So, Herman, I was looking at my phone the other day—I mean, really looking at it, not just scrolling—and I realized I have about fourteen different lists going. It is honestly embarrassing. I have a grocery list in my Notes app, a list of books I want to read in a dedicated reader app, a list of bugs I need to fix for a side project on a Trello board, and then the actual, official work to-do list in Outlook. It is a total mess. It is like having fourteen different people shouting at you at once, all of them claiming their thing is the most important, and all of them standing in different rooms of the house.
Herman Poppleberry here, and Corn, what you are describing is the classic executive function bottleneck. It is a phenomenon we see more and more as our digital lives fragment. It is not that you do not know what needs to be done. You are clearly great at the capture phase. The problem is that the cognitive cost of deciding which one to do first is sometimes higher than the cost of the task itself. Your brain is spending all its energy on the "sorting" processor, leaving nothing left for the "doing" processor.
It is that "paralysis by analysis" feeling. And today’s prompt from Daniel is about exactly that struggle. Daniel is asking about AI technologies that can help sort and triage a to-do list. He wants a tool that can take a massive, messy list, categorize the items, set priorities based on actual logic, and assign deadlines automatically. And he also asked for an essential AI stack for someone with ADHD to manage time and flow. This feels like the holy grail of productivity right now, especially as we move into early twenty twenty-six and these agents are getting more capable.
This is such a timely question because the technology has really shifted in the last eighteen months from simple list-keeping to what we call agentic reasoning. We have finally moved past the era of just having a digital piece of paper that happens to live on a screen. We are now firmly in the era of the cognitive assistant. Daniel mentioned he is good at the capture part—closing those loops like we talked about in previous episodes—but the sifting is where he hits the wall. That is very common for the ADHD brain, where the "importance" filter is often replaced by an "interest" filter or an "urgency" filter.
I remember in episode eight sixty-five, we talked about task drift and how the brain’s internal management system can just get overwhelmed. When you have fifty items on a list, your brain tries to process all fifty at once. It is exhausting. It is like trying to read a fifty-page book where every single sentence is written in bold, red letters. So, Herman, as of February twenty-seven, twenty twenty-six, what is the state of the art? Are there tools that actually do this triage work for us, or are we still just playing with fancy spreadsheets?
We are definitely beyond spreadsheets now. There are a few tools that are getting very close to Daniel's dream. You have platforms like Motion and Reclaim that have been around for a few years, but in the last twelve months, they have integrated much deeper Large Language Model capabilities. They are not just looking at a deadline you typed in anymore. They are looking at your historical velocity. If you usually take three hours to write a report, but you scheduled one hour, the AI will actually flag that and say, "Hey, based on your last four reports, this is not going to fit." It is acting as a reality check against our own optimism.
That is a huge shift because it is moving from reactive to predictive. But what about the categorization part? That is where I struggle. If I just dump twenty things into a voice memo while I am driving, can an AI actually understand the difference between a high-stakes work task and a low-stakes personal errand without me having to manually tag everything?
It can, and it does it through a process called semantic embedding. Most of these modern triage tools use a vector-based approach. They do not just look for keywords like "buy" or "call." They look at the meaning and the weight of the words. If you say, "I need to call the insurance company about the roof leak," and then you say, "I need to buy milk," the AI understands the context. One is a home maintenance emergency with financial and structural implications—that goes into the "High Priority/Home" bucket. The other is a routine domestic errand. It can categorize them into different contexts automatically because it has been trained on millions of examples of how humans prioritize their lives.
So, if Daniel is looking for something that exists right now, would you say a combination of a robust scheduler and a large language model is the way to go? Or is there a specific app that finally nails all of it in one package?
There is not one perfect "unicorn" app that does everything perfectly yet, but we are getting incredibly close. Trevor AI is another good one that focuses on time-blocking and has a very intuitive interface for ADHD users. But the real power comes when you use something like an LLM wrapper that connects to your task manager. For example, if you use Todoist or Things, you can now use custom GPTs or Claude Projects that pull your entire list via API. You can tell the AI, "Sort this based on the Eisenhower Matrix," which is that classic four-quadrant system of urgent versus important. It can then push those sorted tasks back to your calendar as scheduled blocks.
