DAILY RECORD / 2023
365 published entries in 2023.
NOTES
The year AI became personal
Retrospective AI note: The year AI became personal.
The next year
Retrospective AI note: The next year.
What I am taking forward
Retrospective AI note: What I am taking forward.
A public responsibility
Retrospective AI note: A public responsibility.
The information problem
Retrospective AI note: The information problem.
The archive continues
Retrospective AI note: The archive continues.
Time to reflect
Retrospective AI note: Time to reflect.
A quieter ambition
Retrospective AI note: A quieter ambition.
More agency
Retrospective AI note: More agency.
The beginning matters
Retrospective AI note: The beginning matters.
The end of easy assumptions
Retrospective AI note: The end of easy assumptions.
Making the invisible visible
Retrospective AI note: Making the invisible visible.
No substitute for conviction
Retrospective AI note: No substitute for conviction.
The question of authorship
Retrospective AI note: The question of authorship.
A different kind of team
Retrospective AI note: A different kind of team.
Personal systems
Retrospective AI note: Personal systems.
The work that remains
Retrospective AI note: The work that remains.
Less waiting
Retrospective AI note: Less waiting.
The open ecosystem
Retrospective AI note: The open ecosystem.
The ability to focus
Retrospective AI note: The ability to focus.
A year of prototypes
Retrospective AI note: A year of prototypes.
The definition of progress
Retrospective AI note: The definition of progress.
Regulation as a design input
Retrospective AI note: Regulation as a design input.
Europe chooses a direction
Retrospective AI note: Europe chooses a direction.
The world is not only text
Retrospective AI note: The world is not only text.
Gemini raises the multimodal bar
Retrospective AI note: Gemini raises the multimodal bar.
A new creative language
Retrospective AI note: A new creative language.
The learning infrastructure
Retrospective AI note: The learning infrastructure.
More than prediction
Retrospective AI note: More than prediction.
Slow understanding
Retrospective AI note: Slow understanding.
After the first year
Retrospective AI note: After the first year.
The first year
Retrospective AI note: The first year.
A new public square
Retrospective AI note: A new public square.
The next generation
Retrospective AI note: The next generation.
Better than certainty
Retrospective AI note: Better than certainty.
A personal archive
Retrospective AI note: A personal archive.
One year later
Retrospective AI note: One year later.
The work of translation
Retrospective AI note: The work of translation.
Gratitude for access
Retrospective AI note: Gratitude for access.
The technology keeps moving
Retrospective AI note: The technology keeps moving.
A reason to learn governance
Retrospective AI note: A reason to learn governance.
The human story behind the headlines
Retrospective AI note: The human story behind the headlines.
Power needs accountability
Retrospective AI note: Power needs accountability.
Systems depend on people
Retrospective AI note: Systems depend on people.
Leadership matters
Retrospective AI note: Leadership matters.
The pace of expectation
Retrospective AI note: The pace of expectation.
A human standard
Retrospective AI note: A human standard.
The test of usefulness
Retrospective AI note: The test of usefulness.
The return of domain expertise
Retrospective AI note: The return of domain expertise.
More tools, more intention
Retrospective AI note: More tools, more intention.
A new kind of product craft
Retrospective AI note: A new kind of product craft.
The builder’s moment
Retrospective AI note: The builder’s moment.
The value of specificity
Retrospective AI note: The value of specificity.
Everyone can shape a tool
Retrospective AI note: Everyone can shape a tool.
From model to assistant
Retrospective AI note: From model to assistant.
Building gets a new toolkit
Retrospective AI note: Building gets a new toolkit.
Before the next release
Retrospective AI note: Before the next release.
The value of restraint
Retrospective AI note: The value of restraint.
Safety needs builders
Retrospective AI note: Safety needs builders.
Shared problems
Retrospective AI note: Shared problems.
A global conversation
Retrospective AI note: A global conversation.
October’s lesson
Retrospective AI note: October’s lesson.
Safety enters government
Retrospective AI note: Safety enters government.
A public decision
Retrospective AI note: A public decision.
A different kind of literacy
Retrospective AI note: A different kind of literacy.
The discipline of verification
Retrospective AI note: The discipline of verification.
The role of institutions
Retrospective AI note: The role of institutions.
The future is uneven
Retrospective AI note: The future is uneven.
Not just productivity
Retrospective AI note: Not just productivity.
A higher standard for tools
Retrospective AI note: A higher standard for tools.
The unseen labour
Retrospective AI note: The unseen labour.
Creative access
Retrospective AI note: Creative access.
More than aesthetics
Retrospective AI note: More than aesthetics.
Images enter the conversation
Retrospective AI note: Images enter the conversation.
