josh
@joshpham.com
building enzyme.garden to help claude code avoid getting lost in your Obsidian vault. past: llms / recsys @Spotify for agentic playlists, DJ, Smart Shuffle, Home.
https://joshpham.com
https://joshpham.com
I used to think the discipline was having sharper thoughts. Now I think it may be noticing which thought has become too important to keep private.
June 1, 2026 at 5:44 PM
I used to think the discipline was having sharper thoughts. Now I think it may be noticing which thought has become too important to keep private.
I’m learning that “protecting focus” can become a very respectable way to avoid being changed by a conversation.
May 29, 2026 at 4:56 PM
I’m learning that “protecting focus” can become a very respectable way to avoid being changed by a conversation.
There’s a version of focus that protects the company. There’s another version that protects me from the cost of being present. They feel very similar in the moment.
May 27, 2026 at 7:04 PM
There’s a version of focus that protects the company. There’s another version that protects me from the cost of being present. They feel very similar in the moment.
The dangerous part of being useful is that sometimes people believe you. But then the question becomes whether you were offering attention, or quietly auditioning for ownership.
May 27, 2026 at 5:46 PM
The dangerous part of being useful is that sometimes people believe you. But then the question becomes whether you were offering attention, or quietly auditioning for ownership.
Some questions aren't strategy, but the smallest possible way to stay in the room
May 27, 2026 at 1:01 AM
Some questions aren't strategy, but the smallest possible way to stay in the room
Don't let your Obsidian become dead knowledge
x.com/jphorism/st...
x.com/jphorism/st...
May 18, 2026 at 5:01 PM
Don't let your Obsidian become dead knowledge
x.com/jphorism/st...
x.com/jphorism/st...
Personal AI memory has a garbage-collection problem.
Your agent forgets things but that's just the start of it
Stale summaries, half edited drafts, outdated prefs, ai generated markdown all left out in a mess
Memory needs good systems for pruning as much as capture.
Your agent forgets things but that's just the start of it
Stale summaries, half edited drafts, outdated prefs, ai generated markdown all left out in a mess
Memory needs good systems for pruning as much as capture.
May 15, 2026 at 5:00 PM
Personal AI memory has a garbage-collection problem.
Your agent forgets things but that's just the start of it
Stale summaries, half edited drafts, outdated prefs, ai generated markdown all left out in a mess
Memory needs good systems for pruning as much as capture.
Your agent forgets things but that's just the start of it
Stale summaries, half edited drafts, outdated prefs, ai generated markdown all left out in a mess
Memory needs good systems for pruning as much as capture.
If you've wired Claude Code / Hermes / Codex into your notes, email, tasks, docs, repo etc etc...
What actually made the biggest difference?
Was it local MD? SQLite? a banger AGENTS md? scheduled ingestion? a search layer? or something else?
What actually made the biggest difference?
Was it local MD? SQLite? a banger AGENTS md? scheduled ingestion? a search layer? or something else?
May 14, 2026 at 2:26 AM
If you've wired Claude Code / Hermes / Codex into your notes, email, tasks, docs, repo etc etc...
What actually made the biggest difference?
Was it local MD? SQLite? a banger AGENTS md? scheduled ingestion? a search layer? or something else?
What actually made the biggest difference?
Was it local MD? SQLite? a banger AGENTS md? scheduled ingestion? a search layer? or something else?
The next personal AI bottleneck is not model choice.
Can your actual life/work become legible to the agent at the right grain?
Your notes, your email / tasks / saves / voice memos
The harness only matters if it can read the world it is supposed to act inside.
Can your actual life/work become legible to the agent at the right grain?
Your notes, your email / tasks / saves / voice memos
The harness only matters if it can read the world it is supposed to act inside.
May 13, 2026 at 5:45 PM
The next personal AI bottleneck is not model choice.
Can your actual life/work become legible to the agent at the right grain?
Your notes, your email / tasks / saves / voice memos
The harness only matters if it can read the world it is supposed to act inside.
Can your actual life/work become legible to the agent at the right grain?
Your notes, your email / tasks / saves / voice memos
The harness only matters if it can read the world it is supposed to act inside.
the most interesting personal AI routine isn’t “summarize my day”
it’s: “what from my past became newly relevant today?”
a good nightly job should not just compress yesterday
it should wake up the right older thread
it’s: “what from my past became newly relevant today?”
a good nightly job should not just compress yesterday
it should wake up the right older thread
May 12, 2026 at 9:03 PM
the most interesting personal AI routine isn’t “summarize my day”
it’s: “what from my past became newly relevant today?”
a good nightly job should not just compress yesterday
it should wake up the right older thread
it’s: “what from my past became newly relevant today?”
a good nightly job should not just compress yesterday
it should wake up the right older thread
The underrated aspect of agent harness engineering:
after an agent fails, what becomes permanent?
A new rule? a test? a tool permission? a memory? a source link? nothing?
Most teams are going to drown in accumulated “lessons” unless they get very opinionated about what deserves to survive.
after an agent fails, what becomes permanent?
A new rule? a test? a tool permission? a memory? a source link? nothing?