I love the Eisenhower Matrix in theory, but the problem for many people, especially those with ADHD, is that everything feels urgent and everything feels important. It is like a crowded train station where everyone is screaming for the same train at the same time. How does the AI decide what actually matters if the user is not clear on it?
That is where the historical data and the external context come in. An advanced agentic system can see your calendar. It knows you have a meeting with your boss on Thursday at ten a.m. If you have a task called "prepare slides," and it is Wednesday afternoon, the AI can infer that this is now the highest priority task because of the external hard deadline. It is doing the "connective tissue" work that our brains sometimes struggle to do when we are feeling overwhelmed. It is looking at the "if-then" logic of your entire life, not just the single task.
Okay, let us pivot to the second part of Daniel’s prompt. He said if this doesn't exist perfectly yet, how would we build it? If we were going to sit down and architect the ultimate triage agent for someone who needs that heavy-duty executive function support, what would that architecture look like?
Oh, this is a fun thought experiment. First, I would start with a multi-modal capture layer. It needs to be frictionless. If you have to open an app, click a button, and type, you have already lost the battle. It should be a persistent voice listener—maybe a wearable or a very fast Whisper-based shortcut on your watch. Just a stream-of-consciousness dump. No formatting required.
Right, like the "open loops" Daniel mentioned. You just say it and it is gone from your head. But then what happens to that raw, messy data?
The second layer is the semantic parser. This would use a high-reasoning model like GPT-four-o or Claude three point five Sonnet to take that raw transcript and break it into discrete tasks. But here is the secret sauce: it needs a "Personal Context Window." It should have read-only access to your email, your calendar, and even your previous week’s productivity data. It needs to know what "done" looks like for you. If you say "fix the thing," it should know from your emails that "the thing" is the broken dishwasher you have been complaining about for three days.
So, it is not just identifying the task. It is identifying the effort and the history.
I would build a scoring system. Every task gets a score based on three factors: urgency, which is the hard deadline; impact, which is how much this moves the needle on your long-term goals; and cognitive load, which is how much brainpower it requires. We know from episode four fifty-nine, when we talked about autonomous scheduling, that people with ADHD often have energy fluctuations. You do not want to schedule a high-load task for four p.m. on a Friday if you know your focus usually dips then.
That is brilliant. So the AI would say, "Hey Daniel, I see you have this complex coding task. I am putting it on Tuesday morning because that is when you are usually most productive, and I am putting the laundry and the quick emails in the afternoon slump."
Precisely. And the third layer would be the "Nudge Engine." It is not enough to just have a list. The AI needs to be a "body double." It should check in via a small notification or a voice prompt and say, "You are supposed to be working on the project proposal right now. Are you doing that, or are you on Wikipedia looking up the history of salt?" It needs to be a gentle, non-judgmental accountability partner.
I think the non-judgmental part is huge. Daniel mentioned the cycle of shame that often comes with missed deadlines. An AI doesn't get frustrated with you. It doesn't roll its eyes if you push a task for the fifth time. It just recalculates the path. It is like a GPS for your life—if you miss a turn, it doesn't scream at you; it just says "recalculating."
It is just math to the AI. It says, "Okay, we missed the window, here is the new optimal path." It removes the emotional weight of failure. That is a massive psychological benefit for anyone, but especially for those with ADHD who have spent a lifetime feeling like they are "behind."
So if we were building this, we would need a persistent database—probably a vector database like Pinecone for the long-term context—an LLM for the reasoning, and a tight integration with common APIs like Google Calendar or Outlook. It is actually very buildable right now. The biggest hurdle is probably the latency and the battery life of a persistent listener, but with dedicated hardware like the newer AI pins or even just optimized smartphone chips, that is becoming less of an issue.
It really is. And you could even include a feature where it breaks down a big task into smaller steps automatically. We call this "chunking." If you put in "write a book," the AI immediately expands that into a sub-list of fifty smaller, manageable steps like "outline chapter one" or "research character names." That prevents the "Wall of Awful" we discussed in episode four fifty-nine, where a task is so big you just cannot find a way to start.
