- Do you use on front door? Our VMs that don't are working fine, but our app services that do aren't.
- On our end, our VMs are still working, so our gitlab instance is still up. Our services using Azure App Services are available through their provided url. However, Front Door is failing to resolve any domains that it was responsible for.
- Oh I love this oh so much. It lines up with a lot of things I've wanted or am planning. So I'm definitely going to take a dive into your source for inspiration
One thing I wish for encompasses the idea that I might want to share different slices of my life with different groups
- It sounds like different levels of influence/control/responsibility to me.
Fancy validator selection sounds like the individual financial institutions are still responsible for managing and maintaining their nodes, which gives them a fair (as in balanced not fair as in a lot) amount of liability/responsibility/control.
A distributed database, afaik, while geographically distributed, entails more centralization of power/control.
- Super cool! I'm excited to explore it as an excalidraw alternative for a lot of the diagrams I make.
Some initial notes:
- On the very first initial load, ~~there's no default tool selected~~ it selects the 'hand' tool in the bottom left, which hides the cursor. It seems to select the pointer tool after a refresh. I'm able to replicate in an incognito browser or by clearing site data. After scanning your documents, it appears its touchscreen related. My guess is that it got confused because I'm using one of the ASUS Duo laptops. - I personally think it makes more sense to start docs on "What is Context" (on the documentation drawer), but that's a matter of opinion. The sandwich icon isn't as obvious as I'd like though, and it'd be nice to have a link to the next page at the bottom of every article. - my initial thought from the sequence diagrams panel was that I could type to generate an initial image, then drag that into my canvas for position changes. - this actually makes me think about mermaid's rendering engine, and how hard it would be to support moving and "pinning" an element, such that further diagram changes need to work around the pinned elements. - Yes yes yes please so much yes.
I love the idea of epicenter. I love open source local-first software.
Something I've been hacking on for a minute would fit so well, if encryption wasn't a requirement for the profit model.
But uh yes thank you for making my life easier, and I hope to return the favor soon
- I love the timing of how, over the past week or so I've been digging into the ecosystem and options around this, and now here's another article!
- > As a side note, it's becoming increasingly important to write down this info in places where LLMs can access it with the right context. Unfortunately commit history is not one of those spots.
This is the comment that spawned this tragedy of miscommunication.
My interpretation of this comment is that no current programming agents/llm tooling utilize commit history as part of their procedure for building context of a codebase for programming.
It is not stating that it Cannot, nor is it making any assertion on whether these assistants can or cannot be Trained on commit history, nor any assertion about whether commit history is included in training datasets.
All its saying is that these agents currently do not automatically _use_ commit history when finding/building context for accomplishing a task.
- The users want to read an article, I'm not sure the average user really cares if it's delivered as a document or an app
- In other words, a depository is cold storage
- Makes sense as a broad pattern!
I guess I hadn't considered the existence of generic TCP proxies And I'd probably have concerns around how much latency it would introduce in an environment where we have some requirements on the rate of data collection.
That said our service Does act as a kind of proxy for a few protocols.
- I tried this at work but ran into issues where if the server had some error or issue or took longer than expected, the job queue grew too large and caused OOM issues. I had to turn it into a manual list in order to debug the problem, though.
Plus we have a case where a certain type of request should skip to the front of the queue.
Leaning on promises does cut out a lot of the user space complication though.
- I've had to implement this exact logic at work, because we have to talk to devices using modbus tcp, where a lot of devices only supports having one active request per connection at a time. One device we talk to only supports having 5 active connections to it.
- Reminds me of this project
A game of learning your homelab into a cyberpunk mystery adventure | Hacker News https://share.google/WedMuRgx5WgreNqSN
- MCPs aren't the only way to embed tool calls into an LLM
- It may not even have to be an official service, just something that shows up on a CC statement as something other than a "donation".
Or hey maybe they could charge for these expiration notifications.
- I find it has a lot to do with activation energy, or how much friction there is.
If something is too easy, it happens. If something is too hard/annoying to do relative to the reward, it doesn't.
When I lived in Boston, I was able to get myself to go to Tae Kwon Do classes because it was literally on my commute. When I'd work from home, I'd end up missing the class. When covid hit, same.
When I lived in Palo Alto, I rarely went to the city (and usually only when I could crash at a friend's place for the weekend) because the process of biking to caltrain, take caltrain, bike to wherever I wanna go, then reverse on my way home or catch an Uber, kinda ends up being an annoying or expensive process.
When I moved to SF, I went and partied what was probably more frequently than was healthy. Moving to Oakland brought the rate back down again, but to a level that felt a little too isolating. Getting a car seems to hit a sweet spot; driving isn't too bad, and I have a lot more options. Buuut being responsible for a car curbs the worse behaviors.
- Hey I mean I'm not super likely to use both tiger data and tiger graph, but using both tailscale and timescale has resulted in some awkward mixups
- That's a good reason! I think with UUIDv7, because it's sequential it's indexes are faster than UUIDv4. Still larger than bigints though. I'd like to do a benchmark at some point
There's less suffering, sure. But if I were in their shoes I'd want to have a choice. To be manipulated into wanting something so very obviously and directly bad for us doesn't feel great