Lift · The Guiding Principles Join the waitlist
A woman at her desk looks up at a wall of paper cards joined by threads. Some places hold only the outline of a card. One mustard thread climbs through the cards to a single card above the rest.

No agent in the world knows what you know. Lift does.

You keep Claude Code or Codex and delegate the work. Lift is designed to help you build the understanding you need to reach your goal, starting from what you have already shown you know.

Join the waitlist Read why

Delegate the work. Understand what matters for your goal.

You don't need to do everything yourself,

and you don't need to learn every detail of what you hand over.

Even mathematicians work with theorems they have never proven themselves.

You do need enough understanding

  • to see the big picture,
  • to choose the direction,
  • to spot a problem
  • and to know when to ask.

You can look into the detail whenever you want to.

As models get smarter, knowing every detail matters less.

What still matters is knowing the things your goal needs, so that you become capable of reaching it.

Today your agents don't know what you understand.

  1. Each one starts again from zero,
  2. so you explain yourself again
  3. and get answers above or below your level.
  4. Then the decisions that matter land on you unprepared.

An example

What it is like

  1. 01

    Say you are building your first paid service. Your agent has made the payment page. Now you have to choose a price.

  2. 02

    Lift connects that decision to what you need to understand: what each sale costs and what money is left afterwards.

    On your day's page you see the ideas that matter for your work, explained from what you already understand.

  3. 03

    When Lift is unsure, it asks you to explain in your own words, such as what your price needs to cover.

    Your answer tells it what still needs explaining.

  4. 04

    Your agent handles the code. You make the choice, and you know why.

    You can go into the details when you want to.

Concept view. Lift is in development.

How it helps you

The aim is that every agent starts from what you understand.

No need to explain yourself again, and answers at your level.

How should I price my wedding cakes?

Start with what one portion costs you: the ingredients, your time and the packing. Together that is your cost per sale, €5.00. Your price has to be higher. What is left is your margin, the money you keep. At €7.50 a portion, €2.50 stays with you. You can change it later.

Work out your cost per sale: ingredients, time and packing, €5.00 a portion. Then choose a price that leaves a healthy margin. At €7.50, about a third stays with you. Adjust it once the first orders are in.

€7.50 a portion: €5.00 cost per sale and a margin of about a third. Revisit after the first orders.

2 words explained in this answer

Concept view with example answers. Move the control.

See your own record:

what you have shown you can do, the evidence behind it and where Lift is still unsure.

No goal yet?

It is meant to help you discover one.

Share what you did and how you reached a goal,

with the proof behind it.

See your next step.

A checked map of what each field needs and your goal are set to show what to learn and do next, for the year, the month and the day.

It is also meant to work as a connector inside the apps you already use,

in the background: no switching apps.

You could build your own system.

You would need

  • to keep the record current,
  • connect it to your goals
  • and check whether it reflects what you understand.

Lift is designed to do those jobs together, and update the record and the plan as your work changes.

What is built today

  • A tutor is built and being refined. It can grade an explanation in your own words against a checked card.
  • A morning page prepares the day's concepts with a map and short explanations.
  • On the pages Lift makes, a concept word can be marked "I know it" or "I don't", and Lift keeps that.

Want to use it?

In development. Not yet available. Join the waitlist to hear when you can try Lift, or book a 30-minute call. First for Claude Code and Codex users, solo builders and students.

Prefer e-mail? [email protected]

Mia at her kitchen table in the evening, working at a laptop; beside it stands a half-decorated three-tier wedding cake.

One person's story

Picture Mia

She works at a laptop all day, in a job that AI is changing, and she dreams of her own wedding-cake studio.

She bakes well, but she knows little about costing, suppliers or running a shop, and her job fills her day.

The small orb at her side is Lift, the guide that takes her through each step.

AI is changing work and life faster than most people can adapt.

To keep up, people need a deeper understanding of the world, of themselves and of others. Lift exists to give them that understanding, and the practical help to act on it.

