verdverm
@verdverm.com
dev & entrepreneur interested in atproto, cuelang, machine learning, developer experience, combating misinformation
working on https://blebbit.app | @blebbit.app | #blebbit
personal: https://verdverm.com | https://github.com/verdverm
working on https://blebbit.app | @blebbit.app | #blebbit
personal: https://verdverm.com | https://github.com/verdverm
Pinned
verdverm
@verdverm.com
· 15h
This is now the best session I've had with my custom #ai coding agent setup
I'm seriously impressed, this was days of work and iterations handled in $10 over about an hour
Used gemini-3-pro for this one because I knew that extra high-level, abstract reasoning was needed over daily coding task doer
I'm seriously impressed, this was days of work and iterations handled in $10 over about an hour
Used gemini-3-pro for this one because I knew that extra high-level, abstract reasoning was needed over daily coding task doer
This is now the best session I've had with my custom #ai coding agent setup
I'm seriously impressed, this was days of work and iterations handled in $10 over about an hour
Used gemini-3-pro for this one because I knew that extra high-level, abstract reasoning was needed over daily coding task doer
I'm seriously impressed, this was days of work and iterations handled in $10 over about an hour
Used gemini-3-pro for this one because I knew that extra high-level, abstract reasoning was needed over daily coding task doer
January 10, 2026 at 2:57 PM
This is now the best session I've had with my custom #ai coding agent setup
I'm seriously impressed, this was days of work and iterations handled in $10 over about an hour
Used gemini-3-pro for this one because I knew that extra high-level, abstract reasoning was needed over daily coding task doer
I'm seriously impressed, this was days of work and iterations handled in $10 over about an hour
Used gemini-3-pro for this one because I knew that extra high-level, abstract reasoning was needed over daily coding task doer
Shouldi?
Deciding if you should do something based on a diff is a keystone concept in CI systems
This is an abstract version in @cuelang.org
#devops
Deciding if you should do something based on a diff is a keystone concept in CI systems
This is an abstract version in @cuelang.org
#devops
![// #Shouldi resolves then or else
// based on a diff and patterns
#Shouldi: ref & {
// changes to match against
diff!: #Changes
// patterns for matching
include: [...string]
exclude: [...string]
// just do it!
force?: bool
// what to do
then!: _
else?: _
}](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:2jtyqespp2zfodukwvktqwe6/bafkreig4dslo52tp3hco6bbwmu6zmwdizkniqimnmtpolqx3pkecbbllam@jpeg)
January 9, 2026 at 9:41 AM
Shouldi?
Deciding if you should do something based on a diff is a keystone concept in CI systems
This is an abstract version in @cuelang.org
#devops
Deciding if you should do something based on a diff is a keystone concept in CI systems
This is an abstract version in @cuelang.org
#devops
@bnewbold.net @roomy.space
what are your thoughts / ideas on message gardening and at:// links?
#atdev
what are your thoughts / ideas on message gardening and at:// links?
#atdev
January 7, 2026 at 11:27 PM
@bnewbold.net @roomy.space
what are your thoughts / ideas on message gardening and at:// links?
#atdev
what are your thoughts / ideas on message gardening and at:// links?
#atdev
oh man, forgot to include the shared dependency caches I built for veg/env when giving it to my agent, so of course it has to download them all again, blowing up my sessions layer sizes and filling my disk... at least it's limited to colima's disk and I can still work
January 6, 2026 at 10:37 AM
oh man, forgot to include the shared dependency caches I built for veg/env when giving it to my agent, so of course it has to download them all again, blowing up my sessions layer sizes and filling my disk... at least it's limited to colima's disk and I can still work
@simonwillison.net @timkellogg.me @eugenevinitsky.bsky.social
what if we make (maybe there is?) a marathon benchmark of some sort, like how long can your agent remain coherent while performing related and follow on tasks
we all do that "just this one little more thing" since before ai, measure it
what if we make (maybe there is?) a marathon benchmark of some sort, like how long can your agent remain coherent while performing related and follow on tasks
we all do that "just this one little more thing" since before ai, measure it
yeah, since I do that, a lot of my session look like this ( 50/50, 50/50, 99/1 ), mind you this is gemini-3-flash
stats from a recent #agent run to 2.3M tokens
stats from a recent #agent run to 2.3M tokens
January 6, 2026 at 8:08 AM
@simonwillison.net @timkellogg.me @eugenevinitsky.bsky.social
what if we make (maybe there is?) a marathon benchmark of some sort, like how long can your agent remain coherent while performing related and follow on tasks
we all do that "just this one little more thing" since before ai, measure it
what if we make (maybe there is?) a marathon benchmark of some sort, like how long can your agent remain coherent while performing related and follow on tasks
we all do that "just this one little more thing" since before ai, measure it
Reposted by verdverm
I want this site to beat twitter, and for twitter to evaporate. For that to happen, people you hate must have a more pleasant experience posting here than posting there.