Okay, let's move into the practical advice for Daniel and everyone else. Daniel asked for a stack of three or four essential AI tools for someone with ADHD. This is like the survival kit for the modern brain. Herman, what is your first pick?
My first pick is a high-quality voice capture tool. I am currently really impressed with tools like Oasis or Otter, but honestly, even just using the native Dictation on a phone with a custom shortcut that sends the text to an LLM is powerful. There is also a tool called "Goblin dot tools" which has a feature called "Magic To-Do." It is specifically designed for neurodivergent brains. You type in a task, and it uses AI to break it down into tiny, non-intimidating steps. It even has a "spiciness" level for how much help you need breaking it down. It is a game-changer for getting over that initial hurdle of starting.
I love Goblin dot tools. The "Judge" feature is also great for checking the tone of emails. My second pick for the stack would be a smart scheduler like Motion. I know we mentioned it earlier, but for someone with ADHD, the fact that it automatically reshuffles your entire day when a meeting runs over is a life-saver. Traditional calendars are static. If you miss a ten a.m. slot, the whole day feels ruined. Motion just slides everything down and says, "No worries, we will start at eleven." It removes the "all or nothing" thinking that often leads to a total productivity collapse.
That flexibility is key. My third pick would be an AI-powered focus tool. There is a tool called Opal that uses AI to block distracting apps, but it goes deeper by categorizing your focus. But I would also suggest looking into AI body doubling platforms like Focusmate, which is now integrating AI partners. If you cannot find a human to work with, you can hop into a virtual room with an AI avatar that just sits there with you while you work. It sounds strange, but the psychological effect of "being watched" or "working alongside" someone is a proven way to increase dopamine and stay on task for ADHD brains.
That is fascinating. It is like a digital supervisor that you actually like.
And for the fourth tool, I have to go with something that handles information triage. We talked about this in episode five fifty-three, the SITREP method. When you have ADHD, the sheer volume of incoming emails, Slack messages, and news can lead to total paralysis. You need an AI that can summarize your morning. I recommend "Shortwave" for email. It uses AI to bundle your emails, summarize long threads, and even draft replies in your voice. It looks at your inbox and says, "Here are the three things you actually need to care about today, and here is the fluff you can ignore." It reduces the "noise" before you even start your day.
So the stack is: a frictionless voice capture tool like Oasis or Goblin dot tools, a dynamic AI scheduler like Motion, a focus guardian like Opal or an AI body double, and an intelligent briefing tool like Shortwave. That covers the whole pipeline from getting an idea, to planning it, to executing it, to managing the noise around it.
It is about building an external scaffolding for the brain. If your internal executive function is struggling, you outsource those specific processes to specialized AI agents. You are still the pilot, but the AI is the flight computer. It handles the millions of tiny adjustments needed to stay on course while you focus on the big picture.
I like that. It really changes the relationship with the to-do list. It goes from being a source of stress to being a supportive partner. Daniel mentioned that he is technically literate, so he could probably even set up some of these automations himself using tools like Zapier or Make to connect an LLM to his task manager.
Oh, absolutely. If Daniel wanted to go the DIY route, he could set up a system where a voice memo in his phone triggers a script. That script sends the transcript to GPT-four-o with a prompt like, "Extract the tasks, estimate the duration based on my past history, and find the first available slot in my Google Calendar that isn't a focused deep-work block." That is entirely possible with today's APIs. You can essentially build a custom "Daniel-Bot" that knows exactly how you work.
And that brings us back to the idea of the personal context. The more the AI knows about you, the better it can triage. It is like a kitchen’s mise en place. Everything is prepped and in its place before the cooking starts. The AI is doing the chopping and the measuring so that when you sit down to work, you can just start "cooking" the actual project.
That is a great way to put it. The goal is to reduce the "activation energy" required for any given task. If the AI has already prioritized the list and broken the first task into three small steps, your brain sees that and thinks, "Oh, I can do that. That is easy." It bypasses the fear and the overwhelm. We are essentially using AI to hack our own dopamine systems.
We have talked about this in several episodes, but it feels like we are reaching a tipping point where these tools are becoming truly invisible. They are just part of the background of our lives. In episode eight seventy, we discussed AI-driven decision apps for emergencies, and this is really the same logic applied to daily life. It is about reducing the cognitive load during high-stress or high-complexity moments.