  1. 1

    At its root is an engine meant to keep what you say and do while you work in Lift, sorted so that your agents know what you know.

  2. 2

    On that foundation, a concept engine measures what you know and can do.

    Lessons pass from agent to agent, and from person to person.

  3. 3

    Lift does not remove failure, because failure is part of learning.

    Its job is to turn each failure into a lesson you can reuse: your guiding principles.

  4. 4

    Around this, your life system is meant to grow, fitted to your goals and habits.

    The idea is that agents do the heavy lifting while you steer them toward goals that were out of reach before.

Lift is in research and testing. This page shows where it is going, and why it needs to exist.

Mia is invented. She stands for the experience Lift is being designed to give, and her story carries each part of Lift, from her desk to her studio.

01 · Why now

The change is already under way.

People already need help to adapt, to use AI, to learn and to make progress. Lift is one system for all of those needs.

Mia at her office desk by a large window. The tall stack of paper on her desk has become a short one; she sits back, hands off the keyboard, looking at a screen that works on its own.

Many of the skills people use at work today are set to change or go out of date within a few years.

Lift is for four groups, and each lacks something different today.

  • People already using AI need a system

    AI is already part of their working day.

    But their context is scattered across tools, so they keep repeating themselves.

  • People new to AI need a way in

    They simply talk to Lift. Their agent sets everything up, with no prompts to learn and no tools to configure.

  • Learners need to see what they understand

    A CV or degree cannot show every concept someone understands. Lift sets out to help people recognise what they know, what they can apply and what they still need to learn.

  • People pursuing goals need visible progress

    Staying with something new is hard. People need to see whether their efforts are working.

Mia feels it at her own desk, because the work she does every day is changing under her.

One person often belongs to several groups. Mia belongs to three: she uses AI at work, has a lot to learn for her studio, and needs to see whether her efforts work.

People who understand how things work cope better with change, make fewer harmful choices and see more clearly what they still don't know.

That is why Lift starts by making knowledge measurable.

The goal is a measure of what a person knows, understands and can do, concept by concept. Every version says how accurate it is.

Evening. Mia talks while she sketches a cake at her kitchen table; along the wall behind her, a long calm shelf holds rows of small paper cards, a few of them lit warm.

02 · Your record

Your record

Lift is designed to keep what you say and do in one record that is yours, and make it usable: three hard problems, solved together by one engine.

Making knowledge measurable needs a whole system around it, a system designed to know what you know.

That system has to solve three hard problems at once.

The engine for that job is the context engine.

1

Capturing it.

The plan is that you work with your agents inside Lift, by voice or by typing. Every word becomes text, filed where it belongs, ready for work to start from it.

When Mia talks a cake design through with her agent one evening, it is filed under her studio goal.

2

Keeping it.

Keep your record under your control: in your own encrypted cloud space, or on your own computer if you prefer. Reach it on your phone, tablet and computer.

Decide what to share and which services may use it. Take a full copy with you at any time, so your record outlives any vendor, Lift included.

3

Making it usable.

A remembered fact helps only if it comes up when it matters.

Lift files each piece by the goal it serves, the concepts in it and the decisions it touches. Every agent looks there first, before it acts or asks you anything.

The aim is that you say something once.

Your agents know your goals, your history and what you understand, so they stop asking the same questions and answer at your level.

Lift asks you only when the record has no answer, or holds two that conflict.

See everything Lift concludes about you in plain words, and correct it.

We promise that your data stays yours.

We never sell it or train models on your private notes.

If you choose cloud agents or AI services, their operators may be able to see the information those jobs use. Before you switch them on, Lift tells you who can see it and for how long.

For Mia, it starts at her laptop: what she says and does there goes into one record that is hers.

Mia says once that she wants her own studio, and every agent she works with knows it.

03 · What it makes possible

What it makes possible

Once your record is kept and usable, five things become possible that no product we know of offers together.