The creator’s responsibility
Retrospective AI note: The creator’s responsibility.
A different search habit
Retrospective AI note: A different search habit.
The trust gap
Retrospective AI note: The trust gap.
Keeping the human in view
Retrospective AI note: Keeping the human in view.
What I want to build
Retrospective AI note: What I want to build.
Better defaults
Retrospective AI note: Better defaults.
The feedback question
Retrospective AI note: The feedback question.
Useful memory
Retrospective AI note: Useful memory.
The price of speed
Retrospective AI note: The price of speed.
Better interfaces for knowledge
Retrospective AI note: Better interfaces for knowledge.
A question of access
Retrospective AI note: A question of access.
The long game
Retrospective AI note: The long game.
The real differentiator
Retrospective AI note: The real differentiator.
Less invisible work
Retrospective AI note: Less invisible work.
A better first step
Retrospective AI note: A better first step.
The problem of dependence
Retrospective AI note: The problem of dependence.
A smaller distance
Retrospective AI note: A smaller distance.
The next layer
Retrospective AI note: The next layer.
September’s lesson
Retrospective AI note: September’s lesson.
The problem of abundance
Retrospective AI note: The problem of abundance.
A useful tension
Retrospective AI note: A useful tension.
More models, more choice
Retrospective AI note: More models, more choice.
Interface is behaviour
Retrospective AI note: Interface is behaviour.
More modalities, more responsibility
Retrospective AI note: More modalities, more responsibility.
The person behind the output
Retrospective AI note: The person behind the output.
The value of provenance
Retrospective AI note: The value of provenance.
A better creative loop
Retrospective AI note: A better creative loop.
Seeing is changing
Retrospective AI note: Seeing is changing.
DALL·E 3 and precision
Retrospective AI note: DALL·E 3 and precision.
The new baseline for images
Retrospective AI note: The new baseline for images.
The human memory
Retrospective AI note: The human memory.
Better than a feature list
Retrospective AI note: Better than a feature list.
What users really need
Retrospective AI note: What users really need.
Learning by making
Retrospective AI note: Learning by making.
The creative partnership
Retrospective AI note: The creative partnership.
No shortcut to trust
Retrospective AI note: No shortcut to trust.
Context creates quality
Retrospective AI note: Context creates quality.
The archive matters
Retrospective AI note: The archive matters.
An assistant should teach
Retrospective AI note: An assistant should teach.
More work, less friction
Retrospective AI note: More work, less friction.
The organisational problem
Retrospective AI note: The organisational problem.
More than consumption
Retrospective AI note: More than consumption.
Developers become central
Retrospective AI note: Developers become central.
A personal edge
Retrospective AI note: A personal edge.
The power of a draft
Retrospective AI note: The power of a draft.
More human questions
Retrospective AI note: More human questions.
The product is the experience
Retrospective AI note: The product is the experience.
A builder’s autumn
Retrospective AI note: A builder’s autumn.
August’s lesson
Retrospective AI note: August’s lesson.
The work people keep
Retrospective AI note: The work people keep.
Trust unlocks use
Retrospective AI note: Trust unlocks use.
AI enters the enterprise
Retrospective AI note: AI enters the enterprise.
Better questions for companies
Retrospective AI note: Better questions for companies.
The operating leverage
Retrospective AI note: The operating leverage.
A language for builders
Retrospective AI note: A language for builders.
Code becomes more accessible
Retrospective AI note: Code becomes more accessible.
Helping people begin
Retrospective AI note: Helping people begin.
Confidence needs evidence
Retrospective AI note: Confidence needs evidence.
Systems, not features
Retrospective AI note: Systems, not features.
Fewer barriers
Retrospective AI note: Fewer barriers.
The compounding question
Retrospective AI note: The compounding question.
Creativity as direction
Retrospective AI note: Creativity as direction.
The new portfolio
Retrospective AI note: The new portfolio.
Useful skepticism
Retrospective AI note: Useful skepticism.
The new workplace layer
Retrospective AI note: The new workplace layer.
Learning the foundations
Retrospective AI note: Learning the foundations.
The simple test
Retrospective AI note: The simple test.
A bigger surface area
Retrospective AI note: A bigger surface area.
The cost of noise
Retrospective AI note: The cost of noise.
Good constraints
Retrospective AI note: Good constraints.
More prototypes, better decisions
Retrospective AI note: More prototypes, better decisions.
Less intimidation
Retrospective AI note: Less intimidation.
A learning companion
Retrospective AI note: A learning companion.
Personal context
Retrospective AI note: Personal context.
A clearer bar
Retrospective AI note: A clearer bar.