A minority of this site decided instead that they want it to be their little club, so that's what we have.
A minority of this site decided instead that they want it to be their little club, so that's what we have.
January 5, 2026 at 6:09 PM
I want this site to beat twitter, and for twitter to evaporate. For that to happen, people you hate must have a more pleasant experience posting here than posting there.
A minority of this site decided instead that they want it to be their little club, so that's what we have.
A minority of this site decided instead that they want it to be their little club, so that's what we have.
I remember having this book as a child, seems it's been appropriately updated lol
January 6, 2026 at 12:13 AM
I remember having this book as a child, seems it's been appropriately updated lol
Reposted by verdverm
Out: reacting without reading
In: reading then reacting
In: reading then reacting
January 5, 2026 at 11:13 PM
Out: reacting without reading
In: reading then reacting
In: reading then reacting
are tokens the metric unit of text length now?
I was thinking of using that on my blog instead of words or est. time
I wonder though if we could, or should, ever get a standard, or maybe a bunch and we can meme that xkcd comic a 16th time
www.youtube.com/watch?v=-R6o...
I was thinking of using that on my blog instead of words or est. time
I wonder though if we could, or should, ever get a standard, or maybe a bunch and we can meme that xkcd comic a 16th time
www.youtube.com/watch?v=-R6o...
The history of the Metric System in 7 minutes
YouTube video by Reece Batts
www.youtube.com
January 1, 2026 at 10:12 AM
are tokens the metric unit of text length now?
I was thinking of using that on my blog instead of words or est. time
I wonder though if we could, or should, ever get a standard, or maybe a bunch and we can meme that xkcd comic a 16th time
www.youtube.com/watch?v=-R6o...
I was thinking of using that on my blog instead of words or est. time
I wonder though if we could, or should, ever get a standard, or maybe a bunch and we can meme that xkcd comic a 16th time
www.youtube.com/watch?v=-R6o...
December 30, 2025 at 3:36 PM
@pfrazee.com it occurs to me that if we can run a full atproto network in @dagger.io, we can also run full network integration testing in ci too
#devops life
#devops life
December 30, 2025 at 11:00 AM
@pfrazee.com it occurs to me that if we can run a full atproto network in @dagger.io, we can also run full network integration testing in ci too
#devops life
#devops life
December 30, 2025 at 7:28 AM
December 30, 2025 at 3:25 AM
I step closer to the edge of the (ci) abyss...
this is something of a blend between make & argo
this is something of a blend between make & argo
December 29, 2025 at 7:44 AM
I step closer to the edge of the (ci) abyss...
this is something of a blend between make & argo
this is something of a blend between make & argo
December 28, 2025 at 6:46 AM
December 28, 2025 at 3:35 AM
<thinking>we have issues with online comms, can we repurpose stuff from ai for out-of-flow context sharing</thinking>
What if humans adopted the <thinking> tags in our online writing for each other? Would that help to convey more internal context so we have fewer disagreements?
What if humans adopted the <thinking> tags in our online writing for each other? Would that help to convey more internal context so we have fewer disagreements?
December 27, 2025 at 11:17 PM
<thinking>we have issues with online comms, can we repurpose stuff from ai for out-of-flow context sharing</thinking>
What if humans adopted the <thinking> tags in our online writing for each other? Would that help to convey more internal context so we have fewer disagreements?
What if humans adopted the <thinking> tags in our online writing for each other? Would that help to convey more internal context so we have fewer disagreements?
You know that advice in Dockerfiles, where you are supposed to...
apt-get update -y && \
apt-get install -y ... && \
apt-get clean && rm -rf /var/lib/apt/lists/*
...? well those days are over
a better option exists with @dagger.io that is faster and saves the same space, better caching of course!
apt-get update -y && \
apt-get install -y ... && \
apt-get clean && rm -rf /var/lib/apt/lists/*
...? well those days are over
a better option exists with @dagger.io that is faster and saves the same space, better caching of course!
December 27, 2025 at 11:37 AM
You know that advice in Dockerfiles, where you are supposed to...
apt-get update -y && \
apt-get install -y ... && \
apt-get clean && rm -rf /var/lib/apt/lists/*
...? well those days are over
a better option exists with @dagger.io that is faster and saves the same space, better caching of course!
apt-get update -y && \
apt-get install -y ... && \
apt-get clean && rm -rf /var/lib/apt/lists/*
...? well those days are over
a better option exists with @dagger.io that is faster and saves the same space, better caching of course!