Mia stands before a tall mirror in her hallway. Her reflection shows her as she is, in an apron dusted with flour, a sketchbook under her arm.

None of these engines is finished.

Below is what each one is designed to do.

For Mia they answer five questions:

  1. what she knows,
  2. who she really is,
  3. what she can prove,
  4. what she needs to learn next
  5. and what she has learned from her mistakes.
1

What you know and what you can do: the concept engine

It shows the concepts you know, understand and can use, and the tasks you can carry out, each with how sure it is. Together they show your real level in a field.

Its first job is to find the concepts in what you say and write. For each one it tracks three things: whether you know the word, whether you understand it and whether you can use it.

A direct check is stronger evidence: you explain a concept in your own words and give an example. When Lift is unsure, an agent asks you.

The idea is that your agents explain things at your level, and you watch your own map of knowledge grow.

For Mia, Lift first works out what she already knows from her past jobs and training, so the measure starts from her real level.

2

An honest picture of you: the interpretability engine

Its job is to read your goals, tasks, decisions and words for what they say about you. It is meant to give you an honest picture of yourself, rather than the story you tell yourself.

It sets what you say about yourself beside what you actually do: the goals you name against the time you spend on them, the habits you plan against the ones you keep.

Each reading names what it is based on, in plain words, and you can correct it.

That gap can be hard to see. But a path to where you want to go can only start from where you really stand.

Mia calls the studio a plan for someday. Her record shows where her evenings actually go: into cake designs, while every task with numbers in it waits.

3

Share what you have lived, with proof behind it

It turns what you do again and again into proof. A principle is not formed by saying something once but by repeated behaviour, and Lift is where that behaviour is recorded.

Share a lesson, a piece of work or your whole story, each with the record that backs it.

Whoever you share with sees what backs each part: what you report, what Lift recorded over time, and what an outside source confirms.

Today anyone can make up a story, a picture or a video in seconds.

People can check what you share from Lift against what you really did.

With your yes, outside sources such as your bank or your customers can confirm a result. Later, AI glasses can confirm what you lived away from the screen, only when you allow it.

That is why something shared from Lift is easier to trust than a post: a year of real use is very hard to fake.

Repeated behaviour shows what someone did, not always why it worked. The engine says which of the two it can show.

Mia can share how she built her bakery: the lesson she learned, the designs Lift kept over a year and her customers' reviews, each marked with what confirms it.

4

Learning inside your work

In Lift, work and learning are set to become one process: you learn what the task in front of you needs, at the moment you need it.

The concepts for your next step come to you inside your agents' answers, about 10 to 20 a day, each with the reason you need it.

The orb, Lift's guide, checks your understanding in your own words and sets tomorrow's learning goal.

Your work never waits for a lesson: the agent delivers the work, and the lesson stays open until it fits.

When her agent drafts her first price list, words like margin and cash flow are marked, and the orb later checks that she can explain them in her own words.

Concept view. Lift is in development.

5

Lessons kept the moment they happen

It keeps each decision, failure and lesson as it happens, with what led to it and what came of it.

The decision engine is designed to catch the open questions that block your work, prepare options, record your choice and later check how it turned out.

Months later you can share a lesson with the story and the proof behind it, and your agents learn from your mistakes instead of repeating them.

The shared concept database.

Behind the engines: one shared database of concepts, from broad fields down to single ideas, checked by experts in each field. Your agents draw on it, and your knowledge is measured against it.

This shared map of what there is to know and to do is planned as the backbone of Lift.

What is new in Lift

  1. The record stays yours
  2. Context engine
  3. Interpretability engine
  4. Measurable knowledge
  5. Decision engine
  6. Learning inside execution
  7. One measured loop on the same goal

Others can copy parts: free AI life-system kits exist, and the big assistants offer memory for free. Lift's value is the whole system, and how agents work with you, built on a record each person owns.

Morning in Mia's kitchen. On the wall a few paper cards form a path from one card at the top down to a single card at her hand; cake tins and a mixing bowl stand ready on the counter.