From answer to action
Retrospective AI note: From answer to action.
The overlooked workflow
Retrospective AI note: The overlooked workflow.
Independent thinking
Retrospective AI note: Independent thinking.
The work changes shape
Retrospective AI note: The work changes shape.
July’s lesson
Retrospective AI note: July’s lesson.
A public conversation
Retrospective AI note: A public conversation.
Ambition gets practical
Retrospective AI note: Ambition gets practical.
The problem behind the prompt
Retrospective AI note: The problem behind the prompt.
Better than busy
Retrospective AI note: Better than busy.
The voice of the product
Retrospective AI note: The voice of the product.
The validation trap
Retrospective AI note: The validation trap.
More room to experiment
Retrospective AI note: More room to experiment.
The first version is cheap
Retrospective AI note: The first version is cheap.
Safety as a product decision
Retrospective AI note: Safety as a product decision.
Commitments are only a start
Retrospective AI note: Commitments are only a start.
The responsibility of access
Retrospective AI note: The responsibility of access.
Capability travels
Retrospective AI note: Capability travels.
Llama 2 changes the mood
Retrospective AI note: Llama 2 changes the mood.
Distribution meets openness
Retrospective AI note: Distribution meets openness.
Open questions
Retrospective AI note: Open questions.
The little things
Retrospective AI note: The little things.
A collaborator, not an oracle
Retrospective AI note: A collaborator, not an oracle.
The race for reliability
Retrospective AI note: The race for reliability.
Different models, different values
Retrospective AI note: Different models, different values.
More than one frontier
Retrospective AI note: More than one frontier.
The right kind of leverage
Retrospective AI note: The right kind of leverage.
A more open future
Retrospective AI note: A more open future.
What feels durable
Retrospective AI note: What feels durable.
The quality of attention
Retrospective AI note: The quality of attention.
Building from the problem
Retrospective AI note: Building from the problem.
Human motivation
Retrospective AI note: Human motivation.
A new kind of apprenticeship
Retrospective AI note: A new kind of apprenticeship.
The power of iteration
Retrospective AI note: The power of iteration.
A personal benchmark
Retrospective AI note: A personal benchmark.
The second half
Retrospective AI note: The second half.
June’s lesson
Retrospective AI note: June’s lesson.
I want better questions
Retrospective AI note: I want better questions.
Making knowledge usable
Retrospective AI note: Making knowledge usable.
The real bottleneck
Retrospective AI note: The real bottleneck.
Less fear of the blank page
Retrospective AI note: Less fear of the blank page.
Careful optimism
Retrospective AI note: Careful optimism.
More possible paths
Retrospective AI note: More possible paths.
Information versus insight
Retrospective AI note: Information versus insight.
The work after automation
Retrospective AI note: The work after automation.
The human checkpoint
Retrospective AI note: The human checkpoint.
A builder’s responsibility
Retrospective AI note: A builder’s responsibility.
Scale is not understanding
Retrospective AI note: Scale is not understanding.
What students need
Retrospective AI note: What students need.
The confidence problem
Retrospective AI note: The confidence problem.
A public infrastructure question
Retrospective AI note: A public infrastructure question.
Regulation is not the enemy
Retrospective AI note: Regulation is not the enemy.
Rules arrive
Retrospective AI note: Rules arrive.
AI starts using tools
Retrospective AI note: AI starts using tools.
The research gap
Retrospective AI note: The research gap.
Incomplete answers
Retrospective AI note: Incomplete answers.
The system behind intelligence
Retrospective AI note: The system behind intelligence.
The useful friction
Retrospective AI note: The useful friction.
Depth over volume
Retrospective AI note: Depth over volume.
Building for the edge case
Retrospective AI note: Building for the edge case.
The overlooked people
Retrospective AI note: The overlooked people.
The personal operating system
Retrospective AI note: The personal operating system.
More power, more responsibility
Retrospective AI note: More power, more responsibility.
A new expectation
Retrospective AI note: A new expectation.
The temptation to skip
Retrospective AI note: The temptation to skip.
The first real habits
Retrospective AI note: The first real habits.
May’s lesson
Retrospective AI note: May’s lesson.
Serious play
Retrospective AI note: Serious play.
The gap is not fixed
Retrospective AI note: The gap is not fixed.
Creative direction
Retrospective AI note: Creative direction.
Agency over engagement
Retrospective AI note: Agency over engagement.
The learning opportunity
Retrospective AI note: The learning opportunity.
A thousand small assistants
Retrospective AI note: A thousand small assistants.
The best questions are specific
Retrospective AI note: The best questions are specific.
Speed needs standards
Retrospective AI note: Speed needs standards.
