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
the dogpiling is why when you zoom out, the metrics don't look good
bskycharts.edavis.dev/static/dynaz...
bskycharts.edavis.dev/static/dynaz...
December 30, 2025 at 4:40 AM
the dogpiling is why when you zoom out, the metrics don't look good
bskycharts.edavis.dev/static/dynaz...
bskycharts.edavis.dev/static/dynaz...
Looking at all the stuff that goes in as one list puts some perspective on why getting it setup can be a challenge
December 30, 2025 at 3:28 AM
Looking at all the stuff that goes in as one list puts some perspective on why getting it setup can be a challenge
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
![two commands showing (1) the config and service for the atproto network and (2) the storage components to go with them
---
# tony @ hydrogen in ~/hof/hof/lib/env/devex on git:_next x [0:26:47]
$ hof env list ./testnet.cue % server config -s kind
NAME KIND PATH
plc-config hostFile testnet.plc.config
relay-config hostFile testnet.relay.config
jetstream-config hostFile testnet.jetstream.config
pds-config hostFile testnet.pds.config
plc-server service testnet.plc.server
relay-server service testnet.relay.server
jetstream-server service testnet.jetstream.server
pds-server service testnet.pds.server
# tony @ hydrogen in ~/hof/hof/lib/env/devex on git:_next x [0:26:54]
$ hof env list ./testnet.cue -k cache
NAME KIND PATH
plc-pg-vol cache testnet.plc."pg-vol"
relay-pg-vol cache testnet.relay."pg-vol"
relay-data cache testnet.relay.data
jetstream-data cache testnet.jetstream.data
pds-pg-vol cache testnet.pds."pg-vol"
pds-data cache testnet.pds.data
pds-blobs cache testnet.pds.blobs](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:2jtyqespp2zfodukwvktqwe6/bafkreicxozzbj5msrmtnz6eezayocmty4wbad3amk63jm6p4pt7t7bjc7y@jpeg)
December 28, 2025 at 6:46 AM
December 28, 2025 at 3:35 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!
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!
do you mean the typical chat interface, if you are using Ai to write code, using an IDE with that chat on the side is the way to go
What my setup looks like now
chat | code & file diffs | git & ai changeset tree
What my setup looks like now
chat | code & file diffs | git & ai changeset tree
December 25, 2025 at 11:02 PM
do you mean the typical chat interface, if you are using Ai to write code, using an IDE with that chat on the side is the way to go
What my setup looks like now
chat | code & file diffs | git & ai changeset tree
What my setup looks like now
chat | code & file diffs | git & ai changeset tree
Confirmed, the cloning and splicing appears to work
December 25, 2025 at 2:08 AM
Confirmed, the cloning and splicing appears to work
my hacky code caused the model to switch from pro to flash on the second or third message, but it three tapped this feature I've been meaning to implement
The difference, my agents can finally run tests! I'm not even to the good tools yet lol
The difference, my agents can finally run tests! I'm not even to the good tools yet lol
December 24, 2025 at 10:22 PM
my hacky code caused the model to switch from pro to flash on the second or third message, but it three tapped this feature I've been meaning to implement
The difference, my agents can finally run tests! I'm not even to the good tools yet lol
The difference, my agents can finally run tests! I'm not even to the good tools yet lol
For those of you who don't know, I'm a big fan of Go, CUE, and Dagger and I'm finally starting to put all three together.
I just crafted this image for me and my little helper #agent so we can work on all the things. Go, Node, Python, and even ZSH
github.com/hofstadter-i...
#cuelang #dagger
I just crafted this image for me and my little helper #agent so we can work on all the things. Go, Node, Python, and even ZSH
github.com/hofstadter-i...
#cuelang #dagger
December 24, 2025 at 1:36 PM
For those of you who don't know, I'm a big fan of Go, CUE, and Dagger and I'm finally starting to put all three together.
I just crafted this image for me and my little helper #agent so we can work on all the things. Go, Node, Python, and even ZSH
github.com/hofstadter-i...
#cuelang #dagger
I just crafted this image for me and my little helper #agent so we can work on all the things. Go, Node, Python, and even ZSH
github.com/hofstadter-i...
#cuelang #dagger
Iterated back and forth with them getting to the same point
flash is way faster, very nice for iterating on UI
pro is slower, but certainly has a noticeable, though hard to quantify, grasp of the larger picture (?)
unclear who the real winner is here
which do you think is which... and/or better?
flash is way faster, very nice for iterating on UI
pro is slower, but certainly has a noticeable, though hard to quantify, grasp of the larger picture (?)
unclear who the real winner is here
which do you think is which... and/or better?