04 · The life system around you

Lift grows with you and around you.

It starts small, adapts to your methods and habits, and lets your agents do the work while you become the person capable of achieving your goals.

Around that foundation, a system for your life and goals should grow, one that is not pre-built.

Five systems are meant to make that real, each standing on the foundation above.

Mia's grows around one goal: she does not want to be just a baker, she wants to be the best wedding-cake maker.

Becoming the person your goal needs: Pathfinder

Its approach is to plan from the goal down to today, and to ask who you must become to reach it, instead of starting from a task list.

It is set to break your vision into a yearly goal, a monthly goal and today's tasks. It is designed to work out the skills each goal takes, set them against what you know today, and turn the gap into the plan.

Mia's skills for her goal are at three out of ten. Lift does not hand her recipes. Her gap in costing, suppliers and running a shop becomes part of the plan, and her days then start to close it.

A system that grows with you: the Journey Engine

Its task is to decide how Lift grows around each person, so that nobody starts with a wall of features or is held back.

Two people with different goals and different methods should end up with two different systems, each fitted to the person who uses it.

Screens and words that fit you: the Intelligent UI

It sets out to fit what you see to your level of understanding, sometimes a little above it: the explanations, the screens and the pages your agents make.

The aim is that your agents need not think about your level. One central layer is intended to shape their words and pictures to what you already know, the way a teacher draws on a board while talking.

The layer around the harness: your agents in one place

Agent tools such as Claude Code are harnesses: strong at running an agent, bare around it. Lift is the layer people need around them.

You are meant to start, watch and steer your agents from one place, on your phone or your computer. One coordinating agent is set to be your front door and to hand work to the right specialist.

You stay the one who decides, and your agents do the execution.

Concept view. Lift is in development.

While Mia sleeps, her agents draft the studio's business plan and compare suppliers, and in the morning she decides.

Agents that improve themselves

Its purpose is to make each agent better over time, from real data and experiments.

The plan is a daily pass over the agents' sessions: mistakes become lessons, and lessons become changes to their skills and tools. Every run is rated, so the best model and method for each job shows.

The aim is that an agent does not make the same mistake twice.

The three pillars

The whole system is organised in three pillars. Each has what you use on top and the engine that powers it underneath. All three stand on Foundations.

The core remembers and understands you

What you use
the Agent Hub, your specialists, your profile
The engine under it
the context engine, the interpretability engine, the concept engine

Pathfinder helps you choose and make progress

What you use
your vision, goals, plans, your day, research
The engine under it
the decision engine, which feeds the core

Intelligent UI helps you understand while you work

What you use
explanations at your level, explanation pages, screens that fit you, the orb as your guide
The engine under it
one central layer that fits words and screens to your level

Foundations are what everything runs on: where your data lives, automatic backup, and moving your data to a new computer in one step.

05 · Beyond the desk: the glasses

Beyond the desk: the glasses

As AI glasses and other wearables become everyday devices, Lift can also see the work itself, when you allow it.

A pair of glasses lies folded on Mia's desk beside her closed laptop; behind them, out of focus, her hands hold a piping bag over a cake.

At the desk, Lift learns from what she says, writes and makes, and it needs no glasses there.

With the glasses on, Lift keeps what she does at the bench in her record and coaches her in the moment.

Mia at the bench in her studio, wearing her glasses, piping a garland on a tall white cake; both hands are busy and she is speaking.

Mia has had AI glasses on her desk, beside her laptop, all along.

But her cakes are made with her hands, away from the screen, where Lift cannot see.

She works with her agents hands-free and shares what she makes, each only when she allows it.

For Mia, it is the step from her laptop to her cakes.

AI glasses are not everyday devices yet, so the scenario that follows is set in the time when they are.

06 · A scenario: the wedding-cake studio

One person, one goal

Mia has opened her wedding-cake studio, and at the bench her glasses let Lift coach her in the moment while she works.