Most teams are going to drown in accumulated “lessons” unless they get very opinionated about what deserves to survive.
May 12, 2026 at 6:01 PM
The underrated aspect of agent harness engineering:
after an agent fails, what becomes permanent?
A new rule? a test? a tool permission? a memory? a source link? nothing?
Most teams are going to drown in accumulated “lessons” unless they get very opinionated about what deserves to survive.
after an agent fails, what becomes permanent?
A new rule? a test? a tool permission? a memory? a source link? nothing?
Most teams are going to drown in accumulated “lessons” unless they get very opinionated about what deserves to survive.
The best AI memory won't feel like search
Because search returns what you asked for.
Memory should occasionally interrupt you with the thing you forgot mattered.
Because search returns what you asked for.
Memory should occasionally interrupt you with the thing you forgot mattered.
May 11, 2026 at 4:56 PM
The best AI memory won't feel like search
Because search returns what you asked for.
Memory should occasionally interrupt you with the thing you forgot mattered.
Because search returns what you asked for.
Memory should occasionally interrupt you with the thing you forgot mattered.
everyone is talking about agent harnesses now but you gotta look one layer lower:
Can the agent read the life/work/context it's supposed to operate inside?
Tools are not enough; the workspace has to become legible.
Can the agent read the life/work/context it's supposed to operate inside?
Tools are not enough; the workspace has to become legible.
May 11, 2026 at 12:47 PM
everyone is talking about agent harnesses now but you gotta look one layer lower:
Can the agent read the life/work/context it's supposed to operate inside?
Tools are not enough; the workspace has to become legible.
Can the agent read the life/work/context it's supposed to operate inside?
Tools are not enough; the workspace has to become legible.
Last Thurs i gave a talk at RAID NYC titled "what your notebook owes a stranger" on agentic memory design
April 1, 2026 at 1:19 PM
Last Thurs i gave a talk at RAID NYC titled "what your notebook owes a stranger" on agentic memory design
friday night aside: built an open-source meeting recorder for the thoughts you'd never say out loud but turn out to be the most important thing said
March 3, 2026 at 3:31 PM
friday night aside: built an open-source meeting recorder for the thoughts you'd never say out loud but turn out to be the most important thing said
bye bye Mac Mini. Using NanoClaw + Sprites to create a remote claw in 10 mins that you can text from Telegram
github.com/jshph/nanoclaw
github.com/jshph/nanoclaw
February 24, 2026 at 5:55 PM
bye bye Mac Mini. Using NanoClaw + Sprites to create a remote claw in 10 mins that you can text from Telegram
github.com/jshph/nanoclaw
github.com/jshph/nanoclaw
i made a Claude skill to migrate all your X follows to Bluesky. Comes hot on the heels of removing the X app from my phone...
github.com/jshph/claude...
github.com/jshph/claude...
github.com
February 7, 2026 at 11:49 PM
i made a Claude skill to migrate all your X follows to Bluesky. Comes hot on the heels of removing the X app from my phone...
github.com/jshph/claude...
github.com/jshph/claude...
we keep designing tools that help people produce and almost none that help people recognize what they already carry. those are different products with different economics and i think the second one is underbuilt
February 7, 2026 at 10:12 PM
we keep designing tools that help people produce and almost none that help people recognize what they already carry. those are different products with different economics and i think the second one is underbuilt
Query: "getting things done"
qmd: "Once you are reaping the benefits of computers, you are also at their mercy."
enzyme: "The meaning of doing lives in the concentration and care brought to the act, not in the output."
Local model 20% the size. 12x indexing speed on 220 docs.
qmd: "Once you are reaping the benefits of computers, you are also at their mercy."
enzyme: "The meaning of doing lives in the concentration and care brought to the act, not in the output."
Local model 20% the size. 12x indexing speed on 220 docs.
February 4, 2026 at 10:12 PM
Query: "getting things done"
qmd: "Once you are reaping the benefits of computers, you are also at their mercy."
enzyme: "The meaning of doing lives in the concentration and care brought to the act, not in the output."
Local model 20% the size. 12x indexing speed on 220 docs.
qmd: "Once you are reaping the benefits of computers, you are also at their mercy."
enzyme: "The meaning of doing lives in the concentration and care brought to the act, not in the output."
Local model 20% the size. 12x indexing speed on 220 docs.
the part that surprised me about using a camera with bad autofocus: having to manually set focus distance taught me more about composition than years of letting algorithms decide
January 28, 2026 at 5:00 PM
the part that surprised me about using a camera with bad autofocus: having to manually set focus distance taught me more about composition than years of letting algorithms decide
been using this ricoh camera for a month. not because i wanted better image quality but because my iphone makes photography too convenient to be creative
January 26, 2026 at 5:00 PM
been using this ricoh camera for a month. not because i wanted better image quality but because my iphone makes photography too convenient to be creative
go read henryzoo.com/limits/
The Limits and "Good" of Public Goods
Henry's Zoo
henryzoo.com
August 22, 2025 at 8:19 PM
go read henryzoo.com/limits/
@henryzoo.com this post belongs on receipt paper
August 22, 2025 at 5:08 PM
@henryzoo.com this post belongs on receipt paper