It really is. And for someone with ADHD, every day can feel like a series of high-complexity moments because the brain is constantly trying to sort through the noise. Using AI to filter that noise is not just a productivity hack; it is a mental health tool. It reduces that shame and anxiety Daniel was talking about. When you have a system you trust, you stop beating yourself up for forgetting things.
I think that is a really important point. This is not just about doing more work. It is about feeling better while you do the work you choose to do. It is about reclaiming your time and your peace of mind. It is about moving from "surviving the day" to actually "designing the day."
Well said, Corn. I think Daniel has a lot of great options here, whether he uses off-the-shelf tools like Motion and Shortwave or decides to build a custom solution using the architecture we discussed. The technology is finally catching up to the needs of the neurodivergent community. We are seeing a shift from "one size fits all" software to "one size fits you" AI.
It is exciting to see. And honestly, even for people without ADHD, these tools are becoming essential. The world is just getting too fast and too loud for the unassisted human brain to handle alone. We all need a little bit of help with our triage these days.
We are all becoming a bit more reliant on our digital exoskeletons, aren't we? It is the evolution of the "Second Brain" concept that Tiago Forte popularized, but now that second brain has its own prefrontal cortex.
We definitely are. Before we wrap up, I want to remind everyone that if you are interested in the deeper mechanics of how the brain handles tasks, you should definitely check out episode eight sixty-five on task drift. It provides a lot of the scientific foundation for why these AI tools are so effective.
And if you want to see how we apply this to physical objects, episode seven eighty-six on AI-powered inventory management is a fun look at the same principles applied to the physical world. It is the same logic, just with boxes instead of tasks.
Great suggestions. Well, this has been a fascinating deep dive. Daniel, thank you for the prompt. It really forced us to look at the intersection of technology and psychology in a very practical way. I hope this helps you find some clarity in the chaos.
Yes, thanks Daniel. And if you are out there listening and you have found a specific AI tool that has changed your workflow, especially if you have ADHD, we would love to hear about it. We are always looking for new tools to test in the lab.
Definitely. You can reach the show at show at my weird prompts dot com. And if you are enjoying these conversations, we would really appreciate it if you could leave us a quick review on your podcast app or Spotify. It genuinely helps other people find the show and keeps us going.
It really does make a difference. You can find all our past episodes and a contact form at my weird prompts dot com. We have a huge archive of topics like this, ranging from AI ethics to DIY bio-hacking.
And just a reminder, our show music is generated with Suno. It is pretty amazing what you can do with these generative tools these days. We are living in the future, folks.
It really is. Well, I think that is it for today. I have a to-do list that needs some serious AI triage myself.
Thanks for listening to My Weird Prompts. I am Corn.
And I am Herman Poppleberry.
We will see you next time.
Goodbye everyone!
So, Herman, I was thinking about that nudge engine you mentioned. Do you think there is a risk of it becoming too annoying? Like, if the AI is constantly telling you what to do, don't you eventually just start ignoring it? It is like the "Boy Who Cried Wolf" but with productivity.
That is a classic problem in human-computer interaction called alert fatigue. The key is that the AI has to be smart enough to know when to shut up. This is where "Contextual Awareness" comes in. If it sees you are actually in a flow state—maybe your typing speed is high and you haven't switched tabs in twenty minutes—even if you are not doing exactly what was scheduled, it should leave you alone. The worst thing an AI can do is interrupt a genuine moment of productivity to tell you that you should be productive.
Right, it needs to be able to "read the room." Or read the brainwaves, eventually.
We are not quite there yet for the mass market, but with biometric integration—like heart rate variability from your watch or even eye-tracking from your webcam—the AI could start to infer your cognitive state. It could see that your heart rate is up and your eyes are darting around, and it might realize you are stressed and need a break, rather than another task. It could suggest a three-minute breathing exercise instead of showing you your next deadline.
That is the level of support that would really change the game. It is moving from a task manager to a wellness manager. It is looking at the human behind the list, not just the list itself.
It is about holistic support. We are moving away from the "grind culture" of the early two thousands and toward a more sustainable, AI-assisted way of living.