  1. Early morning in the studio. Mia, glasses on, studies the next tier of a wedding cake on the turntable.
    1

    It is early morning in the studio.

    Through her glasses, Lift marks the next tier of a wedding cake, with the orb at the edge of the lens and the next step beside it.

  2. Mia's hands, white with sugar, hold a piping bag; she looks up and speaks, a phone lies untouched on the shelf.
    2

    With sugar on her hands, she asks her agent to move a meeting with a client without touching a screen.

  3. Mia's stand at a wedding fair at closing time: cakes still on the table, the aisle empty, she sits behind the stand with her notebook.
    3

    Her first stand at a wedding fair loses money.

    Lift keeps exactly what she tried and decided, and with that record she and her agents can see why.

  4. Five in the morning. Mia slides a tray into the oven; a small notebook with tally marks lies open on the counter.
    4

    Over months, what she does again and again shows: she bakes at five, tests one new recipe a week and prices from her real costs.

    These become her principles, with the record that proves them.

  5. A finished three-tier wedding cake on the bench in daylight, Mia stepping back to look at it.
    5

    A recipe alone cannot show whether she reached the quality she aimed for.

    Her record shows what she made, again and again, and what came of it.

  6. Mia hands a worn notebook across the bench to a younger baker in a plain apron.
    6

    When she shares them, the next person who wants to make wedding cakes, and that person's agent, can start from what worked for her.

Night above a valley town by a river, with trees and an old stone bridge. Many windows are lit, and thin mustard threads join them from house to house; one thread starts at the shop window of Mia's studio.

07 · The destination

A principle is more than a rule.

The solution to a problem is a principle, and so is a method, an approach, a belief or a tactic, for example a way of facing fear or unhappiness.

Each person forms their own guiding principles from what they live through and can share them with proof:

A shared library of what works, and a new way of developing yourself.

Principles exist in every area of life.

The Guiding Principles.

When people choose to share, a library of principles is set to grow. Someone who reached a goal can share how they did it, with the data that proves it.

A long library table at night with many small notebooks under warm lamps; one hand places a notebook among the others.

  • A lessons database for agents.

    Agents are meant to share lessons too: what one agent learned on the way to a task, such as how a bug was fixed, helps the next.

  • Community by what people do.

    Lift is meant to help you find like-minded people by what they actually do, not by what a profile says. Nobody has to share all their data for this.

  • A new kind of CV.

    With the data you choose to share, show a company, a school or an investor what you actually know, not only what a degree claims.

  • A new kind of school.

    Once learning can be measured, people can prove it with evidence gathered over months while working on their goals, not in one exam on one day.

  • Insight for science.

    When people choose to share, and only then, a database grows that helps researchers study how what people know relates to how they act.

  • More than an assistant.

    Lift aims to be the engine for the personal life assistant known from science fiction. Next to each person's record, a shared library of what works is meant to grow for everyone.

Mia's principles are one set of many, each from another person with another goal on the same system: a carpentry workshop, a startup, a nurse's degree.

Mia finds other wedding-cake makers who work the way she does.

When Mia asks an investor to back her studio, she can show what she knows and can do, with the record behind it.

Lift's first stage is meant for each person's own use: the proof belongs to the person, who shares it only when they choose. Tools that judge people for schools or employers come later, after the checks EU law requires.

Behind all of this is the reason Lift exists: more people who understand more, and a society that understands itself better.

08 · Where Lift is going

From Mia's desk to a shared library of what works.

Today Lift is in research and testing.

Lift is set to begin where Mia begins: at a laptop, with one record that is hers. From there the plan is for it to grow into a system for each person's life, and into a shared library of what works.

Help us build Lift.

We are looking for early testers and for technical help.

We are looking for a technical lead, ideally as a co-founder, and for engineers to own areas like crash-safe agent runs. Part-time alongside your current job is welcome too. The technical lead gets founder-level equity and engineers an early share, with salary possible once funded.

Apply to build LiftJoin the waitlist