with, until I dumped what I was struggling with AND the worksheet into Chat - and what came out the other side was...actually a pretty emotional read for me, and when I was discussing it with my wife (a Clinical Psychologist by trade...) this morning she was like, 'Huh. Chat's doing therapy without being asked now - and doing a decent job of it, too.'
Context: Step 2 (Reflect on what you actually want and why it is important.) smacked me in the face with a half-brick. I'm GenX (the embodiment of 'slack') and have ADHD, projects and goals like "write more, write better" have always come pretty easy, but WHY they're important? Oof.
Dumped the worksheet copy from my Drive into Chat, added that existential "I don't feel like anything I do is actually important or has any lasting value anyway..." and said "given everything you know about me, fill this worksheet out for me.
A: The worksheet is *goal-first* and assumes that you can actually name values, desired feelings, and outcomes. The right move here is not “pick better rocks,” but to build a scaffold whose job is discovering what is worth caring about, by running small, low-stakes experiments and measuring signal (energy, curiosity, relief, engagement) instead of meaning or legacy.
Rock 1: Rebuild energy and enthusiasm - one activity per week that produces noticeable engagement.
Rock 2: Curiosity Sampling - Curiosity is my strongest motivator, but it's currently starved.
Rock3: Prototype my Future - You can't choose a future you haven't touched. Do one exploratory action per month (short trips, deep research sprints - reconnaissance, not committment)
Anti-Rocks
1. NO LEGACY PROJECTS
2. No 'reinventing your life' narratives
3. No forcing 'purpose'.
OK... *that* is infinitely more impactful than any rocks I'd have picked like "TOGAF Certification" or something...
Thanks for sharing. Using LLMs, I constantly find valuable insights about topics I took for granted. I agree that it can allow some profound introspection, and I find it a good exercise to have an LLM interview me to help me refine and clear up my own thinking about myself.
Because of that, I'm curious about your experience: is there anything specific you do in your prompting to ensure the model forces you out of your comfort zone and challenges your existing perspective?
I spent a lot of time with GPT-4o when that was the frontier model, and it was infamous for just being overly agreeable (it would 'glaze the user' incessantly). I got in the habit of using the IT Security phrase "red team" to get the LLM (or even different LLMs as I copy/pasted back and forth) to be pretty adversarial if I use things like "red team this for accuracy and factfulness" or "red team this against best practice" or (especially with Chat's memory functions) "red team this against everything you know about me and tell me if it's workable (or makes sense, or is too ambitious for my limitations, etc.)
Looks great, definitely worth adapting - its' now on my schedule for next week.
But you're missing the step where you load all your recent emails and in-process papers into an LLM and have it fill out the form so you don't need to do any of the thinking or reflection, much less admitting what you might be doing poorly. ;)
(And then I can tell it to start executing on the new plans while I spend even more time on Twitter, of course.)
This is the right 'thing' at the right time for me. Thanks for this. Last fall, the platform I work on for my company had an "AI Days" thing that I attended and (having already been steeped in AI since 2021...) a lot of it was rehash - but the one takeaway that resonated with me is that with the pace of AI development, an Annual Review is no longer sufficient, and Quarterly is the way to go.
This will go a long way, I think...
Thanks for sharing, this is helpful!
I have a related question. Within this context of personal planning and goal setting, do you use LLMs or other AI tools, and to what extent? I've recently been experimenting with custom Gems to assist me in prioritizing work and tasks based on my long-term goals, so I'm wondering whether you've been doing something similar. If so, do you find it helpful? Do you have any tips/tricks to share?
I do use LLM tools for a lot of the data collection and processing (plus vibe coding cool dashboard), and then also use LLMs as a second opinion sometimes, but I make sure to allocate a bunch of time for non-LLM reflection and intention setting as well.
(Maybe Overshare alert :D )
This was something I was struggling with, until I dumped what I was struggling with AND the worksheet into Chat - and what came out the other side was...actually a pretty emotional read for me, and when I was discussing it with my wife (a Clinical Psychologist by trade...) this morning she was like, 'Huh. Chat's doing therapy without being asked now - and doing a decent job of it, too.'
Context: Step 2 (Reflect on what you actually want and why it is important.) smacked me in the face with a half-brick. I'm GenX (the embodiment of 'slack') and have ADHD, projects and goals like "write more, write better" have always come pretty easy, but WHY they're important? Oof.
Dumped the worksheet copy from my Drive into Chat, added that existential "I don't feel like anything I do is actually important or has any lasting value anyway..." and said "given everything you know about me, fill this worksheet out for me.
A: The worksheet is *goal-first* and assumes that you can actually name values, desired feelings, and outcomes. The right move here is not “pick better rocks,” but to build a scaffold whose job is discovering what is worth caring about, by running small, low-stakes experiments and measuring signal (energy, curiosity, relief, engagement) instead of meaning or legacy.
Rock 1: Rebuild energy and enthusiasm - one activity per week that produces noticeable engagement.
Rock 2: Curiosity Sampling - Curiosity is my strongest motivator, but it's currently starved.
Rock3: Prototype my Future - You can't choose a future you haven't touched. Do one exploratory action per month (short trips, deep research sprints - reconnaissance, not committment)
Anti-Rocks
1. NO LEGACY PROJECTS
2. No 'reinventing your life' narratives
3. No forcing 'purpose'.
OK... *that* is infinitely more impactful than any rocks I'd have picked like "TOGAF Certification" or something...
What a time to be alive, yeah?
Thanks for sharing. Using LLMs, I constantly find valuable insights about topics I took for granted. I agree that it can allow some profound introspection, and I find it a good exercise to have an LLM interview me to help me refine and clear up my own thinking about myself.
Because of that, I'm curious about your experience: is there anything specific you do in your prompting to ensure the model forces you out of your comfort zone and challenges your existing perspective?
I spent a lot of time with GPT-4o when that was the frontier model, and it was infamous for just being overly agreeable (it would 'glaze the user' incessantly). I got in the habit of using the IT Security phrase "red team" to get the LLM (or even different LLMs as I copy/pasted back and forth) to be pretty adversarial if I use things like "red team this for accuracy and factfulness" or "red team this against best practice" or (especially with Chat's memory functions) "red team this against everything you know about me and tell me if it's workable (or makes sense, or is too ambitious for my limitations, etc.)
Looks great, definitely worth adapting - its' now on my schedule for next week.
But you're missing the step where you load all your recent emails and in-process papers into an LLM and have it fill out the form so you don't need to do any of the thinking or reflection, much less admitting what you might be doing poorly. ;)
(And then I can tell it to start executing on the new plans while I spend even more time on Twitter, of course.)
Thanks, that looks quite good, let's see if it can help me overcome my magpie syndrome 😁
This looks really useful; I’ll give it a go and report back!
Please do!
Thank you, helpful.