The Verbosity of Echo
Sam Ruby has called for people to start experimenting with the current (very early) Echo example feeds, and the response has been pretty impressive; check out these feeds from Joel Spolsky, Phil Ringnalda and Mark Pilgrim. Now that Echo has progressed to a stage where there are concrete feeds to examine, I have some serious concerns over the verbosity of the format. As they stand, Echo feeds contain a lot of duplicated information. Considering that the default behaviour of aggregators is to poll a feed for updates once an hour, any unnecessary information in the feed itself is going to have a very real monetary cost in terms of burnt bandwidth.
I’ve posted some further thoughts on this subject in Sam Ruby’s comments, but I think the Echo development community need to spare some thought for the size of the feed. I’ve seen the justification before that requiring elements rather than making them optional makes it easier to write Echo implementations and I agree that this is a worthwhile consideration, but I think the trade-off in terms of additional bandwidth costs needs to be given a higher priority.
To that end I’ve proposed a small alteration to Echo to remove redundant author information from entries. If you are following the development of Echo I encourage you to take a look.
On a related note, this Java blogger has some interesting thoughts on the WikiNature (or lack there-of) of the Echo Wiki.
More recent articles
- Understanding GPT tokenizers - 8th June 2023
- Weeknotes: Parquet in Datasette Lite, various talks, more LLM hacking - 4th June 2023
- It's infuriatingly hard to understand how closed models train on their input - 4th June 2023
- ChatGPT should include inline tips - 30th May 2023
- Lawyer cites fake cases invented by ChatGPT, judge is not amused - 27th May 2023
- llm, ttok and strip-tags - CLI tools for working with ChatGPT and other LLMs - 18th May 2023
- Delimiters won't save you from prompt injection - 11th May 2023
- Weeknotes: sqlite-utils 3.31, download-esm, Python in a sandbox - 10th May 2023
- Leaked Google document: "We Have No Moat, And Neither Does OpenAI" - 4th May 2023
- Midjourney 5.1 - 4th May 2023