Work becomes conversational
Retrospective AI note: Work becomes conversational.
No-code changes shape
Retrospective AI note: No-code changes shape.
The founder advantage
Retrospective AI note: The founder advantage.
Context is care
Retrospective AI note: Context is care.
AI becomes more personal
Retrospective AI note: AI becomes more personal.
A fast mirror
Retrospective AI note: A fast mirror.
The work is still human
Retrospective AI note: The work is still human.
Better feedback loops
Retrospective AI note: Better feedback loops.
The risk of sameness
Retrospective AI note: The risk of sameness.
The new literacy
Retrospective AI note: The new literacy.
Useful, not magical
Retrospective AI note: Useful, not magical.
Distribution changes everything
Retrospective AI note: Distribution changes everything.
AI becomes platform-level
Retrospective AI note: AI becomes platform-level.
The pressure to move
Retrospective AI note: The pressure to move.
What cannot be generated
Retrospective AI note: What cannot be generated.
Creative confidence
Retrospective AI note: Creative confidence.
The source beneath the summary
Retrospective AI note: The source beneath the summary.
A tool for the curious
Retrospective AI note: A tool for the curious.
The system matters
Retrospective AI note: The system matters.
A better tutor
Retrospective AI note: A better tutor.
Attention is a scarce resource
Retrospective AI note: Attention is a scarce resource.
Beyond the novelty
Retrospective AI note: Beyond the novelty.
April’s lesson
Retrospective AI note: April’s lesson.
The real compounding
Retrospective AI note: The real compounding.
Access is not understanding
Retrospective AI note: Access is not understanding.
The cost of being early
Retrospective AI note: The cost of being early.
One person, many roles
Retrospective AI note: One person, many roles.
A larger imagination
Retrospective AI note: A larger imagination.
Learning how to verify
Retrospective AI note: Learning how to verify.
The quiet advantage
Retrospective AI note: The quiet advantage.
Better inputs
Retrospective AI note: Better inputs.
A new creative baseline
Retrospective AI note: A new creative baseline.
The answer is not the end
Retrospective AI note: The answer is not the end.
Trust is the product
Retrospective AI note: Trust is the product.
The human layer
Retrospective AI note: The human layer.
Small teams, larger ambition
Retrospective AI note: Small teams, larger ambition.
A reason to read deeper
Retrospective AI note: A reason to read deeper.
The interface is changing
Retrospective AI note: The interface is changing.
My new research assistant
Retrospective AI note: My new research assistant.
Not everything should be automated
Retrospective AI note: Not everything should be automated.
Building in public
Retrospective AI note: Building in public.
The work behind the answer
Retrospective AI note: The work behind the answer.
A strange kind of access
Retrospective AI note: A strange kind of access.
The student version of this
Retrospective AI note: The student version of this.
Intelligence without responsibility
Retrospective AI note: Intelligence without responsibility.
Questions become leverage
Retrospective AI note: Questions become leverage.
The first draft changes
Retrospective AI note: The first draft changes.
More than a chatbot
Retrospective AI note: More than a chatbot.
Judgment stays expensive
Retrospective AI note: Judgment stays expensive.
The danger of the demo
Retrospective AI note: The danger of the demo.
A new default
Retrospective AI note: A new default.
The speed is unsettling
Retrospective AI note: The speed is unsettling.
AI meets regulation
Retrospective AI note: AI meets regulation.
Deployment is the real test
Retrospective AI note: Deployment is the real test.
The public catches up
Retrospective AI note: The public catches up.
Scale is not a plan
Retrospective AI note: Scale is not a plan.
Intelligence needs friction
Retrospective AI note: Intelligence needs friction.
Tools need judgment
Retrospective AI note: Tools need judgment.
The agent question
Retrospective AI note: The agent question.
A model with hands
Retrospective AI note: A model with hands.
AI gets tools
Retrospective AI note: AI gets tools.
The pause debate begins
Retrospective AI note: The pause debate begins.
The conversation expands
Retrospective AI note: The conversation expands.
An assistant needs permission
Retrospective AI note: An assistant needs permission.
Agency inside the office
Retrospective AI note: Agency inside the office.
The danger of effortless output
Retrospective AI note: The danger of effortless output.
The work behind the work
Retrospective AI note: The work behind the work.
AI moves into the documents
Retrospective AI note: AI moves into the documents.
Capability and responsibility
Retrospective AI note: Capability and responsibility.
GPT-4 changes the conversation
Retrospective AI note: GPT-4 changes the conversation.
Before the next model
Retrospective AI note: Before the next model.
A faster first draft
Retrospective AI note: A faster first draft.