Well, on that note, I think I am going to go listen to some of those voice memos I have been ignoring and see if I can get an AI to make sense of them. I think I have a few "open loops" that are starting to feel like "open wounds."
Good luck with that, Corn. Let me know if you find anything life-changing in there. Or at least something that explains why you have fourteen different lists.
I will. Probably just a lot of reminders to buy more coffee and fix that one bug in the CSS that has been haunting me since twenty twenty-four.
A noble goal. Coffee is the fuel of the modern mind, after all.
Truly. Alright, take care everyone. We will be back next week with more weird prompts and AI deep dives.
See you!
I was just thinking about the categorization piece again. You know how some tasks are not really tasks, but they are more like projects? Like, "plan the summer vacation" is not a single thing you can do. It is a hundred things.
Right, that is the difference between a horizontal list and a vertical project. Most people fail because they put horizontal items on a vertical list. They put "Write Novel" right next to "Buy Bread." One takes ten minutes, the other takes ten months.
And I think the real power of the next generation of AI will be the ability to automatically convert those vague, horizontal goals into vertical, actionable plans. It is like having a project manager who works for you for free, twenty-four hours a day.
And it never gets tired of your questions or your procrastination.
That is the best part. No matter how many times I ask, "What should I do next?" it will always have an answer that is based on logic, not emotion.
And hopefully, it is the right answer.
That is the goal. Alright, for real this time, thanks for listening. We have a lot of exciting stuff coming up in the next few episodes, including a look at AI in the kitchen.
Bye!
One last thing, Herman. Did you see that new paper on using recursive reward modeling for personal agents? It came out just last week.
I did! It is fascinating. The idea that the AI can learn your preferences by watching which of its suggestions you actually follow and which ones you ignore. It is essentially fine-tuning itself in real-time on your behavior. It is like a tailor making a suit while you are wearing it.
That is exactly what we were talking about with the nudge engine. If the AI sees you always ignore the laundry at two p.m., it will stop suggesting it then. It learns your resistance patterns and your "path of least resistance."
It is almost like it is learning your personality. Which is a little bit creepy if you think about it too long, but also incredibly useful for someone who needs that high level of personalization.
It is a fine line, as always, between "helpful assistant" and "creepy stalker." But for someone struggling with executive function, that kind of deep personalization is the dream. It is the only way to build a system that actually sticks.
It really is. Okay, I think we have truly covered it all now. My own internal "done" sensor is finally pinging.
I think so too. Thanks again, Daniel, for such a thought-provoking prompt. It is one of our favorites so far this year.
Take care, everyone. Stay productive, or at least, stay kind to yourselves while you try to be.
See you in the next one.
Bye!
Hey, wait, Herman, before we go, I just remembered. Daniel mentioned in his email that his son Ezra was born recently. We should probably mention that an AI stack is even more important when you have a newborn in the house. The sleep deprivation alone is enough to wreck anyone's executive function, ADHD or not.
Oh, absolutely! Congratulations, Daniel! When you are sleep-deprived, your prefrontal cortex is basically offline. You are living entirely in your lizard brain. Having an AI to remember when the last feeding was or when the next doctor's appointment is, is not just a luxury, it is a survival strategy. It is like having a backup brain when your primary one is running on two hours of sleep.
I can only imagine the chaos. Daniel, if you are listening, we hope you and Hannah and little Ezra are doing well. And hopefully, some of these tools can give you a few extra minutes of sleep or at least take the "mental load" off your plate.
Or at least make the minutes you are awake a little less chaotic. Use those voice memos, Daniel! Don't try to remember anything right now. Let the AI do the heavy lifting.
Alright, now we are really done. For real. No more "one last things."
Goodbye!
Bye everyone!
See you!
Peace.
Cheers.
Alright, I am hitting the stop button now. I can see the waveform stopping.
About time. We are way over our usual limit, but it was worth it.
Just making sure we covered everything! I didn't want to leave Daniel hanging on the architecture side.
You always do. You are the king of the "one more thing."
Guilty as charged. It is my own brand of task drift, I suppose.
See you later, brother. Go get that coffee.
On it. See you, Herman.
Bye!
Bye!
