sia.hackernoon.com

how we put ai on like another jacket, threw a course on top, and ended up with burnout 2.0

for the last two years you’ve been sold the same dream on repeat:

“plug in ChatGPT / Cursor / a dozen prompts — and you’ll finally breathe.”

the punchline for a lot of people is the opposite:

one more task system,
one more chat,
one more course on “how to prompt correctly,”
and that sticky feeling that you’ve got less energy and more chaos.

“i’m doing everything right: watching webinars, saving ‘100 prompts for productivity,’ running tasks through ai… and by evening it still feels like i didn’t move anything forward.”

sound familiar?

in this piece we’re offering a different diagnosis:

often you’re not burned out.
you’ve got context obesity.

and yeah — ai, courses, and “magic prompts” are basically tossing more firewood on the pile.

why you can trust this text

we’re the team behind AI Mindset and {context} lab. for the last two years we’ve been running labs and communities for people who build an ai stack and a personal operating system — not just “mess around with prompts.”

more than 700 people have gone through our formats already: product folks, developers, hr, consultants, founders.

from the latest lab:

60+ participants in a single cohort,
15 deep interviews about their experience,
segmentation into four types of people (from “tech skeptics” to founders).

and they all share the same plot: people come in “for ai tools” — and hit a wall where their brain can’t handle the context configuration they’re living in.

this article is our attempt to describe that layer honestly.

classic burnout vs context obesity

classic burnout is about depletion.
context obesity is about overloaded working memory.

if we simplify hard: your head has a limited “context window.” only a few items fit in there at once. when those items get closed — the system chills out. when they don’t — you drag tail-ends behind you.

psychology has been pointing at a couple key effects for a while:

the zeigarnik effect — unfinished tasks stick in memory and hum louder than finished ones (read more);
attention residue — part of your brain stays in the previous task even after you switch (sophie leroy’s research).

in a world where you have one project and one notebook, it’s manageable. but when you’ve got:

a work project,
a pet project,
ai experiments,
side education,
several active chats with colleagues and communities,

those tails start eating your ram.

what burnout looks like

classic burnout:

exhaustion, cynicism, lower performance;
“treated” with vacation, job change, therapy;
emotional + physical “i can’t do this anymore.”

context obesity:

the feeling of a stuffed head and a “second job” maintaining systems;
treated by rebuilding your context configuration, not by one more vacation;
about the architecture of tasks, tools, and expectations — not just hours in the office.

these two can overlap. but a lot of people who self-diagnose burnout are actually living in chronic context overload.

what a day looks like with context obesity

let’s call them Alex, a product manager in a big company. but feel free to swap in yourself.

you’re not burned out, you’ve got context obesity

09:10 Alex opens the laptop and, on autopilot, boots up:

corporate messenger,
email,
notion,
ChatGPT.

somewhere in the background there’s a thought looping: “i should finish that experiment concept we came up with last week.”

10:30 on a call they discuss a new hypothesis. someone drops a link: “10 ways to use ai in analytics.” Alex opens it, scrolls, saves it to “read later.” meanwhile, a personal todo gets a new entry: “think about an ai assistant for reports.”

12:15 Alex decides to “speed up” and asks ChatGPT to draft an email to a client. the text is fine, but:

facts need checking,
it needs to be translated into company language,
it needs to match the old thread.

so instead of 15 minutes for an email — it’s 25 minutes of review.

14:40 sprint planning time. Alex has:

a backlog in jira,
product ideas in personal obsidian,
notes from the last retro in notion.

some tasks are duplicated, some contradict each other. Alex promises: “tonight i’ll consolidate everything into one system.”

18:30 Alex closes the laptop with this feeling:

“i was spinning all day, but not a single track feels finished.”

this isn’t about Alex “doing nothing.” it’s about nothing closing a loop: not the email, not the experiments, not the task system.

in interviews, our lab participants described it almost word for word:

“my notes and notion turned into a second job.”
“i open my system and realize i don’t understand my own system.”

and almost everyone showed up with: “i think i’m burned out.”

why “100 prompts for productivity” makes it worse

you’re not burned out, you’ve got context obesity

the industry responds to this state like this:

“here are 50 prompts that’ll make you a super-manager,”
“ChatGPT course: 3x your speed,”
“5 tools that will replace your assistant.”

what happens when someone in Alex-mode walks into that content?

every prompt list = another layer of expectations and guilt. now you’re not only behind on projects, you’re also “not using ai to the max.”
every prompt creates new context. you generate dozens of drafts, ideas, lists — and all of them demand review and decisions. congrats, you’ve just opened new unfinished loops.
ai doesn’t plug into a system — it plugs into chaos. if you don’t have clear architecture for tasks and meaning, ai simply increases the amount of text your brain has to chew.

people said it like this in interviews:

“i used ai every day, but it felt like another layer of work. the only thing that helped was when we started discussing the architecture of context — not just touching tools.”

so yeah: “100 prompts” isn’t medicine. it’s a new set of calories for an already overloaded context.

ai as a context-obesity amplifier

ai tools amplify the system they land in.

if you’ve got a clean process and sane load — they really do remove routine.

if your task configuration is already chaotic — they amplify the chaos.

you’re not burned out, you’ve got context obesity

there’s a useful term here: the “hidden tax of context switching.” when you jump between tasks and apps, your brain pays a reboot cost. across different reviews, estimates say it can eat up to 20–40% of work time.

some breakdowns here:

add research on technostress — digital stress from too many tools and notifications. meta-reviews in recent years show that introducing digital tech without changing processes often increases stress and burnout risk at first, not lowers it (the review, and a popular retelling).

now let’s look at three layers of taxes in the “you + ai” bundle.