The source is still the work
Retrospective AI note: The source is still the work.
The age of assistants
Retrospective AI note: The age of assistants.
Organisations are not datasets
Retrospective AI note: Organisations are not datasets.
Context is the advantage
Retrospective AI note: Context is the advantage.
The office is changing
Retrospective AI note: The office is changing.
AI enters business software
Retrospective AI note: AI enters business software.
Workflow over prompt
Retrospective AI note: Workflow over prompt.
A prototype is not a company
Retrospective AI note: A prototype is not a company.
The application layer
Retrospective AI note: The application layer.
AI becomes infrastructure
Retrospective AI note: AI becomes infrastructure.
The model becomes an API
Retrospective AI note: The model becomes an API.
One month of acceleration
Retrospective AI note: One month of acceleration.
The wrapper question
Retrospective AI note: The wrapper question.
Evaluation becomes a profession
Retrospective AI note: Evaluation becomes a profession.
Capability spreads
Retrospective AI note: Capability spreads.
The model leaves the building
Retrospective AI note: The model leaves the building.
The future is multi-model
Retrospective AI note: The future is multi-model.
Intelligence wants to escape the lab
Retrospective AI note: Intelligence wants to escape the lab.
Open models change the map
Retrospective AI note: Open models change the map.
The company inside the model
Retrospective AI note: The company inside the model.
The anthropomorphism trap
Retrospective AI note: The anthropomorphism trap.
A tool is not a friend
Retrospective AI note: A tool is not a friend.
The boundary matters
Retrospective AI note: The boundary matters.
When chat becomes strange
Retrospective AI note: When chat becomes strange.
Answers are not decisions
Retrospective AI note: Answers are not decisions.
A new kind of search habit
Retrospective AI note: A new kind of search habit.
Real-time intelligence
Retrospective AI note: Real-time intelligence.
Every answer needs a source
Retrospective AI note: Every answer needs a source.
The race is now public
Retrospective AI note: The race is now public.
Search needs humility
Retrospective AI note: Search needs humility.
The demo is not the product
Retrospective AI note: The demo is not the product.
Speed changes behaviour
Retrospective AI note: Speed changes behaviour.
The browser becomes an assistant
Retrospective AI note: The browser becomes an assistant.
Search is awake
Retrospective AI note: Search is awake.
Capability is not intimacy
Retrospective AI note: Capability is not intimacy.
The first personal tool
Retrospective AI note: The first personal tool.
The assistant economy
Retrospective AI note: The assistant economy.
The price of availability
Retrospective AI note: The price of availability.
Intelligence becomes a subscription
Retrospective AI note: Intelligence becomes a subscription.
What comes after ChatGPT?
Retrospective AI note: What comes after ChatGPT?
Trust is the bottleneck
Retrospective AI note: Trust is the bottleneck.
A literacy problem
Retrospective AI note: A literacy problem.
Not replacement
Retrospective AI note: Not replacement.
The new generalist
Retrospective AI note: The new generalist.
The future of teams
Retrospective AI note: The future of teams.
The system behind the answer
Retrospective AI note: The system behind the answer.
Intelligence as infrastructure
Retrospective AI note: Intelligence as infrastructure.
The race becomes visible
Retrospective AI note: The race becomes visible.
The AI team
Retrospective AI note: The AI team.
Taste after automation
Retrospective AI note: Taste after automation.
Code is becoming conversation
Retrospective AI note: Code is becoming conversation.
Data is still people
Retrospective AI note: Data is still people.
An assistant needs boundaries
Retrospective AI note: An assistant needs boundaries.
The first product instinct
Retrospective AI note: The first product instinct.
Work is moving
Retrospective AI note: Work is moving.
The machine is not the point
Retrospective AI note: The machine is not the point.
The first layer of agency
Retrospective AI note: The first layer of agency.
A tool for the ambitious
Retrospective AI note: A tool for the ambitious.
The confidence problem
Retrospective AI note: The confidence problem.
A tutor without shame
Retrospective AI note: A tutor without shame.
The new research habit
Retrospective AI note: The new research habit.
The speed of curiosity
Retrospective AI note: The speed of curiosity.
Intelligence without memory
Retrospective AI note: Intelligence without memory.
The prompt is a prototype
Retrospective AI note: The prompt is a prototype.
The cost of a first answer
Retrospective AI note: The cost of a first answer.
A model needs a job
Retrospective AI note: A model needs a job.
The blank page changed
Retrospective AI note: The blank page changed.
The old internet
Retrospective AI note: The old internet.
The first question
Retrospective AI note: The first question.
The year starts with a chat box
Retrospective AI note: The year starts with a chat box.