Chris Lloyd

I've completed "Plutonian Pebbles" - Day 11 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/11

I just completed "Print Queue" - Day 5 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/5

I just completed "Ceres Search" - Day 4 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/4

I just completed "Mull It Over" - Day 3 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/3

I just completed "Red-Nosed Reports" - Day 2 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/2

I just completed "Historian Hysteria" - Day 1 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/1

I decoded my daily message on topsecret.spatie.be and got a nice prize!

I can't comment, because I haven't been able to play 🛩️ yet due to the issues! 🙂 Maybe I will try again today...

I hear it did, I'm hoping the same thing happens to #MSFS2024 eventually as they're having a bit of a launch disaster too 😉

Omg Hache is going into administration this is a disaster 🥺

This is a UDP joke but you might not get it.

This is probably the third thing I tried to tackle with string parsing recently that I ended up just sending to an LLM.

Works incredibly well, and cost effectively even if you need to parse hundreds of thousands of them.

"You have been provided with an email from a ticketing system. Please respond with the "body" of the email, removing any headers, greetings (such as Hi), signoffs (such as Thanks), case information, signatures etc. Retain the original formatting. Do not rewrite the content, reproduce it verbatim."

Tools like github.com/mailgun/talon/ and github.com/github/email... try to do this, but they're not perfect, it's too hard to deal with all the variations. But, a prompt like:

Where the thread is full of replies (quoted in various ways), attachment CIDs, signatures, headers about how the sender is outside your org, footers about how you need to delete the message if you're not the right recipient, and all in random formations.

I'm finding that LLMs (or SLMs) are really good at tasks that you might have tried to do with loads of regular expressions. For example, trying to parse out the "body" from the top of an email thread captured by a case management system.

Need to get started with this somehow. Maybe the dog?