December 22, 2025 at 8:43 AM
Iterated back and forth with them getting to the same point
flash is way faster, very nice for iterating on UI
pro is slower, but certainly has a noticeable, though hard to quantify, grasp of the larger picture (?)
unclear who the real winner is here
which do you think is which... and/or better?
flash is way faster, very nice for iterating on UI
pro is slower, but certainly has a noticeable, though hard to quantify, grasp of the larger picture (?)
unclear who the real winner is here
which do you think is which... and/or better?
another +1 for flash
better tooltip style and most importantly not centered on the mouse, pro covered the graph...
follow up for both: "make the tooltip instantaneous and draw a highlighted vertical line"
notably, flash vs pro follow up response time: 4s vs 49s, 15% fewer tokens so far
better tooltip style and most importantly not centered on the mouse, pro covered the graph...
follow up for both: "make the tooltip instantaneous and draw a highlighted vertical line"
notably, flash vs pro follow up response time: 4s vs 49s, 15% fewer tokens so far
December 22, 2025 at 6:26 AM
another +1 for flash
better tooltip style and most importantly not centered on the mouse, pro covered the graph...
follow up for both: "make the tooltip instantaneous and draw a highlighted vertical line"
notably, flash vs pro follow up response time: 4s vs 49s, 15% fewer tokens so far
better tooltip style and most importantly not centered on the mouse, pro covered the graph...
follow up for both: "make the tooltip instantaneous and draw a highlighted vertical line"
notably, flash vs pro follow up response time: 4s vs 49s, 15% fewer tokens so far
The code...
1. Code is equivalent, nearly identical
2. They failed with the same error
From just reading it, I like a bit of both, but flash overall better... generally more succinct, name is more descript for a vertical lines denoting certain events (also considering more may come)
1. Code is equivalent, nearly identical
2. They failed with the same error
From just reading it, I like a bit of both, but flash overall better... generally more succinct, name is more descript for a vertical lines denoting certain events (also considering more may come)
December 22, 2025 at 6:03 AM
The code...
1. Code is equivalent, nearly identical
2. They failed with the same error
From just reading it, I like a bit of both, but flash overall better... generally more succinct, name is more descript for a vertical lines denoting certain events (also considering more may come)
1. Code is equivalent, nearly identical
2. They failed with the same error
From just reading it, I like a bit of both, but flash overall better... generally more succinct, name is more descript for a vertical lines denoting certain events (also considering more may come)
Took some code from a blog on the internet, one file for making svg graphs. Hacked on it to support multiple series and threshold lines, to draw token usage.
Now a/b testing my #agent with gemini-3-flash/pro (46s|74k / 2m|87k). They did the exact same thing func call wise, haven't delved code
1/
Now a/b testing my #agent with gemini-3-flash/pro (46s|74k / 2m|87k). They did the exact same thing func call wise, haven't delved code
1/
December 22, 2025 at 5:11 AM
Took some code from a blog on the internet, one file for making svg graphs. Hacked on it to support multiple series and threshold lines, to draw token usage.
Now a/b testing my #agent with gemini-3-flash/pro (46s|74k / 2m|87k). They did the exact same thing func call wise, haven't delved code
1/
Now a/b testing my #agent with gemini-3-flash/pro (46s|74k / 2m|87k). They did the exact same thing func call wise, haven't delved code
1/
Just did my first session having two #agents working in tandem, to implement a button so I can click to see what the system prompt will look like with the current config & state, as markdown in VS Code
1. focused on the backend
2. focused on the frontend
Sent the last backend to the frontend, oops
1. focused on the backend
2. focused on the frontend
Sent the last backend to the frontend, oops
December 21, 2025 at 3:55 AM
Just did my first session having two #agents working in tandem, to implement a button so I can click to see what the system prompt will look like with the current config & state, as markdown in VS Code
1. focused on the backend
2. focused on the frontend
Sent the last backend to the frontend, oops
1. focused on the backend
2. focused on the frontend
Sent the last backend to the frontend, oops
Billing in cloud looks reasonably up to date
December 20, 2025 at 12:04 AM
Billing in cloud looks reasonably up to date
Usage graphs for 1, 7, 28, 90 days, billing tomorrow (24h delay)
December 19, 2025 at 11:48 PM
Usage graphs for 1, 7, 28, 90 days, billing tomorrow (24h delay)
2 days with gemini-3-flash, def a great model
- implemented autocomplete for its input
- environment as vscode scm with diffs
- sorted agents.md files, even discovered and adjusted its own prompt without prompting
- enhanced the ui around tool calls
changes spanning languages and front/backends
- implemented autocomplete for its input
- environment as vscode scm with diffs
- sorted agents.md files, even discovered and adjusted its own prompt without prompting
- enhanced the ui around tool calls
changes spanning languages and front/backends
December 19, 2025 at 11:27 PM
2 days with gemini-3-flash, def a great model
- implemented autocomplete for its input
- environment as vscode scm with diffs
- sorted agents.md files, even discovered and adjusted its own prompt without prompting
- enhanced the ui around tool calls
changes spanning languages and front/backends
- implemented autocomplete for its input
- environment as vscode scm with diffs
- sorted agents.md files, even discovered and adjusted its own prompt without prompting
- enhanced the ui around tool calls
changes spanning languages and front/backends
So I kept going, instead of the usual advice to start fresh, because it kinda feels like it's in that butter zone of context, having seen a bunch of the code already, still less than 50k context... so I had it make some more changes, 4 turns ($0.20)
December 19, 2025 at 6:05 AM
So I kept going, instead of the usual advice to start fresh, because it kinda feels like it's in that butter zone of context, having seen a bunch of the code already, still less than 50k context... so I had it make some more changes, 4 turns ($0.20)