1/ integration tax

every new service means:

figuring out how it works,
fitting it into your current stack,
designing how you won’t drown in its notifications.

in our interviews, many participants spent the first weeks not solving tasks, but maintaining the tool zoo.

2/ cognitive tax

ai writes — you review. the responsibility for meaning is still on you:

check it didn’t hallucinate,
adapt it to context,
decide what to do with it.

that’s another stream clogging your working memory.

3/ the “the system will save me” illusion

especially nasty when ai lands on top of an existing graveyard: notion, obsidian, multiple todo apps.

instead of admitting “my system can’t hold reality,” we slap one more smart layer on top.

if the foundation is rotten, no ai penthouse will save it.

why vacation or a job change doesn’t fix it

you’re not burned out, you’ve got context obesity

when you say “i’m burned out,” you usually get advice like:

take a vacation,
change jobs,
delete social media,
“start from a clean slate.”

all of that can be fine, with one catch:

you bring the same context habits with you.

on vacation your brain keeps looping:

unclosed promises,
hanging projects,
decisions you didn’t make.

a job change swaps the set of tasks — but doesn’t change how you:

open too many fronts,
keep everything in your head,
are afraid to admit something isn’t important and let it go.

a new notion workspace without new logic turns into the same landfill in a month.

this isn’t willpower.
this is missing an explicit system prompt for yourself.

humans need a system prompt too

large language models have a system prompt — a text that defines:

who it is,
what style it answers in,
what it won’t do,
what its default goals are.

you’re not burned out, you’ve got context obesity

most people don’t have that explicit layer. every morning we boot life in a mode like:

“take everything that lands,”
“don’t say no to tasks,”
“reply immediately so you don’t feel guilty.”

at one lecture on context, a participant said:

“when i realized my head has a context window too, i stopped shoving everything in there.”

that’s basically the first sketch of a personal system prompt:

what’s allowed for me right now — and what isn’t;
what tasks don’t get to enter my ram at all;
what rules i set for myself and for my ai assistants.

how to work with context obesity: three levels

spoiler: you don’t need to “burn it all down and go back to a paper notebook.”

you’re not burned out, you’ve got context obesity

you need to rebuild your context configuration honestly.

in our labs, three mandatory layers show up almost every time.

1/ capture: get the context out of your head

as long as your head is the only database, you’re doomed.

exercise: “15 unfinished loops”

open a note or grab a piece of paper.
for 7–10 minutes, write down everything that responds to “i should…”: “send the documents,” “talk to that colleague,” “figure out this service.”
don’t sort or judge. the job is to offload.

people usually freak out at the volume. one participant said after doing it:

“i realized i’m not burned out — i’ve got a whole jira board open in my head.”

the simple fact these things moved from brain to text already brings relief.

2/ structure: give it a skeleton

after the offload, you don’t need to “optimize everything.” you need to format the chaos.

frameworks help here — not as religion, as a grid:

the wheel of life — see where you’re objectively overdoing it (work, study) and where you’re crashing (health, relationships). breakdown, for example: https://quenza.com/blog/wheel-of-life-assessment/
dilts levels — realize you’re fixing a skill when the problem is environment/values. description: https://www.skillsyouneed.com/lead/logical-levels.html
yearly reviews like yearcompass — turn a chaotic year into a coherent story (site: https://yearcompass.com/).
weekly cycles (for us it’s the arc method: appreciate → reflect → create) — end the week not with “i survived,” but with a set of concrete artifacts.

micro-format: burn / delegate / keep

once a week:

open your list of unfinished loops.
for each item, mark it honestly:
- burn — i won’t do it, and i’m brave enough to admit it.
- delegate — a person or ai can do it (drafts, data prep).
- keep — it’s truly important and belongs to the next few weeks.
everything you “keep” goes into one system (any system) and stops living in your head.

often at this step the list shrinks 2–3x — and the guilt shrinks with it.

3/ your personal system prompt 2026

the last layer is turning all of this into explicit rules.

exercise: three rules for my brain-llm

imagine your brain is an llm and you can write its system prompt. answer in writing:

what three rules do i put into my system prompt for 2026?

it might look like this:

don’t hold more than three active projects in my head. everything else goes to the backlog with an honest “not now” date.
don’t plug in a new ai tool until i’ve lived in the current configuration for a full week. no new shiny toy until the old stuff is routine.
every friday, close at least one loop: do / delegate / burn. so the week ends with completion, not hanging tabs.

then:

run your current tasks through these three rules;
see what falls off automatically, what needs to be handed off, and what actually stays yours.

for many lab participants this step alone changes their week configuration — without adding a single new service.

where the context lab came from

you’re not burned out, you’ve got context obesity

context lab was born from a painfully grounded question:

why do smart, motivated people with “good systems” in notion and obsidian still feel like Alex from this article?

we started gathering small groups and testing a hypothesis: if you give people language for context, offloading practices, and basic frameworks — does their “i’m not getting anything done” feeling change?

interviews show that this exact layer — understanding context and working with it regularly — is what people most often name as their main “click.”

context lab is a place where you don’t get another set of “secret prompts.” you build your context stack and your system prompt — and only then hang ai assistants and automations onto it.

second article in the series:

the productivity museum: why notion, ai, and task courses only make it harder to live

and in the third we’ll share a personal story of someone who drowned in obsidian, crawled out through their own system prompt, and along the way helped invent the context lab formats.

Ray Svitla
stay evolving 🐌